Examining framings of geoengineering using Q methodology

    Examining framings of geoengineering using Q methodology Rose Cairns Climate Geoengineering Governance Working Paper Series: 002. Published online 6 June 2013       Climate Geoengineering Governance (CCG) Climate Geoengineering Governance (http://geoengineeringgovernanceresearch.org) is a research project which aims to provide a timely basis for the governance of geoengineering through robust research on the ethical, legal, social and political implications of a range of geoengineering approaches. It is funded by the Economic and Social Research Council (ESRC) and the Arts and Humanities Research Council (AHRC) - grant ES/J007730/1 CGG Working Papers The CGG Working Paper series is designed to give a first public airing to a wide range of papers broadly related to the project’s themes. Papers published in this series may be, but are not necessarily, early outputs from the project team; equally they may be from other authors, and reflect different perspectives and different issues from those directly pursued by the project itself. The aim is to promote vigorous and informed debate, in a spirit of pluralism. What the working papers have in common is that they will all be at an early stage of development, prior to full publication. Comment and response, at any level of detail, is therefore doubly welcome. Please send all responses in the first instance to the authors themselves - each paper contains a correspondence address. We will be looking for opportunities to use the website or other project activities to give a wider airing to any dialogues and debates that develop around a paper or issue. About the author Rose Cairns ([email protected]) is a research fellow at SPRU – Science and Technology Policy Research, at the University of Sussex. Her primary research interests are in environmental politics and governance, participatory research methods and discourse analysis, and the theory and practice of interdisciplinary research for sustainability. Rose completed her PhD on conservation politics on the Galápagos Islands from Leeds University in 2012. She also holds an MSc in Conservation and Biodiversity from Exeter University, and a BA in Social Anthropology from Cambridge University. Prior to her recent roles in academia Rose worked for a number of years in environmental campaigning and the voluntary sector.       Examining framings of geoengineering using Q methodology Rose Cairns SPRU – Science and Technology Policy Research University of Sussex June 2013 Abstract Despite (or perhaps reflecting) widespread awareness of its ambiguity, the term ‘geoengineering’ has in recent years become massively more prominent. Academic, policy and civil society circles routinely use this term to describe, support or oppose a diverse range of techniques and ideas. This study aims to contribute to understandings of ways in which variously envisaged approaches to ‘geoengineering’ of the global climate are currently being framed. It asks not only about variously viewed implications of geoengineering itself, but also what these diverse framings can reveal about wider politics in contemporary debates around climate change, science and technology. The paper applies Q methodology to analyse geoengineering as a subjective discursive construct, the bounds of which are continually negotiated and contested. 35 participants from a variety of disciplinary and institutional backgrounds in the UK, US, Canada and Japan undertook a ‘Q sort’ of 48 opinion statements about geoengineering between December 2012 and February 2013. Four distinctive framings emerged from this analysis, labelled: ‘At the very least we need more research’; ‘We are the planetary maintenance engineers’; ‘Geoengineering is a political project’; and ‘Let’s focus on Carbon.’ Results indicate a strong polarity around divergently-construed pros and cons of geoengineering as a whole – underscoring the political salience of this term. But additional axes of difference suggest a more nuanced picture than straightforward pro/anti positioning. The ambiguity of the term is argued to offer interpretive flexibility for articulating diverse interests within and across contending framings. The paper questions whether increasing terminological precision will necessarily facilitate greater clarity in governance discussions or public engagement, and argues that the merits of any given form of precision will depend on particular framings. Much ambiguity in this area may thus be irreducible, and hence the challenge lies rather in realising the wider implications of the political pluralities this reveals. Introduction This paper explores divergent societal framings of variously-envisaged approaches to ‘geoengineering’ of the global climate (henceforth geoengineering). A diverse array of technologies are commonly referred to as geoengineering, including those that have been collectively labelled as solar     radiation management methods (such as stratospheric aerosol injection and marine cloud brightening), and those referred to as carbon dioxide removal methods (such as ocean iron fertilization, or direct air capture) (Shepherd et al. 2009). Although the term geoengineering has become increasingly prominent in discussions of these approaches in scientific, policy, and civil society circles, there is evidence of a growing sense that the label itself may be so broad and ambiguous as to be unhelpful, or even incoherent. Thus the recent report from the IPCC ‘expert meeting on geoengineering’ draws attention to the ‘fuzzy’ boundary between geoengineering and other approaches to dealing with climate change, and suggests that ‘because of the longstanding ambiguity surrounding the term geoengineering … the individual methods discussed might be referred to more specifically’ (Edenhofer et al. 2012, p.3). This refrain about the need to look at different technologies and approaches separately runs through many other reports on geoengineering, and yet many of them (like the IPCC report) retain the word geoengineering in their titles (Shepherd et al. 2009; GAO 2010). In one such report by the US think tank, the Bipartisan Policy Centre (Long et al. 2011), debates around whether the term geoengineering ‘was too imprecise…[or] too controversial’ (Sarewitz 2011, p.7), actually resulted in the appearance of the additional (equally imprecise) term ‘climate remediation’ being used alongside geoengineering in the title. Given the widespread awareness of the ambiguity of the term, and the difficulties this poses for meaningful governance, is it the case that the term geoengineering can be said to have simply outgrown its usefulness? Or is it that, as has been argued to be the case for terms such as ‘sustainability’ or ‘sustainable development’ (Baker et al. 1997) it is the very ambiguity of the terms that might make them particularly useful to certain actors and give them a resilience they might not otherwise have had? In this study rather than seeing ambiguity as ‘a linguistic veil which can be lifted to reveal the truth’ (Rydin 1999, p.468), and attempting to remove it by carrying out our own boundary work to define a sub-set of technologies or approaches as our object of study, we embrace the ambiguity of the term geoengineering. This study is thus distinct from previous work on frames and framing of geoengineering, much of which starts by offering a definition of geoengineering as the object of study (Sikka 2012; Luokkanen, Huttunen & Hildén 2013; Huttunen & Mikael Hilden 2012; Scholte et al. 2013; Nerlich & Jaspal 2012).1 This paper takes a fundamentally different approach: rather than treating geoengineering as an object, a ‘novel controversial technology’ (Luokkanen, Huttunen & Hildén 2013), or even a set of technologies about which there exists an array of sometimes conflicting opinions, or for which there is support or opposition, this study treats geoengineering as a discursive phenomenon, the bounds of which are continually being negotiated. This is in                                                                                                                         1  An  exception  might  be  the  work  of  Bellamy  et  al.(Bellamy  et  al.  2012)  who  do  explicitly  explore  the  ambiguity   in  the  term  at  the  outset  of  their  paper,  however  there  is  a  subsequent  attempt  to  minimise  this  ambiguity  by   adopting  a  singular  definition  (ibid  p.  7).       line with insights from post-positivist policy analysts such as Hajer who have drawn attention to the fact that environmental conflict should not be understood as ‘a conflict over a pre-defined unequivocal problem with competing actors pro and con,’ but rather should be seen as ‘a complex and continuous struggle over the definition and meaning of the environmental problem itself’ (M. A. Hajer 1997, p.14). Focusing analytical attention on the inherently ambiguous, undifferentiated category ‘geoengineering’ (a term that has been referred to as a ‘quasi-stable meta-label’ (Porter & Hulme 2013, p.3)), is argued to be the best way to identify (rather than impose) the most significant axes for distinction, as these relate to key differences in divergent perspectives. Rather than working to remove ambiguity from the term geoengineering then, this study shifts the focus to an exploration of the kinds of work it does (in spite of, or because of the ambiguities within it), the boundary work it prompts and the tensions and ambivalences it inspires and reveals. We ask what these can reveal about the kinds of politics at work in contemporary debates around climate change, science and technology. Framing geoengineering There is a small but growing body of social scientific literature examining discourses and framing of geoengineering, including work examining framings of geoengineering in the media (Porter & Hulme 2013; Scholte et al. 2013; Luokkanen, Huttunen, Hildén, et al. 2013), academic literature (Bellamy et al. 2012; Huttunen & Mikael Hilden 2012), work examining the use of metaphor in framings of geoengineering (Nerlich & Jaspal 2012), or examining the framings within particular influential texts (Gardiner 2011). While a number of common themes (for example the importance of ‘climate emergency’ as a framing device) emerge from this work, a diversity of findings have been presented regarding the relative openness or otherwise of the discourse around geoengineering, or the relative importance of strategic framing to the issue. Given that the term is arguably still unfamiliar to many people, some have argued that the ‘first impression, frame, and narrative has yet to be set’ (Leiserowitz 2010, cited by Buck 2013), or suggested that there is a need for more active and strategic framing of the issue by scientists in particular ways (Buck 2013). On the other hand others have suggested that the way that appraisals of geoengineering options have been carried out to date, provide evidence of a premature ‘closing down’ around particular ‘sets of values and assumptions with respect to the instrumental framing effects of contexts, methods and criteria and options’ (Bellamy et al. 2012, p.28), or cite evidence from analysis of the metaphors used to describe geoengineering as indicative of ‘restrictions in the interpretative flexibility’ of the term (Luokkanen, Huttunen & Hildén 2013). Sikka takes a particularly strong view of the strategic nature of the framing of geoengineering to date, arguing that ‘special interests, including private corporations, conservative think tanks and scientists affiliated with both have drawn on a variety of discursive frames to limit, shape and mould the current debate     surrounding geoengineering’ (Sikka 2012, p.173). Conversely others have drawn evidence from an analysis of the changing frames of geoengineering apparent in English speaking newspapers in recent years, to argue that there is evidence of a progressive ‘opening up’ of the debate around geoengineering (Scholte et al. 2013). This study falls broadly under the description of a frame-reflective analysis outlined by Schon and Rein (Schön & Rein 1995), and complements and builds upon the corpus of work on framing of geoengineering by bringing a distinctive focus on the ambiguity of the term, as outlined above. Within this study, frames are understood as ‘schemata of interpretation’ (Goffman 1974, p.21), or narratives of understanding that ‘help to render events meaningful and thereby function to organize experience and guide action’ (Benford & Snow 2000, p.614). Crucially, frames have both ontological and normative dimensions in that they ‘link causal accounts of policy problems to particular proposals for action, and so link accounts of ‘is’ and ‘ought’ (Rein and Schon cited in Fischer and Forester 1993, p. 11). Hoppe (1999) emphasises that frames are necessary for judgement and action, acting as ‘a sort of mental grappling hook’ (p.207) to enable people to make sense of and act on a given situation. Through selectively emphasizing certain facets of a given issue over others, and linking interpretation with action, frames in and of themselves can be understood to perform particular functions (c.f. Entman 2004). However there may be a distinction made between the use of the verb framing and the resulting frames or framings as nouns, with the former denoting ‘an active processual phenomenon that implies agency and contention at the level of reality construction’(Benford & Snow 2000, p.614). Although not exclusively a conscious phenomenon, a rich literature from the study of social movements has highlighted the ways in which framing can be used as a tool to affect change. In this sense, framing can be understood as ‘the conscious strategic efforts by groups of people to fashion shared understandings of the world and of themselves that legitimate and motivate collective action’ (Snow, in McAdam, McCarthy and Zald, 1996:6 Emphasis added). Although there are several different quantitative and qualitative methods applicable to framing analysis, this study uses Q method (see Box 1), to create a typology of frames in a systematic way so as to render these shared subjective constructions of the world observable (c.f. Dayton 2000). While various framing analyses have explored discursive elements in isolation (e.g. metaphors), in its application of Q method, this work aims to use an operant approach to allow participants autonomy to bring these framing elements together to constitute frames. Data collection     The ‘Q sort’ is the basic unit of data in a Q study, and consists of a selection of subjective statements about the topic of interest, rank-ordered by a participant according to a particular instruction (for example according to those that are most-to-least like the participant’s point of view). To select the statements that are to be sorted, the researcher first gathers as comprehensive as possible a selection of statements to reflect the diversity of opinions about the subject. This selection of statements is known as the ‘concourse’, and represents an attempt to capture the ‘volume of discussion’ on a given topic (Brown 1986, p.58). In this case, subjective statements about the topic of geoengineering were sought from a diverse range of sources, including academic papers, government policy documents, NGO reports, scientific and popular news media sources, television and radio interviews, blog posts and comments on online news sites. It is common practice within Q studies to include some statements that are ‘deliberately ambiguous’ (Dryzek & Berejikian 1993) or contain ‘excess meaning’ (Brown 1970). Because the statements do not just have one objective meaning set by the researcher, but are intended to act as stimuli to reveal the internal frames of reference of the participants, this ambiguity is not problematic as it would be in, for example a questionnaire design. Box 1. Q method overview Q is a ‘quali-quantitative’ method that can be used to examine the subjectivity inherent in any given topic around which there is social contestation, and as such, lends itself particularly well to the study of frames (c.f. Dayton 2000). Q has disciplinary roots in psychology (Stephenson 1935), but is now commonly applied across a range of disciplines including political sciences (Brown 1980; Dryzek & Berejikian 1993), geography (Eden et al. 2005); ecological economics (Swedeen 2006) and environmental policy analysis (Webler et al. 2009; Addams & Proops 2000). Q is an intensive ‘small n’ method in which a number of purposively selected participants (usually between 20 – 50) are asked to rank order a number of statements about a given topic. Outcomes are then statistically analysed using factor analysis to look for patterns in ways of thinking and talking about the topic (i.e. frames). Interview data and comments from the participants are used to aid interpretation of these patterns, and potentially to gain an insight into the more or less active processes of framing that different actors undertake. Although it has quantitative features, the method has a large qualitative component, and as with any other methodology, must be ‘employed reflexively and creatively, with full awareness of its interpretative dimensions and not as a number-crunching exercise’ (Eden, Donaldson, & Walker, 2005, p. 421). Q methodology does not impose categories of discourse onto the data, a priori, or position participants with respect to some predefined framework. Rather it asks its participants to decide what is or is not meaningful and relevant to their opinion by the process of sorting of statements. It has been argued that the method’s potential to reveal discourses that might otherwise be obscured, might facilitate processes of ‘opening up’ policy to reflexive appraisal (Stirling 2007; Ockwell 2008), and proponents of the method have claimed that by allowing the researcher to ‘surrender the monopoly of control in their relationship with the researched’ Q method     can contribute to a more democratic research design and implementation (Robbins & Krueger 2000, p.636). The concourse in this study consisted of 322 statements, after which point the research team decided that the addition of further statements did not add to the diversity of opinions present, and that ‘saturation point’ (Eden et al. 2005) had been reached. To narrow down the concourse to the sample of statements to be presented to participants, a structured approach was adopted whereby statements were categorised into a number of themes that were observed in the concourse as a whole. These were: 1) context (the nature of ‘the problem’); 2) definitions and characteristics of geoengineering; 3) appraisals of geoengineering; 4) the relationship between science/research and deployment; and 5) governance concerns. Approximately equal numbers of statements from each category were sought, with the aim that each would capture a particular dimension of the issue around which opinion might be divided. The choice of the number of statements to be included in the final Q sample must balance the need to incorporate as great a diversity of statements as possible, against the need not to overly tax the abilities and patience of the participants. In line with a rule of thumb that suggests a Q sample size of between 20 – 60 statements (Webler et al. 2009, p.15), the final sample consisted of 48 statements. A pilot was carried out with 7 project members and colleagues in order to test the clarity of the statements, the comprehensiveness of the themes and topics covered by the statement sample, and the ease with which it was possible to sort them. Following the pilot a number of statements were removed because they were felt to be confusing or to duplicate existing themes in the sample, others were paraphrased for greater clarity, and a number of additional statements were added to cover themes that were felt by pilot participants to be missing. The final set of 48 statements is listed in Table 1. Participants and the Q sort The aim of this study was to uncover the different framings of the term geoengineering, hence the approach to participant selection was to focus on people who are or have been involved in debates and discussions around geoengineering, or those (self- selected) who would consider themselves to ‘have an opinion’ on geoengineering. The aim was not to try to elicit the views of some imagined wider ‘public’ as such, and no claim can be made that the sample group was in any sense representative of a larger population, but this is not the aim of a Q study (See Box 2). Rather, participants were selected on the basis that it was felt that they had the potential to reveal something interesting about the way in which debates around geoengineering are structured and the existing frames and framing strategies that are being employed. Based on an initial review of the academic and non-academic literature on the topic, a list of participants was drawn up to encompass a range of people involved in making     statements about geoengineering, and attempts were made to incorporate as diverse as possible a group of people (from different disciplinary backgrounds, sectors, nationalities and genders). This was complemented by a snowballing approach whereby participants were asked to identify other possible recruits with opinions that might differ from their own. Additionally an invitation to participate was circulated to the geoengineering Google list (an online forum for discussion on geoengineering 2), Geoengineering Net Forum (a Japanese discussion forum3) and the African Technology Policy Studies Network4. Given that the interest of the study was in revealing framings, rather than making any attempt to test the validity of viewpoints, the sole criterion for participation for those people responding to the general invitation was that participants considered themselves to ‘have an opinion’ about geoengineering. This criterion was applied based on the rationale that in order to have formulated an opinion on geoengineering an individual would need to have engaged with debates and discussions around the topic in some way, and their opinion would thus be as representative as anyone else’s of the types of framings of the issue that are emerging. Box 2. The notion of ‘sample size’ and generalizability of findings in a Q study Within traditional quantitative (‘R’) methods, the term ‘sample size’ refers to the number of participants, and is ideally as large as possible in order to be statistically representative of the larger population from which participants have been sampled, and thus to make it possible for inferences to be made about that population on the basis of the results of measurements of the sample. In Q methodology, the concept of sample size is more applicable to the concourse of statements, and the notion of representation is relevant in as far as the statements in the concourse should be representative of the total range of statements being made about the topic. Thus participants for a Q study are not randomly sampled from a population, but are deliberately chosen for their relevance to the topic in question (Brown 1980). The most important principle of participant selection is diversity of opinion, so that ideally if a particular discourse exists, even if is very marginal, the process would hope to reveal it. It is also important that the participants are familiar with the topic and have ‘well-formed opinions’ (Webler et al., 2009 p. 9 ). Given that the aim of a Q study is to search for distinct subjective viewpoints or framings of a given issue, and no claim is made about the proportions of the views uncovered in a wider population, the same need for large sample sizes does not apply in a Q study. Rather as Dryzek & Berejikian point out ‘our units of analysis, when it comes to generalization, are not individuals but discourses’ (Dryzek & Berejikian 1993). Thus although no claim can be made that the subjects who carried out the Q test are                                                                                                                         2  https://groups.google.com/forum/?fromgroups#!forum/geoengineering    http://geoeng.brs.nihon-­‐u.ac.jp/   4  http://www.atpsnet.org/   3     statistically representative of some larger population this is not the aim of a Q study. Instead in so far as the concourse is ‘representative’ of the breadth of opinion on the topic each factor described should ‘prove a genuine representation of that discourse as it exists within a larger population’ (Dryzek and Berejikian, 1993 p. 52). And thus although it cannot be asserted that the factors uncovered by this study are the only viewpoints that exist on the topic, the discovery of factors other than those described (for example through the participation of an additional individual with a unique point of view) should ‘in no way influence description’ of the existing factors (Brown, 1980 p. 67). Furthermore, within a Q study, individual cases are not treated as anomalies, or insignificant, but can provide valuable insights to the topic in question. As Brown explains, given that ’the interest of Q methodology is in the nature of the segments [discourses] and the extent to which they are similar or dissimilar, the issue of large numbers, so fundamental to most social research, is rendered relatively unimportant. In principle as well as practice, single cases can be the focus of significant research’ (Brown, 1993 p. 93). Participants were asked to sort the statements into a grid along a scale from +4 (most like their point of view) to -4 (least like their point of view). As is common in Q studies, the grid had a pyramidal or ‘quasi-normal’ shape, which limited the number of statements that could be placed in each category (See Figure 1). Although the imposition of this distribution shape is not necessary for the technique to work (Brown 1971; Burt 1972; Barry & Proops 1999), it is considered good practice as it encourages the participants to consider the relative placement of the statements more carefully and hence to reveal their preferences more thoroughly (Webler et al. 2009). In the majority of cases participants carried out a Q sort during a face-to-face interview with the researcher, and (with participant consent) comments made at the time of the Q sort were recorded to aid interpretation. Interviews lasted between 25 minutes up to 1.5 hours, averaging approximately 1 hour. In addition, to maximise the diversity of the participant group and facilitate participation from international participants, there was an option for participants to take part via an online interface using Q-Assessor (http://q-assessor.com ), a tool specifically designed for online Q studies. The online study can be accessed (and carried out) by visiting http://q-assessor.com/studies/753/responses/new. The use of a combination of face-to-face Q sorts and online sorts has precedents in the literature on Q method (e.g. Gruber 2011) and is supported by empirical work which has shown there to be no apparent difference in the reliability or validity of face-to-face sorts and those carried out remotely by mail (Van Tubergen & Olins 1979); between paper sorts and online sorts in general (Hogan     2010); and between paper sorts and the specific online sorting program we applied in this study, Q-Assessor (Reber et al. 2000).5 Thirty-five people, including 5 people involved with the Geoengineering Governance Project in various ways, carried out a Q sort. The inclusion of researchers’ own Q sorts is considered good practice in scholarship on Q methodology (Robbins & Krueger 2000; Swedeen 2006). Given that there is no such thing as a truly ‘frame neutral position’ (Schön & Rein 1995, p.38) from which to examine framings, in this study it was decided that to enhance both the reflexivity and transparency of the Climate Geoengineering Governance Project, it would be desirable both for the lead researcher and others less directly involved in this component of the project to reflect upon and reveal their own positionality within these debates. With participant permission, the list of participants and their institutional affiliations (where applicable) are listed in Appendix I. Twenty-one individuals carried out face-to-face interviews, while thirteen people carried out the Q sort online6. Participants who carried out a web-based sort are distinguished in the results table by the letter (W), while individuals associated with the Climate Geoengineering Governance project, including the author, are distinguished by the letters CGG. -4 -3 -2 -1 0 +1 Least like my point of view view +2 +3 +4 Most like my point of Figure 1. The distribution shape onto which participants were asked to sort the statements in the Q sample.                                                                                                                         5  Although  more  qualitative  data  can  be  collected  from  face-­‐to-­‐face  interviews  as  compared  to  online  sorts   (and  hence  a  possible  bias  emerges  in  the  fact  that  these  sorts  might  be  more  fully  interpreted  than  the  online   sort  patterns),  the  benefit  to  the  overall  study  of  being  able  to  incorporate  a  greater  diversity  of  participants   was  felt  to  outweigh  this  concern.       6  The  author’s  own  Q  sort  brought  the  total  number  of  sorts  to  35.       Analysis The 35 sorts were analysed with the freely available software PQMethod (Schmolck 2002). Each sort was correlated with every other sort, and a 35 by 35 correlation matrix was generated. Principal components analysis was then used to identify clusters of similarly performed Q sorts, and the resulting factors were rotated using a varimax rotation that aimed to find the simplest structure in the data and to explain the greatest amount of variance.7 It is important to bear in mind that there is not just one objectively ‘correct’ or ‘mathematically superior’ solution regarding the number of factors that emerge from a Q study (Watts & Stenner 2005, p.80). Rather, although the data itself is ‘fixed’ in the sense that the correlation scores between individual Q sorts do not change, there could be many vantage points from which to view and describe the similarities and differences between views, that are largely dependent on what one is interested in (for example, whether one is interested particularly in revealing minority views, or examining more dominant discourses). In this study a solution was sought that maximised the simplicity, clarity, distinctness and stability of the emerging framings (Webler et al. 2009, p.31), and ensured that at least 2 individual Q sorts correlated uniquely with each factor (cf. Brown, 1980 p. 293). Based on these criteria, a three-factor solution was selected as optimal. However, scrutiny of the results revealed that one of the factors was ‘bipolar’: certain individuals’ sorts were highly positively correlated with this factor, while others were highly negatively correlated, indicating the presence of two groups of people who sorted the statements in more or less opposite ways. In line with standard practice in Q studies (Brown 1980), the bipolar factor was split into two separate factors (highly negatively correlated with one another) resulting in a final solution consisting of the four factors that will be described below. Correlations between an individual’s Q sort and a given factor were deemed as being statistically significant at the p<0.01 level, if they exceeded a factor loading of +/-0.38, based on the equation: 2.58 x (1/√n), where n = the number of statements in the Q sample: 2.58(1/ √48) = 0.3723 (Brown 1980). Sorts that were significantly correlated with a factor (i.e. those that load at +/0.38 for that factor) were considered representative of that view, and the weighted average of those sorts were used to calculate an idealised sorting pattern for that factor along the original response scale (-4 to +4). These idealised sorting patterns are listed in Table 1. The degree to which each                                                                                                                         7  It  is  also  possible  to  extract  factors  using  the  centroid  method,  and  to  rotate  factors  ‘by  hand’  in  order  to   explore  the  data  based  on  particular  theoretical  hypotheses  about  how  particular  clusters  relate  or  how   particular  individual’s  sorts  fit  into  the  overall  picture,  and  some  researchers  argue  that  this  is  the  preferable   approach  (Brown,  1980).    In  this  instance  both  procedures  were  carried  out  and  the  results  compared.    While   not  substantially  different,  principal  components  analysis  and  varimax  produced  the  clearest  and  most  stable   and  interpretable  result.       participant’s sort correlated with each factor is described is given in Table 2. The degree of correlation between factors is given in Table 3. Narrative descriptions of each factor were drafted by examining these idealised sorting patterns and analysing the interview comments made by those people’s whose sorts were significantly correlated with that factor. Draft descriptions of these factor narratives were sent to all participants, who were asked to comment on whether they felt that their views had been appropriately represented. These comments were used to test the validity of the views described.     Table of statements sorted by participants, and the idealised sorting pattern (from -4 to +4) for each factor Statement:   Idealised  sort  pattern     1   2   3   4   1.  Geoengineering  is  about  fruitlessly  trying  to  solve  problems  with  the  same  mind-­‐ set  that  created  them:  attempting  to  control  nature.   -­‐3   -­‐4   2   -­‐3   2.  Geoengineering  can  only  ever  be  fundamentally  undemocratic.   -­‐2   -­‐1   1   -­‐4   3.  The  risks  of  not  doing  geoengineering  research  outweigh  the  risks  of  doing  it.   3   2   -­‐2   1   4.  Hasty  pursuit  of  international  regulation  of  geoengineering  risks  lock-­‐in  to   commitments  that  might  soon  be  regretted,  such  as  a  total  ban  on  research  or   testing,  or  burdensome  vetting  of  even  innocuous  research  projects.   1   3   -­‐2   0   5.  Calls  for  more  science  on  geoengineering  don't  really  make  sense:  full  trials  are   unethical  and  small  scale  experiments  are  pointless  (since  any  impacts  are  drowned   in  the  noise  of  global  weather).   -­‐4   -­‐3   2   -­‐3   6.  We  might  once  have  trusted  nature  to  look  after  the  environmental  regulation,   but  not  any  more:  like  it  or  not,  we  are  the  planetary  maintenance  engineers.     1   3   -­‐2   -­‐1   7.  Deliberate  geoengineering  of  the  climate  has  been  happening  for  decades;  the   current  suggestion  that  it’s  all  about  dealing  with  climate  change  is  just  a  front   intended  to  legitimize  on  going  activities  that  have  other  motivations.     -­‐4   -­‐2   0   -­‐1   8.  One  of  the  central  motivations  for  the  growing  interest  in  geoengineering   research  is  the  potential  for  strategic  military  applications  of  these  technologies.       -­‐4   -­‐2   1   1   9.  The  belief  that  technological  solutions  can  be  found  to  social  problems,  and  to   problems  arising  in  earlier  technological  development,  is  a  dangerous  illusion  which   fails  to  address  political  and  social  drivers  and  implications.   -­‐1   -­‐1   3   -­‐1   10.  Geoengineering  is  a  bit  like  chemotherapy:  the  decision  to  undertake  it  would   be  difficult,  but  it  could  turn  out  to  be  the  least  bad  option  we  are  going  to  have.     2   1   -­‐3   -­‐2   11.  Carbon  emissions  will  never  be  reduced  to  zero  because  this  would  require  a   complete  change  in  the  way  humans  are.   0   2   0   -­‐4   12.  At  the  very  least,  we  need  to  do  more  research  in  order  to  learn  what   approaches  to  avoid  even  if  we  become  desperate.   4   2   0   3   13.  It  is  inappropriate  to  make  geoengineering  research  decisions  subject  to   'upstream  engagement'  or  public  control.   -­‐2   0   -­‐4   -­‐2   14.  Geoengineering  is  the  most  revolutionary  and  potentially  valuable  new  idea  in   climate  policy  today.   -­‐3   2   -­‐1   -­‐3   15.  Geoengineering  is  potentially  the  key  to  unlock  the  mitigation  puzzle:  a  way  of   controlling  climate  risks  during  the  many  decades  that  it  will  take  to  transform  the   global  energy  system.   0   1   -­‐3   1   16.  The  idea  of  actually  deploying  a  geoengineering  system  would  be  very   controversial,  but  the  narrower  question  of  a  research  program  should  not  be.   3   -­‐1   -­‐1   -­‐1   17.  The  climate  system  is  too  complex  and  chaotic  to  judge  cause  and  effect  of   various  geoengineering  ideas,  even  as  we  try  them  out.   -­‐2   -­‐2   1   4   18.  After  realizing  that  our  actions  en  masse  affect  the  climate,  anything  we  do  to   address  it  (including  nothing),  is  geo-­‐engineering.   -­‐3   1   -­‐2   3   19.  The  technical  community  has  a  responsibility  to  explore  back-­‐up  strategies  for   dealing  with  an  unexpected  climate  emergency.   1   2   -­‐1   1   20.  We  use  technology  to  clean  up  land  and  water,  so  why  not  clean  up  the   1   1   0   1       contaminated  sky?   21.  Attempts  by  us  to  regulate  the  Earth's  climate  and  chemistry  would  condemn   humanity  to  a  Kafkaesque  fate  from  which  there  may  be  no  escape.   -­‐1   -­‐4   2   -­‐2   22.  The  laudable  goal  of  combating  climate  change  has  no  business  in  the  definition   of  geoengineering,  as  it  suggests  that  technologies  do,  in  fact,  combat  climate   change  giving  the  whole  suite  of  planet-­‐altering  technologies  a  veneer  of   respectability  they  have  not  earned.       0   -­‐1   1   -­‐1   23.  Some  geoengineering  technologies  might  be  able  to  provide  win-­‐win  solutions,   allowing  economic  growth  and  climate  change  mitigation  to  proceed  hand  in  hand.   0   2   0   2   24.  More  than  a  set  of  technologies,  geoengineering  is  a  political  strategy.   -­‐1   0   3   2   25.  Until  there  has  been  a  full  debate  on  the  course  all  countries  wish  to  go,  it  is   common  sense  to  institute  a  moratorium  on  all  geoengineering  activities  outside  the   laboratory.     -­‐2   -­‐4   4   2   26.  Much  current  policy  discussion  of  'geoengineering  governance'  makes  no   reference  to  the  possibility  that  the  world  could  decide  not  to  go  down  this  path,   and  is  thus  little  more  than  a  marketing  exercise.   -­‐2   -­‐1   1   0   27.  Commercial  involvement  in  geoengineering,  including  competition,  may  be   positive  in  that  it  mobilizes  innovation  and  capital  investment,  which  could  lead  to   the  development  of  more  effective  and  less  costly  technologies  at  a  faster  rate  than   in  the  public  sector.   1   1   -­‐2   1   28.  'Encapsulated'  geoengineering  technologies  are  ethically  preferable  to  non-­‐ encapsulated  technologies.   2   -­‐1   0   1   29.  In  a  geoengineered  world,  rather  than  finding  deep  meaning  in  our  natural   surroundings,  humanity  might  start  to  view  them  as  a  constant  potential  threat.   -­‐1   -­‐3   0   -­‐2   30.  Complex  control  systems  never  work  perfectly,  humans  can  make  mistakes  in   design,  manufacturing,  and  operation  (think  of  Chernobyl,  the  Exxon  Valdez,   airplane  crashes  etc);  given  this  fallibility,  it  is  unwise  to  stake  so  much  on  a  more   complicated  arrangement  than  anything  attempted  before.   0   -­‐2   4   3   31.  There  are  such  things  as  morally  bad  research  projects.   3   0   2   3   32.  We  don't  have  the  luxury  of  waiting  around  for  a  change  in  attitude  that  is  never   going  to  happen,  or  until  it  is  far  too  late  for  action.   1   4   -­‐2   2   33.  Government  support  for  geoengineering  research  is  important,  because  good   policy  decisions  depend  on  good  science.   3   2   0   2   34.  A  ban  on  geoengineering  would  be  unenforceable  and  counter-­‐productive  as   those  carrying  out  tests  would  do  so  in  secrecy.   0   1   -­‐3   -­‐2   35.  Geoengineering  technologies  will  not  "solve"  the  climate  change  "problem";   rather,  they  will  redesign  major  Earth  systems  -­‐  including  not  just  natural  but  human   and  built  systems  -­‐  powerfully,  unpredictably  and  potentially  irreversibly.   0   -­‐3   3   0   36.  Substantial  investment  in  geoengineering  research  will  encourage  political   inertia  on  mitigation  and  adaptation,  and  also  facilitate  the  actual  deployment  of   geoengineering  "solutions".     -­‐1   -­‐2   2   -­‐1   37.  In  a  nutshell:  someone,  somewhere  is  going  to  do  geoengineering  research,  so  it   might  as  well  be  someone  responsible.   2   0   -­‐1   -­‐2   38.  Given  that  geoengineering  experiments  are  already  underway  -­‐  uncontrolled,   unmonitored,  illegal  -­‐  then  it  is  time  for  the  world  to  have  a  serious  conversation   about  geoengineering.   2   0   2   0   39.  Even  if  we  stopped  emitting  any  greenhouse  gases  right  now,  we'd  still  have   4   0   -­‐1   0       locked  in  at  least  1.5  degrees  of  warming,  probably  more,  and  there's  no  sign  that   we're  going  to  stop  emitting  any  time  soon.    So  some  kinds  of  geoengineering  might   be  a  necessity,  not  as  a  solution,  but  as  a  postscript  to  a  solution.   40.  Although  there  is  a  lot  of  variety  in  the  geoengineering  approaches  being   proposed,  at  the  end  of  the  day  they  are  technological  fixes  -­‐  and  so  flawed.   -­‐3   -­‐3   1   -­‐4   41.  It  is  likely  that  some  new  technologies  will  be  needed  to  move  to  lower-­‐polluting   energy  systems  and  to  adapt  to  likely  climatic  changes;  but  the  types  of   technologies  pursued,  the  interests  they  favour  and  the  future  societal  vision  they   serve  (or  suppress)  ought  to  be  the  subject  of  public  deliberation  and  ultimately   public  control.     2   -­‐1   4   4   42.  The  governance  challenges  of  controlling  the  global  climate  through   geoengineering  are  more  difficult  to  overcome  than  those  of  transforming  the   global  energy  system.   1   -­‐2   1   0   43.  Since  we  know  that  accumulating  carbon  dioxide  in  the  atmosphere  is  a  problem   requiring  action,  then  we  need  to  focus  our  energies  on  finding  ways  of  taking  it  out   and  storing  it  somewhere  safe  and  permanent,  the  same  way  we  do  with  problems   like  nuclear  waste  and  arsenic  in  water  supplies.   0   3   -­‐1   4   44.  The  anti-­‐geoengineering  people  are  ideologically  motivated  and  rely  on  doom   and  gloom  stories  about  the  environment;  the  last  thing  they  want  to  see  is  that   some  of  these  techniques  might  work.     -­‐1   0   -­‐4   -­‐1   45.  All  of  the  technologies  needed  to  meet  global  2020  emissions  reductions  exist   today.   4   0   3   -­‐3   46.  It’s  not  a  question  of  if  but  when  humanity  will  be  compelled  to  use   geoengineering.   -­‐1   4   -­‐4   0   47.  There  is  an  important  difference  between  geoengineering  field  experiments  that   do  not  cause  damage  to  the  environment,  and  those  that  specifically  aim  to  perturb   natural  systems;  so  a  blanket  ban  on  field  testing  doesn't  make  sense.   2   1   -­‐1   2   48.  Right  now,  the  Arctic  (and  hence  the  Earth)  is  in  a  state  of  dire  emergency,  and   only  immediate  drastic  action,  including  some  forms  of  geoengineering,  can  save  us   from  catastrophe.     -­‐2   4   -­‐3   0       Table 2. Degree to which each participant’s sort correlated with each factor Participants  (by  professional  sector)   Degree  of  correlation  of  Q  sorts  with  each  factor     1   2   3   4   Participants  whose  sorts  correlate  with  just  one  factor:   Media   0.7146*   0.2620   -­‐0.2620   0.3663   Academia  (physical/natural  sciences)   0.5894*   0.2479   -­‐0.2479   0.1008   Academia  (physical/natural  sciences)   0.6133*   0.1971   -­‐0.1971   0.1116   Government   0.8248*   -­‐0.0039   0.0039   0.0494   Academia  (physical/natural  sciences)   0.6957*   0.0514   -­‐0.0514   0.2540   Academia  (physical/natural  sciences)   0.5064*   0.3414   -­‐0.3414   0.0585   NGO   0.5218*   -­‐0.2054   0.2054   0.3222   Academia  (physical/natural  sciences)   0.5637*   -­‐0.2639   0.2639   0.0650   Academia  (physical/natural  sciences)   0.6423*   -­‐0.0186   0.0186   0.3463   Industry   0.6968*   0.2824   -­‐0.2824   0.2885   NGO   -­‐0.6110   0.2209   W 0.3744   0.6110*   NGO   0.1652   0.7993*   -­‐0.7993   0.2589   NGO   -­‐0.1203   -­‐0.7913   0.7913*   0.1451   -­‐0.3013   -­‐0.7216   0.7216*   0.0222   -­‐0.0880   W Academia  (social  sciences/humanities)   W Media   -­‐0.1471   -­‐0.7851   0.7851*     0.2630   -­‐0.6437   0.6437*   0.2930   Academia  (social  sciences/humanities)   -­‐0.0386   -­‐0.8237   0.8237*   -­‐0.0222   Academia  (physical/natural  sciences)   0.3661   -­‐0.2500   0.2500   0.6441*   Industry   -­‐0.0618   0.5440*   Academia  (social  sciences/humanities) (CGG) W 0.1517   0.0618   Industry   0.3012   -­‐0.1675   0.1675   0.5893*   Academia  (social  sciences/humanities)   -­‐0.2322   0.1116   -­‐0.1116   0.7681*   Participants  whose  sorts  correlated  with  more  than  one  factor:   Government   0.6669*   0.4148*   -­‐0.4148   -­‐0.0534   W NGO   W Industry   W Academia  (physical/natural  sciences)   W Industry   W Academia  (social  sciences/humanities)   0.6418*   0.0730   -­‐0.0730   0.4445*   0.405*   0.4904*   -­‐0.4904   0.4047*   0.4325*   -­‐0.0165   0.0165   0.4227*   0.4206*   0.5229*   -­‐0.5229   0.4998*   0.4087*   -­‐0.4449   0.4449*   0.1662   Academia  (physical/natural  sciences)   0.5006*   -­‐0.3921   0.3921*   -­‐0.0582   Academia  (physical/natural  sciences)   0.5248*   0.2192   -­‐0.2192   0.5806*   Industry   0.5671*   0.5335*   -­‐0.5335   0.0511   0.5572*   0.0330   -­‐0.0330   0.5167*   0.6124*   0.4552*   -­‐0.4552   0.3127   0.519*   0.4099*   -­‐0.4099   0.3618     0.0775   -­‐0.5757   0.5757*   0.5035*     0.1935   -­‐0.6088   0.6088*   W Academia  (social   W(CGG sciences/humanities) )   W Industry   Academia  (physical/natural  sciences) (CGG) Academia  (social  sciences/humanities) (CGG) Academia  (social  sciences/humanities)   W(CGG) 0.5097*   w * Indicates that a sort correlates significantly with the factor at the p < 0.01 level; ( ) indicates that an individual carried out a web based Q sort using Q assessor; (CGG) indicates that an individual is associated with the Climate Geoengineering Governance research project, including the author.     Table 3. Correlations between factors.   Factor  1   Factor  2   Factor  3   Factor  4   Factor  1   1         Factor  2   0.4015   1       Factor  3   -­‐0.2080   -­‐0.7225   1     Factor  4   0.3588   0.2007   0.0726   1   Results Narrative descriptions of the four factors that emerged from analysis are given below. The factors were assigned names drawn from statements that were ranked particularly highly for that factor, and act as an abbreviated story line, capturing some essence of the larger narrative. The numbers in square brackets within the text refer to the statement upon which the interpretation is based (see Table 1). It will be noted (in Table 2) that a number of individuals’ Q sorts correlated with more than one factor which suggests that ‘there is not necessarily radical discontinuity across discourses’ (Dryzek & Berejikian 1993), and that many individuals have access to, and may move between discourses or framings (c.f. Collins & Yearley 1992). The relatively high number of these individuals might also be illustrative of the fundamentally ambiguous nature of the term geoengineering, or perhaps suggestive of the fact that around and between the more stable framings described below, the discursive landscape is characterised by a high degree of ambivalence as people struggle to formulate opinions incorporating contradictory normative positions. Although the framings will be described below as discrete in order to facilitate an exploration of some of the different tensions within and between them, it should be remembered that this is not actually the case Factor 1: “At the very least we need more research” Ten participants’ sorts were correlated significantly with this factor, including six individuals from academic (natural/physical science) backgrounds, one journalist, one government employee, one non-governmental organisation professional, and an individual from an industrial background. This framing has been summarised as follows: Action on climate change is clearly urgent [39], but arguments that frame the need for geoengineering in terms of an emergency are unhelpful and counterproductive [48]. Geoengineering is certainly not the most revolutionary new idea in climate policy [14], however we shouldn’t rule any options out, and at the very least we need more research in this area to understand what approaches won’t work and should be avoided at all costs [12]. Research is the only way to determine the potential impacts of different technologies, and we have now achieved the level of scientific sophistication to make research in this     area worthwhile [5, 17]. Research is clearly distinct from deployment, and if carried out in a responsible manner, should not be overly controversial [16]. Indeed, if responsible parties don’t carry out research, it will be done by less responsible parties [37]. Furthermore, the technical community has a responsibility to explore back-up strategies for dealing with possible future climate emergences [19]. Now is the time for a serious societal conversation about if and how we want to develop these different technologies [38], and public involvement in choices about directions of research and development in this area are crucial [13, 41]. Regulation of research is important, but should be undertaken carefully, as there is a risk that hastily developed regulation might be counter-productive and stifle innovation and scientific freedom [4]. Given the variety of different research activities that might take place a moratorium on all activities outside the laboratory doesn’t make sense [25, 47]. Although the deployment of geoengineering is by no means inevitable [46], and we already have all the technology we need to reduce emissions [45], some kinds of geoengineering will probably be a necessary part of any solution [39]. Geoengineering technologies that are likely to be more ethically preferable are ‘encapsulated technologies’ such as air capture, rather than non-encapsulated techniques such as stratospheric aerosols or iron fertilization [28]. Commercial involvement in geoengineering might be helpful [27], but we should probably be wary of claims of technologies to provide win-win solutions allowing economic growth and mitigation to proceed hand in hand [23]. Factor 2: “We are the planetary maintenance engineers” Two participants’ sorts were correlated significantly with this factor, both of whom are associated with non-governmental organisations. The framing has been summarised as follows: We are currently in an unprecedented planetary emergency brought about by climate change [4], immediate action is urgent [32], and it is likely to be only a question of time before humanity is compelled to use geoengineering [46]. Geoengineering is an important part of the solution to climate change [35, 40], hence, research on geoengineering is both crucial and worthwhile [5, 17], and should be supported by governments as the best basis for sound policy making [33]. Humans have demonstrated their ability to build functioning complex control systems [30], and now need to apply that knowledge to the task of planetary maintenance engineering that (like it or not) now falls upon us [6]. Given the dire state of the climate, neither research nor deployment of geoengineering should be overly controversial [16]. Although there might be some risk associated with research, not carrying out research given what we know about climate change would be riskier [3]. Only through research can we learn what technologies might be helpful, and conversely which shouldn’t be deployed [12]. Investment in geoengineering research isn’t likely to have a significantly negative impact on policies towards mitigation and adaptation, especially when one considers the dire state that mitigation policies are in     already [36], and while the governance of geoengineering brings particular challenges, these are likely to be less difficult to overcome than the challenges of transforming the global energy system [42], which so far appears to have failed. Indeed geoengineering has the potential to revolutionise climate policy [14], opening-up possibilities for economic growth and climate change mitigation to proceed hand in hand [23]. We should not be too hasty in pursuing regulation, which might be stifling to innovation and research [4]. A ban on geoengineering would just be counterproductive [25], likely resulting in research being carried out in secrecy or by less responsible parties [34, 47]. Those carrying out research are motivated by a desire to find solutions to the climate change problem, and for developing ‘back-up’ strategies for dealing with a possible future climate emergency [19], rather than any other motivation [8, 7]. Given the urgency of the problem, commercial involvement in geoengineering might be positive in terms of mobilizing innovation and capital investment, possibly increasing the speed with which these technologies could be developed [27]. Factor 3: “Geoengineering is a political strategy” Five participants’ sorts were significantly correlated with this factor, including 3 individuals from academic (social science/ humanities) backgrounds, one journalist and one individual associated with a non-governmental organisation. This framing has been summarised as follows: Geoengineering won’t solve climate change, but is likely to cause unpredictable and irreversible damage to the planet [35]. Attempts to control the climate through geoengineering are neither feasible [30], nor inevitable [46], and would likely lead humanity to a dystopian future in which we would find ourselves trapped by the consequences of our hubristic actions [21]. Geoengineering proposals stem from the same mind set of attempting to control nature that got us into the environmental mess we are in today [1], and are built on the dangerous illusion that complex social problems can be solved with technology [9]. The idea that all of the proposed technologies can be defined as geoengineering because their stated intent is to deal with climate change, is misleading [22]. More than as a set of technologies defined by a stated shared intent, geoengineering can be thought of as a political strategy [24] that serves the interests of the status quo. The commercial interest in some of these technologies only serves to highlight this, and if we were really serious that geoengineering was about ‘saving the planet’ we wouldn’t leave such a task to business [27]. We have all the technologies we need to mitigate carbon emissions effectively; it’s just a question of using them [45]. More research into new technologies isn’t the most crucial thing [12], indeed, the risks of doing research (including the risk of strategic military applications of these technologies [8]) may well outweigh the benefits [3]. It isn’t possible to separate out research from deployment in any straightforward way, and both should be considered controversial [16]: carrying out research, especially field trials, is the first step toward deployment, and drawing distinctions between different types     of field test only serves to obscure this fact [47]. Since full-scale trials are unethical and small-scale trials can’t produce useful data in the noise of global weather [5], it is common sense to institute a moratorium on all testing activities outside the laboratory [25]. The argument that ‘someone somewhere will do it, so it might as well be us’ [37] is not acceptable, nor are arguments stemming from claims of present day [48] or hypothetical future emergencies [12]. The governance challenges of controlling the global climate through geoengineering would likely be more complex and difficult to overcome than those of transforming the global energy system [42], and given the undemocratic and risky nature of proposals for geoengineering, we probably shouldn’t be going down this path [21]. Publics need to be engaged meaningfully in decisions about research [13], and ultimately have control over which (if any) of these technologies are to be pursued [41]. However, much talk of governance seems to see deployment as inevitable, and is hence a purely instrumental exercise for smoothing this process, rather than allowing genuine dissent to emerge [26]. Factor 4: “Let’s focus on carbon” Four participants’ sorts were significantly correlated with this factor, including two individuals from industrial backgrounds, and two individuals from academic backgrounds (one social scientist, one natural scientist). The framing has been summarised as follows: Action on climate change is urgent [32], and is likely to require the development and deployment of new technologies [45], including some that might be labelled as geoengineering [39]. However, the definition of geoengineering is slippery and after realising that our actions en masse affect the climate, anything we do to address it (including nothing) might be considered geoengineering [18]. Although ambiguous, the concept of geoengineering might be useful as a political strategy to help open up the solution space available to us for dealing with climate change [24]. There is nothing wrong with a technological fix per se [40, 20], but it’s important to remember that technology alone will not ‘solve’ the climate change problem [35]. It is important to ensure that the direction of development of these technologies is the subject of public deliberation and control [41], so that, appropriately managed, Geoengineering does not have to be fundamentally undemocratic [2]. We clearly need research into new technologies, if only to be able to rule out those that shouldn’t be deployed [12], but some research is more morally acceptable [31] than others, and the argument that ‘someone will do it so it might as well be someone responsible’ (i.e. us) is problematic [37]. Research cannot be neatly separated from deployment, and thus it is difficult to defend the idea that only deployment should be controversial [16]. For this reason publics should be engaged ‘upstream’ in the direction of research in this area [13]. The inherent complexities of the climate system limit the human ability to predict and judge cause and effects of interventions [17]. This complexity, coupled with human     fallibility, means that attempts to control the climate system are likely to fail [30]. Hence we should focus our energies on removing carbon dioxide from the atmosphere [43] (an endeavour in which commercial involvement might be helpful [27]), so that with the right technological and social changes, a carbon neutral future for humanity can be achieved [11], and mitigation and continuing economic activity can take place [23].       Discussion No claim is made that the four framings of geoengineering that have been described above, constitute any kind of comprehensive, authoritative or final set of framings. As in any study of discourse (whether acknowledged or not), these might rather be thought of as stylised representations that will hopefully be of heuristic utility in offering ‘tools to think with’ in processes of further enquiry (c.f. Brand & J. Fischer 2012). This discussion will draw out some of the tensions between and within the different framings, starting with an examination of the fluidity and ambiguity of the term itself and the definitions of geoengineering given by participants, before examining how concepts of control, research, novelty and interests all feature within and constitute the different framings of geoengineering. Given the existence of campaigns both for8 and against9 geoengineering in general, one might expect to find opinion around geoengineering highly polarised. The emergence of a bipolar factor (split into factors 2 and 3), indicating highly opposed views, is therefore perhaps unsurprising. The prominence of this axis also confirms the general salience for this purpose of an aggregated concept of ‘geoengineering’. However, the existence of a further two factors clearly indicates that the picture is not as simple as a description of a straightforward ‘pro’ /‘anti’ axis might suggest. With regard to the coherence or ambiguity of the term geoengineering, it appears that (although the most different in terms of their framing of geoengineering) individuals loading on factor 2 (broadly in favour of geoengineering), and factor 3 (decidedly against geoengineering), actually appeared to find it less problematic making statements about geoengineering as a non-differentiated category, than those loading on factors 1 and 4. Thus for example, within the factor-3 framing, a total ban on all geoengineering activities outside the laboratory is a necessary and coherent thing to call for [25]. Likewise within the factor-2 framing, geoengineering (in general) is one of the most revolutionary new ideas in climate policy [14]. Interview data collected at the time of the Q sorts can be instructive in understanding this. Thus a participant whose sort correlated with factor 2 explained why he considered the term geoengineering to be useful: ‘The term has proven to be very useful because of the discussions it catalyses. I view the real utility of geoengineering not really as being the technological interventions but as being so extreme as a concept that it actually provokes imagination and the ability to open up discussions that are otherwise mired in more detailed political positions…it opens up new opportunities for reframing how we deal with climate…’ On the other hand the following quote from a participant whose sort correlated with factor 3 illustrates why he feels it is meaningful to object to geoengineering                                                                                                                         8 9    E.g.  AMEG  (the  Arctic  Methane  Emergency  Group)  www.amegme      E.g.  The  HOME  (Hand’s  Off  Mother  Earth)  campaign  http://www.handsoffmotherearth.org/       is possible, and disaggregating the term is not considered to be of primary importance: ‘I just think there's a broader thing afoot about trying to solve problems through technological solutions rather than the heavy lifting of social change and actually addressing the root causes and so forth, and geoengineering in some way I think is totemic for that… it’s not just about a geoengineering solution, it kind of speaks to, we're going to have a technological solution for this, whether its carbon capture and storage or biofuels or air capture or nuclear power, whatever it is, we're going to craft a technological way round this, such that we don't have to significantly disrupt the economy…’ But while this participant appears to be able to object to geoengineering as illustrative of a broader (in his view) problematic attitude towards the application of technology to social and environmental problem solving, , participants loading on factor 1, appeared much less willing to take such a general or abstract stand. A fairly typical quote from a participant loading on factor 1 illustrates this: ‘My reaction to the term is that it isn’t particularly helpful because it describes different groups of approaches. So there’s negative emissions, taking carbon out of the air, carbon dioxide removal I guess it’s called in the academic world, which is a very different set of interventions than the solar radiation management stuff. So geoengineering is a catch-all term that creates challenges for us to then have a clear position on… and within those there’s lots of different approaches which each have their pros and cons and different risk profile. Another commented: ‘I don't want to have a debate about the terminology too much, I think I want to have a debate more about what the technologies do to our climate and I don't think having huge arguments about what a term is or not really make much difference to that’. Interestingly, although the factor-1 perspective appears to find the ‘catch all’ nature of the term presents a problem for the making of general statements about geoengineering, and prefers to focus attention on individual technologies, this view is very clear about what geoengineering is not. Hence it appears from the negative score awarded to statement 18, that the definition of geoengineering (while encompassing a broad range of technologies) does not include those activities the effects of which were inadvertent. A different perspective still was offered by factor 4, whose agreement with statement 18, suggest a distinctive take on the issue of intent, and a broader understanding of what might constitute geoengineering. For example, one factor-4 participant defined the term to include interventions not generally     classed as technological, such as the implementation of a carbon tax (See definition 7 in Table 4 below). While subscribing to a very broad definition of the term itself (to the extent that it might be difficult to differentiate from other categories of effort such as mitigation), factor-4 participants were also conscious of the possible utility of the term itself as offering something distinctive on the discursive level at least. Hence one participant argued against the so-called ‘moral hazard’ argument against geoengineering research [36], by referred to the fact that arguments against geoengineering such as this act to prematurely close down ‘the solution space, the option space that you want to keep open.’ Understanding ‘the strategically and politically loaded negotiation of definition and meaning’ (Walker & Shrove 2007, p.216) is crucial to understanding framing, and definitions themselves can be understood as key elements of wider frames. Table 4. lists a selection of definitions of geoengineering given by participants.10 Table 4. Selection of participants’ definitions of geoengineering Definitions of geoengineering (factor number(s) with which participant’s sort correlated). 1. ‘Deliberate large scale interventions in the earth’s natural systems’ (1/4) 2. ‘I suppose I’ve been influenced quite a bit by the Royal Society report and its very inclusive definition of geoengineering to include any deliberate attempt to change the climate, to fix the climate… Any deliberate attempt to fix the climate apart from cutting our actual emissions’. (1) 3. ‘To me it means intentional technical interference with atmospherics in order to mitigate some or all of the impacts of anthropogenic climate change…my framing would always be to have something to do with the impacts of climate change …I guess I can’t see why [weather modification for other purposes] wouldn’t be included…but all the discussions I’ve had have been with the climate change community and it has been a climate change framed discussion’(1) 4. ‘I guess the scientific definition would be: the modification of climate variables to offset rising greenhouse gas concentrations and associated rise in temperature’(1) 5. ‘Large scale, intentional attempts to change the climate system and other large scale systems’ (3) 6. ‘All sorts of technological ways to combat climate change, so I’d be including removing green house gases, so particularly things like carbon scrubbing that takes out carbon dioxide and buries it underground, and also what they call solar radiation management, actually reflecting the sun. To me it’s all those things’ (2) 7. ‘Geoengineering is the intentional intervention in the global planetary system to affect weather and climate… I am agnostic on what those interventions are, I don’t differentiate geoengineering by methodology, so if you capture carbon from smokestacks in coal plants, or you implement a carbon tax, or you put particulates in the stratosphere, those are all examples of geoengineering…. (4) 8. ‘What it conjures up to me really is talking about the global climate change, and thinking of technological solutions that can either dramatically speed up the rate of mitigation of                                                                                                                         10  Not  all  participants  provided  definitions  (those  carrying  out  online  sorts  for  example  were  not  asked  to   define  geoengineering),  and  there  was  a  degree  of  repetition  among  those  that  did,  hence  the  table  is  not  a   comprehensive  list,  but  offers  a  selection  indicative  of  the  diversity  of  definitions  given.       climate change or do something different, that is prevent some of the worst effects of climate change’ (1) 9. ‘It’s about humans taking interventions to try and control the climate system, which we’ve never deliberately done before even if we are now changing the climate system through our pollution, so it crosses a particular barrier…’ (1/4) 10. ‘I just stick to the Royal Society definition really: deliberate, large scale manipulation of planetary environment in order to counteract anthropogenic climate change’ (3) 11. [Climate remediation] ‘is trying to have available one or more, preferably a set of technologies such that if the planet gets into a really rather unfortunate situation with respect to the people living on it and the environment more generally you don’t have to live with the situation for hundreds and hundreds of years even if you shut down greenhouse gas emissions’ (1/2) 12. ‘When I say geoengineering I mean something like solar radiation management. Geoengineering via carbon dioxide removal is a lot more blurred between geoengineering and mitigation, because is planting the Amazon rainforest and re-growing lots of trees mitigation? Or is it geoengineering because we're going to be locking up a lot of carbon in those trees? That is a very blurry line’ (1) An examination of the definitions reveals potentially significant ‘grey areas’ and ambiguities in the term. For example: is geoengineering defined as being solely about interventions in the climate system (e.g. 2, 3, 4), or could the definition also include any large-scale natural system (e.g. 1, 5)? Is geoengineering defined as being about purely ‘technical’ or ‘technological’ interventions (e.g. 3, 6), or could it also encompass economic or social changes (e.g. 7)? At what scale does an intervention become geoengineering? Is weather modification included (e.g. 7), or is it only about global change (e.g. 3, 8 10)? Is the attempt to counteract climate change fundamental to the definition of geoengineering (e.g. 3, 4, 6, 10), or could an intervention be defined as geoengineering if carried out with other intent (e.g. 1, 5)? Is it primarily about control (e.g. 9)? Should it actually be called something else, such as climate remediation (e.g. 11)? Or should the definition be narrowed down to just solar radiation management and exclude carbon dioxide removal (e.g. 12)? While the term’s ambiguity is therefore clear, it might be said to have ‘functional malleability’ (Gledhill, 1994 p 216). Hence there appears to be a sense (particularly expressed by factor-2 participants) that the term in all its ambiguity might be politically useful in terms of acting as a catalyst for certain kinds of discussions. Ironically the primary discourse of opposition as represented by factor 3, in its view of geoengineering as emblematic of the fundamentally flawed nature of the global neo-liberal political economy, might actually act to breathe life into it. Within the factor-1 framing, the desire to disaggregate the term might be read as an opposition to the constraints of the term, or as an example of boundary work aimed at reducing ambiguity; while the broad definition of geoengineering offered by factor-4 participants could be read as a different type of boundary work actually aimed at increasing the ambiguity of the term.     Scholte et al (2012) argue that ‘ambivalence’ about geoengineering is a frame in and of itself (characterised by the presentation of arguments for and against geoengineering within one text), and they suggest that the increasing prevalence of ‘the ambivalence frame’ above other framings of geoengineering articles in newspapers, provides hope for increasing reflexivity in the debate. We concur that ambivalence is a characteristic of the discourse as a whole, as indicated by the co-existence of multiple divergent normative positions within the debate. However, counter to the argument made by Scholte et al, ambivalence is not here understood as a singular way of framing geoengineering in and of itself. The relatively high number of so-called ‘confounders’ (individuals whose Q sorts correlated with more than one frame), that emerged from the analysis presented here, could be interpreted as revealing different forms of ambivalence with respect to these framings. However, ambivalence with respect to the framings described here need not correspond to an individual being ambivalent about geoengineering per se, although equally, this may be the case in some instances. The existence of these ambivalences might best be interpreted as a degree of instability in the discourse, suggesting that the meanings attributed to geoengineering are still in some senses quite negotiable. Within the four frames uncovered by this study, there emerges clear (nonambivalent) support for, and opposition to geoengineering represented by the polarised factors 2 and 3. In addition, the positions represented by factors 1 and 4, rather than being distinguished by ambivalence, may actually be interpreted instead as clear efforts to formulate non-ambivalent positions with respect to different issues within the geoengineering discourse. For example, this may be by discriminating in more detail between particular technologies (to allow non-ambivalence on each). Or it may be through seeking to elucidate a discriminating normative position in relation to specific issues such as the controllability/knowability of the climate, the neutrality or otherwise of research, and the role of technology in society, each of which also has the effect of reducing ambivalence in particular instances. The existence of the framing exemplified by factor 4, also problematises the frequent calls for increasing precision around the term geoengineering as a prerequisite for effective governance discussion. This is because it highlights how no one framework for partitioning of the term geoengineering into sub-categories can in itself be thought of as final. The commonly used CDR/SRM distinction, for instance, or even to the level of ‘individual’ technologies, only makes sense from within particular frames. Alternative partitionings of the overarching field, like those defined under factor 4, may cross-cut such a taxonomy – and each other. So any one form of precision may reduce particular ambiguities, but leave others unaddressed – or even compound them. Calls for greater precision must therefore be interrogated as to particular axes of precision involved and their implications. And it cannot be assumed that precision in and of itself will remedy ambiguity. Axes of difference     Hulme (2008) argues that the prospective routes held out to us for dealing with climate change all have ‘connotations of global control and mastery of the climatic future’ (p. 12), and this observation is borne out of the factors that emerged from this study, in which various ideas around the issue of control constitute one of the axes of difference between the framings. Unsurprisingly perhaps, the starkest contrast is between the polarised views of factor 2 and 3, although interestingly notions of control are arguably central to both, with the former affirming the notion that ‘we can and should control the climate’, and the latter affirming the notion that ‘we can and should control the research.’ Hence the factor-2 framing of geoengineering as ‘planetary maintenance engineering’ [6], and its emphasis on the human ability to create complex control systems [30], building on an ever increasing scientific sophistication allowing greater understanding of complexity [17]. Conversely the factor-3 framing emphasizes what is perceived to be the folly of attempts to control nature [1, 15], the irreducible complexity and chaos of the climate system [17], and human fallibility in attempting to create complex control systems in the past [30]. The roles are reversed when the focus becomes control of research and other geoengineering activities, with the factor-2 framing emphasizing what is felt to be the ‘counterproductive’ nature of attempts to ban on geoengineering which would result in testing carried out in secrecy [34], and the inappropriateness of a moratorium on geoengineering [25], while the factor-3 framing emphasizes the perceived necessity of strong controls on research. With regard to the degree to which this control of research is believed to be possible, one participant commented: ‘It may be true that it’s not fully enforceable but it has very powerful norm setting… it’s very important to set that as the standard.’ While rejecting the ‘planetary maintenance’ metaphor, the factor-1 and 4 framings diverge somewhat in their view of the feasibility of achieving climate control, with more reticence being expressed within the factor-4 framing, about the possibilities for either knowledge of complexity [17], and control of the climate [30]. The concept of research represents another fault line or tension between the framings. Given that much discussion of geoengineering occurs in academic journals, and much of the discussion is about research of various types and disciplines, Geoengineering is thus framed by many as being ‘at the research stage’, and in particular factor 1 participants appeared to adhere to this view. Thus for example, a fairly typical factor-1 viewpoint was expressed thus: ‘I would be paranoid and scared of anybody saying we're going to start geoengineering tomorrow, but I'd be just as worried about someone saying we're going to outlaw any research on geoengineering. We need to do this research. Whether or not we actually do anything with the research is another matter. But in case we need to geoengineer, we     should do the research now. Because when you're doing it in a panic and you think you've only got 20 years before London is underwater, you're not going to do science as well as when you think we might not need to do this, you can actually sit back and concentrate and take a slightly longer term view of it’. The emphasis on research is broadly shared by factors 1, 2 and 4, as illustrated by factor scores for statements 3, 12, and 33, but is problematized by factor 3 in particular, and to a lesser extent factor 4 (statements 16 and 37). Thus a factor 3 participant commented about research: ‘[research] creates a dynamic where you're moving towards something, it creates the beginnings of almost an industry of people who have invested in all of that.’ And the same participant was keen to unpick the term, asking ‘what’s hidden in the term research’? and commenting: ‘I think it’s a very deliberate, the term [research] gets kept together, and by being kept together it means that people who actually don't ever want to move to some kind of experimentation should nonetheless feel they have to support that statement [3] … it speaks to scientific freedom and all these kind of things, which of course and if you’re in academia are deeply important and rightly so, but I think it’s a deliberate strategy to keep that language obscure’. The way in which geoengineering is framed (particularly but not exclusively by factor 1 participants) as being at the research stage, also feeds into ideas about the degree to which geoengineering represents something fundamentally new and untried, or is a continuation/the latest manifestation of practices and ideas with a long history. Of the four framings uncovered by this study, the emphasis on continuity is most apparent within the factor-3 and 4 framings, while as outlined above, 1 and 2 appear to emphasize research and novelty. For example, factor 3 was distinctive among the four factors in awarding neither a positive nor negative ranking to statement 7 (that deliberate geoengineering has been happening for decades and was not all about dealing with climate change). Interview data and comments from participants who loaded on factor 3 point to a division over exactly what this means. One view was characterised by the following comments: ‘Geoengineering technologies patented decades ago have been and are being used covertly as political/economic/military weapons. This is obvious to anyone who studies the sky and knows the history of weather/climate modification development. Look up!’     Although not all factor-3 participants shared this view, the issue of continuity with other technologies and the idea that the issue was broader than the current climate change focus might suggest were shared. Hence another factor 3 participant commented: ‘I do think there’s other interests in geoengineering other than climate change, particularly commercial interests. I think there’s an attempt to create new markets in the longer term, there is military interest… I don’t think they’ve been spraying but I do think it’s not all about climate change. There’s more reasons to want to have geoengineering as a tool in the box than climate change’. The distinction between geoengineering, weather modification and so-called ‘chemtrails’ theories is worthy of a brief note at this point. As a subject discussed and debated by governments, think-tanks and academics, geoengineering in all its ambiguity appears to have acquired widespread credibility as a ‘serious’ (Keith & Dowlatabadi 1992) scientific subject. Weather modification on the other hand, has a relatively less authoritative status, occasionally presented as a ‘pseudo-science’, associated with an array of more or less credible characters driven by a variety of more or less honourable intentions (Fleming 2006). While finally the ‘chemtrail’ theory (positing the existence of a global network of weather modification for nefarious ends), lacks credibility and authority and is widely labelled (dismissively) as a conspiracy theory. However, as this examination of the multiple framings of geoengineering has revealed, the boundaries around terms and activities are by no means clear cut or un-ambiguous, and on-going boundary work (c.f. Gieryn 1983) is required to maintain the distinction between terms in such a way as to maintain the epistemic authority of certain actors. Interview comments from factor-4 participants reveal a view that is more explicit about the fluid and blurred nature of the boundaries between different activities: ‘We’re already geoengineering the climate … I mean we do a lot to try and change the climate system, we dam rivers, we irrigate large parts of farmland that changes the local climate, we deforest. In the western US I think there’s 169 weather modification project that try to improve rainfall, China does it systematically…’ And when questioned further about the distinctiveness of weather modification from geoengineering, the same participant highlighted the continuum between them and the constructed character of notions of ‘climate’, by commenting: ‘Weather is events and climate is statistics…’ The distinction between weather modification and geoengineering is also brought into question by the following comment made by a factor-1 participant:     ‘[The Chinese] are raising from 70 million to 500 million a year the amount they’re spending on their weather modification program, and once you get to half a billion dollars a year, you’re actually talking about something that on aggregate could have a significant effect.. Assume it’s effective , you’re now at a stage where you’re modifying local weather sufficiently over a long enough period that it’s kind of like a geoengineering intervention […] I think we’re going to back-step into geoengineering in that way.’ The emphasis on novelty or continuity in different framings of geoengineering is intimately connected to different framings of the interests and motivations at play, and this is another axis of difference along which the framings uncovered by this study can be seen to diverge. Again, factors 3 and 4 are united by a shared framing of the potential for non-climate change related application of geoengineering technologies, including military applications. But beyond more radical uses of geoengineering technologies for purposes other than combating climate change, a key distinguishing characteristic of the factor3 perspective is that geoengineering – both the technologies it comprises and the attitude it is understood to represent – is an explicitly political project. The framing of the issue is understood to be a key component of that project. As a participant explained: ‘On the pro-geoengineering side I think there is a small core of ideologically motivated and politically smart and active people who are moving people intentionally, particularly in the whole framing game in very careful ways … while there is a lot of naivety and good intention throughout the discussion there’s also some very active interests… I can see some evidence of that. That sounds conspiratorial, it’s not … it’s just looking at the political economy of discussions around climate change’. Finally, various authors have noted the use of a real or hypothetical climate emergency as a powerful framing device within which geoengineering interventions are situated, and similarly the existence of a climate emergency was an important element of one of the framings (factor 2) that emerged from this study. The following comment typifies this element of the framing: ‘The risks from the climate are infinitely worse than the risks from geoengineering, I mean that’s absolutely obvious. I say infinitely because that means the end of everything, end of civilisation possibly all human life, I mean it’s as serious as that […] Long term it’s a catastrophe. However, although interviews reveal that the urgency of the climate predicament is clearly important for many people, it appears that the framing of the issue in terms of emergency is being consciously rejected by all but factor-2 participants [statement 48]. For example, one factor-1 participant commented: ‘I think the     whole idea of a climate emergency is really kind of counter-productive.’ Another hinted at a more strategic view of framing by commenting that it was not a question of whether emergency was a reality or not, but whether or not the emergency frame was helpful for achieving particular ends: ‘I think people are consciously stepping away from [the emergency framing] because it’s become clear that different ideas about what emergency means makes the term useless… It’s difficult to use emergency to promote particular actions.’ However, although emergency was rejected as a valid framing of the issue by participants that loaded on factors 1 and 4, the idea of a hypothetical future emergency still featured within these views as a rationale for research [statement 19]. Participants loading on factor 3 rejected any emergency rationale (either present or future) for geoengineering. One participant explained why he considered the climate emergency framing to be problematic: ‘The dangerous things to do with geoengineering, is to frame it … only to be a climate discussion, because if you do then it becomes this unidimensional, you know, climate change has got terribly bad, we need to have a fix for it, everything gets arbitrated within this very narrow climate thing, but what your changing is the planet, or you know, large parts of it, which are much more than about climate, climate is just one factor.’ Significant silences It is worth noting that a number of people involved with critical environmental NGO’s, who were invited to take part in this study did not respond to invitations to participate, and hence it is likely that there may be a number of significant silences or gaps in the research presented. The reasons for individuals’ reticence about involvement (whether about the subject matter, this particular study, or the Geoengineering Governance Project more broadly) were not specified and thus can only be the subject of conjecture. However, Walker and Shrove point out that involvement of a broad range of stakeholders in participatory projects and processes, can raise a number of issues, with the potential for inclusion to be ‘re-interpreted as a process of co-option and neutering of dissent, producing deeply problematic tensions for those taking part’ (Walker & Shrove 2007, p.221). Indeed the issue of co-option was one that was raised explicitly by a factor 3 participant, who argued that much of the discussion around geoengineering was being manipulated by people interested in slowing down and confusing governance of climate change; and that hence even being drawn into these discussions would be to play into the hands of these interests. He commented: ‘There's a lot of well-intentioned people, who are caught up in the discussion and I think to some extent are being used, and some of them are letting themselves be used...’     If then, one views the conversation itself as a massive distraction from existing governance discussions around climate change, then perhaps silence and nonparticipation in that conversation, as embodied by refusals to participate in just such processes and projects as this one, can be read as an effective form of dissent (c.f. J. Whelan & Lyons 2005). Conclusion Fischer and Hajer (1999, p.2) argued that although conceptually weak, the term ‘sustainable development’ created a generative metaphor or story-line around which different interests could converge, and thus proved to be a very functional concept. Arguably the same might be said of the term ‘geoengineering’ (on a smaller, subordinate and more idiosyncratic canvas). As this study has illustrated, geoengineering has a fluid and ambiguous set of meanings and is framed by different actors in a number of ways. Interestingly (and unlike ‘sustainability’), the convening power of the term seems equally potent in two opposing directions. This evident polarity within the debate as revealed by the existence of factors 2 and 3, appears to indicate a ‘framing gulf’ across which actors are likely to ‘talk past one another’ rather than engage meaningfully (c.f. Hoffman 2011). However, it is also the case that the existence of additional framings not defined purely along this axis of difference suggests an emerging resistance among certain actors to the debate becoming polarised in this way. These alternative framings appear to be seeking either to increase (in the case of factor 4) or decrease (in the case of factor 1) the ambiguity of the term, but given the multiple framings and meanings within the term, the latter is unlikely ever to be fully realizable. Unlike the picture suggested by Scholte et al. (2013), who suggest that what they call ‘the ambivalence frame may prove to be less powerful than other frames that evoke strong positive or negative feelings’, our findings suggest that ambivalence is not best seen as a frame in and of itself able to garner or lose support. Rather, along with ambiguity, it is a more pervasive and fundamental feature of the discursive landscape of geoengineering. Interviews have highlighted the diversity of actors broadly ascribing to shared framings of geoengineering, which might suggest the coming into existence of various discourse coalitions (F. Fischer & Forester 1993) around the term, linking otherwise disparate actors and networks through certain shared narratives and the utilisation of certain discursive resources (e.g. the narrative of the neutrality/ normative desirability of ‘research’ linking disparate groups within factor 1, or the narrative of the essentially undemocratic nature of engineering at the planetary scale, linking groups within factor 3). And that the ambiguous nature and interpretative flexibility of the term facilitates this process by allowing individuals with perhaps little in common, to speak the same language or     advance shared interests. Clearly there is also a danger here of co-option of certain actors by others utilizing particular framing devices strategically to garner support for a particular view. Finally there may be significant silences in the picture presented, and evidence of a desire for non-engagement in debates around geoengineering, which might have implications for future work on public engagement. References Addams, H. & Proops, J., 2000. Social discourse and environmental policy: an application of Q methodology, Cheltenham: Edward Elgar. Baker, S. et al., 1997. The Politics of Sustainable Development, London: Routledge. Barry, J. & Proops, J., 1999. Seeking sustainability discourses with Q methodology. Ecological Economics, 28(3), pp.337–345. Bellamy, R. et al., 2012. Appraising Geoengineering. Tyndall Centre for Climate Change Research, Working Paper, (June). Benford, R.D. & Snow, D.A., 2000. Framing processes and social movements: An overview and assessment. Annual Review of Sociology, 26, pp.611 – 639. Brand, R. & Fischer, J., 2012. Overcoming the technophilia/technophobia split in environmental discourse. Environmental Politics, (November 2012), pp.1– 20. Brown, S.R., 1970. On the use of variance designs in Q methodology. Psychological Record, 20, pp.179–189. Brown, S.R., 1980. Political Subjectivity: Applications of Q methodology in political science, New Haven and London: Yale University Press. Brown, S.R., 1986. Q technique and method: Principles and procedures. In W. D. Berry & M. S. Lewis-Beck, eds. New tools for social scientists: Advances and applications in research methods. Beverly Hills: Sage, pp. 57–76. Brown, S.R., 1971. The forced-free distinction in Q technique. Journal of Educational Measurement, 8(4), pp.283–287. Buck, H.J., 2013. Climate engineering: spectacle, tragedy, or solution? A content analysis of news media framing. In C. Methmann, D. Rothe, & B Stephen, eds. (De-)Constructing the Greenhouse: interpretative approaches to global climate governance. London: Routledge.     Burt, C., 1972. The reciprocity principle. In S. R. Brown & D. J. Brenner, eds. Science, psychology and communication. New York: Teachers College Press. Collins, H.M. & Yearley, S., 1992. Epistemological Chicken. In A. Pickering, ed. Science as practice and culture. Chicago, USA: University of Chicago Press, pp. 301–326. Dayton, B.W., 2000. Policy Frames, Policy Making and the Global Climate Change Discourse. In H. Addams & J. Proops, eds. Social discourse and environmental policy: an application of Q methodology. Cheltenham, UK: Edward Elgar Publishing, Inc. Dryzek, J.S. & Berejikian, J., 1993. Reconstructive Democratic Theory. The American Political Science Review, 87(1), pp.48–60. Eden, S., Donaldson, A. & Walker, G., 2005. Structuring subjectivities? Using Q methodology in human geography. Area, 37(4), pp.413–422. Edenhofer, O. et al., 2012. Meeting Report of the Intergovernmental Panel on Climate Change Expert Meeting on Geoengineering, Lima. Entman, R., 2004. Projections of power: Framing news, public opinion, and US foreign policy, Chicago: University of Chicago Press. Fischer, F. & Forester, J., 1993. The argumentative turn in policy analysis and planning, Routledge. Fleming, J.R., 2006. The pathological history of weather and climate modification : Three cycles of promise and hype. Historical Studies in the Physical and Biological Sciences, 37(1). GAO, 2010. Climate Change: A Coordinated Strategy Could Focus Federal Geoengineering Research and Inform Governance Efforts. Report to the Chairman, Committee on Science and Technology, House of Representatives, Washington D.C. Gardiner, S.M., 2011. Some Early Ethics of Geoengineering the Climate : A Commentary on the Values of the Royal Society Report. Environmental Values, 20, pp.163–188. Gieryn, T.F., 1983. Boundary-Work and the Demarcation of Science from NonScience : Strains and Interests in Professional Ideologies of Scientists. American Sociological Review, 48(6), pp.781–795. Goffman, E., 1974. Frame analysis: An essay on the organization of experience, Harper & Row. Gruber, J., 2011. Perspectives of effective and sustainable community-based natural resource management: an application of Q methodology to forest projects. Conservation and Society, 9(2), pp.159–171.     Hajer, M. & Fischer, F., 1999. Living with nature: Environmental politics as cultural discourse. In F. Fischer & M. A. Hajer, eds. Living with nature: environmental politics as cultural discourse. New York: Oxford University Press, pp. 1–20. Hajer, M.A., 1997. The Politics of Environmental Discourse: Ecological Modernization and the Policy Process, Oxford: Oxford University Press. Hoffman, A.J., 2011. Talking Past Each Other? Cultural Framing of Skeptical and Convinced Logics in the Climate Change Debate. Organization & Environment, 24(1), pp.3–33. Hogan, A., 2010. A comparison of offline and online Q factors. Paper read at the International Society for the Scientific Study of Subjectivity (ISSSS) 26th Annual Q Conference, Akron, OH, USA, October 7-9, pp.1–6. Hoppe, R., 1999. Argumentative turn. Policy analysis, science and politics: from “speaking truth to power” to “making sense together”. Science and Public Policy, 26(3), pp.201–210. Hulme, M., 2008. The conquering of climate: discourses of fear and their dissolution. The Geographical Journal, 174(1), pp.5–16. Huttunen, S. & Hilden, Mikael, 2012. Framing geoengineering in academic literature (DRAFT). Keith, D. & Dowlatabadi, H., 1992. A serious look at geoengineering. Eos, Transactions American Geophysical Union, 73(27), pp.292 –293. Long, J. et al., 2011. Geoengineering: A national strategic plan for research on the potential effectiveness, feasibility, and consequences of climate remediation technologies, Washington D.C. Luokkanen, M., Huttunen, S., Hildén, M., et al., 2013. Geoengineering, news media and metaphors: Framing the controversial. Public Understanding of Science. Luokkanen, M., Huttunen, S. & Hildén, M., 2013. Geoengineering, news media and metaphors – framing the controversial. Public Understanding of Science. Nerlich, B. & Jaspal, R., 2012. Metaphors we die by? Geoengineering, metaphors, and the argument from catastrophe. Metaphor and Symbol, 27(2), pp.131–147. Ockwell, D.G., 2008. “Opening up” policy to reflexive appraisal: a role for Q Methodology? A case study of fire management in Cape York, Australia. Policy Science, 41(4), pp.263–292. Porter, K. & Hulme, M., 2013. The emergence of the geoengineering debate in the UK print media: a frame analysis. The Geographical Journal, online ver.     Reber, B.H., Kaufman, S.E. & Cropp, F., 2000. Assessing Q-Assessor: A Validation Study of Computer-Based Q Sorts versus Paper Sorts. Operant Subjectivity, 23(4), pp.192 – 209. Robbins, P. & Krueger, R., 2000. Beyond bias? The promise and limits of Q method in human geography. The Professional Geographer, 52(4), pp.636– 648. Rydin, Y., 1999. Can we talk ourselves into sustainability? The role of discourse in the environmental policy process. Environmental Values, 8, pp.467–484. Sarewitz, D., 2011. The voice of science: let’s agree to disagree. Nature, 478(7), p.2011. Schmolck, P., 2002. PQMethod 2.11. Downloaded from http://www.lrzmuenchen.de/~schmolck/qmethod/downpqx.htm. Scholte, S., Vasileiadou, E. & Petersen, A., 2012. Opening Up the Societal Debate on Climate Engineering: How Newspaper Frames are Changing. SSRN Electronic Journal, pp.1–28. Scholte, S., Vasileiadou, E. & Petersen, A.C., 2013. Opening up the societal debate on climate engineering: how newspaper frames are changing. Journal of Integrative Environmental Sciences, (February 2013), pp.1–16. Schön, D.A. & Rein, M., 1995. Frame reflection: Toward the resolution of intractable policy controversies, Basic Books. Shepherd, J. et al., 2009. Geoengineering the climate: science, governance and uncertainty, London: The Royal Society. Sikka, T., 2012. A critical discourse analysis of geoengineering advocacy. Critical Discourse Studies, 9(2), pp.163–175. Stephenson, W., 1935. Correlating persons instead of tests. Character and Personality, 4(1), pp.17–24. Stirling, A., 2007. “Opening Up” and “Closing Down”: Power, Participation, and Pluralism in the Social Appraisal of Technology. Science, Technology & Human Values, 33(2), pp.262–294. Swedeen, P., 2006. Post-normal science in practice: A Q study of the potential for sustainable forestry in Washington State, USA. Ecological Economics, 57(2), pp.190–208. Van Tubergen, G.N. & Olins, R.A., 1979. Mail vs personal interview administration for Q sorts: a comparative study. Operant Subjectivity, 2(2), pp.51 – 59.     Walker, G. & Shrove, E., 2007. Ambivalence, sustainability and the governance of sociotechnical transitions. Journal of Environmental Policy and Planning, 9(3-4), pp.213 – 225. Watts, S. & Stenner, P., 2005. Doing Q methodology: theory, method and interpretation. Qualitative Research in Psychology, 2(1), pp.67–91. Webler, T. et al., 2009. Using Q Method to Reveal Social Perspectives in Environmental Research. Whelan, J. & Lyons, K., 2005. Community engagement or community action: Choosing not to play the game. Environmental Politics, 14(5), pp.596–610.     Appendix: Participant list Name Adrian Tuck Andrew Lockley Andy Boston Angus Ferraro Ben Parker Catherine Scott Cathy Johnson Clare Heyward (CGG) Hayley Stevenson Hiroshi Mizutani Holly Buck James Greyson Professional self-description and discipline (where applicable) Professor, physical chemistry and meteorology Internet Marketing Consultant/ moderator of Geoengineering Google group Technical Head, Business Modelling PhD researcher, meteorology and climate science Post-doctoral research fellow, climate modelling PhD researcher, atmospheric sciences Climate science researcher Research Fellow, political philosophy Lecturer, politics Professor of biogeochemistry and sociogeochemistry PhD researcher, development sociology Consultant Jim Thomas Research Fellow/ Deputy Director, Centre for Engineering Policy, physics/policy Research programme manager John Nissen Chairman Jon Taylor Jolene Cook Climate change programme manager Researcher, climate science Masahiro Sugiyama Researcher Matt Woodhouse Research Fellow, atmospheric chemistry Freelance journalist Jason Blackstock (CGG) Max Mogren Institution/organisation (where applicable) Imperial College London, UK Business Angel consulting, UK E.ON, UK Reading University, UK University of Leeds, UK University of Leeds, UK Department for Energy and Climate Change, UK Oxford University, UK Sheffield University, UK Nihon University, Japan Cornell University, US Blindspot thinktank & Carbon Gold, UK University College London, UK Etc Group, Canada Arctic Methane Emergency Group, UK WWF, UK Department for Energy and Climate Change, UK Central Research Institute of Electric Power Industry, Japan University of Leeds, UK NA Friends of the Earth, UK Paul Appleby Head of Policy, Research and Science Head of Energy Economics Peter Davidson Consultant chemical engineer Davidson technologies, UK Peter Healey (CGG) Research Fellow, sociology Oxford University, UK Phil Renforth Oxford University, UK Piers Forster Research Fellow, civil engineering/geochemistry Professor, atmospheric sciences Richard Mountford Founder/ director 2percent for the planet, UK Mike Childs     BP, UK University of Leeds, UK Roger Pielke Professor, environmental studies University of Colorado, US Ronal Larson Consultant/ coordinator of Yahoo discussion group on biochar PhD researcher, atmospheric physics Science journalist Larson Consulting , US Programme manager /James Martin Fellow, Oxford Geoengineering Programme Lecturer, communication Oxford University, UK Research Fellow, geoengineering governance Oxford University, UK Simon Driscoll Stephen Battersby Tim Kruger Tina Sikka Nils Markusson (CGG) Oxford University, UK New Scientist, UK Simon Fraser University, Canada (CGG) indicates a participant associated with the Climate Change Governance project.