Gene Drives as Interventions into Nature: the Coproduction of Ontology and Morality in the Gene Drive Debate

Gene drives are potentially ontologically and morally disruptive technologies. The potential to shape evolutionary processes and to eradicate (e.g. malaria-transmitting or invasive) populations raises ontological questions about evolution, nature, and wilderness. The transformative promises and perils of gene drives also raise pressing ethical and political concerns. The aim of this article is to arrive at a better understanding of the gene drive debate by analysing how ontological and moral assumptions are coproduced in this debate. Combining philosophical analysis with a critical reading of the gene drive literature and an ethnographic study of two leading research groups, the article explores the hypothesis that the development of and debate about gene drives are characterized by a particular intervention-oriented mode of coproduction. Based on the results of this exploration, we highlight the need for a broadening of the perspective on gene drives in which empirical, moral, and ontological concerns are addressed explicitly in their interplay rather than in (disciplinary) isolation from each other.


Introduction
Gene drives are inheritance-biasing biotechnologies, considered promissory for their potential to revolutionize approaches to disease eradication and prevention, protection of wild populations, species, ecosystems and biodiversity, and protection of agricultural crops [1].Projected applications are based on the gene drive construct's ability to copy itself such that it allows spreading genetic alterations through wild populations [2].As for many related biotechnological fields, the advent of CRISPR as a genome editing technique has led to recent claims of revolutionizing development of synthetic (human-made) gene drives 1 .After a first article outlining the technological specifics and potential fields of application of CRISPR-based gene drives in 2014 [3], proof-ofconcept under laboratory conditions quickly followed for yeast [4], fruit flies [5], mosquitos [6][7][8][9], bacteria [10], and mice [11]  2 .
The revolutionary promises and core features of gene drive technologies give rise to both moral and ontological (or metaphysical 3 ) issues.Starting with the former, a noteworthy feature of the gene drive debate is how explicitly morally charged it has been from the outset (since 2014).The commonly highlighted potential benefits in health, conservation and agriculture-contributing to the values of health through disease eradication, biodiversity through population and ecosystem management and food safety and security through pest control-are often presented as not only supporting the use of gene drives but as making their further development and testing close to morally obligatory [12][13][14][15][16][17].At the same time, gene drives raise strong moral concerns.The technology's potential for triggering rapid, possibly irreversible spread of genetic modifications to target populations in the wild [1] led to concerns about security and unintended consequences that have become widely discussed in the literature [1,15,[18][19][20][21][22][23][24][25][26].Recognition of these concerns triggered policy, regulation, and governance proposals from an unusually wide range of contributors, who at times also collaborated in preparing these proposals: policy and regulation makers, (popular science) media, academics from the social sciences and humanities (particularly those interested in governance and policymaking and those interested in ethics), and-of particular noteworthiness-individual scientists working on the technology [1, 14, 15, 21-23, 25, 27-42].The main risks initially focussed on were the accidental, premature or ill-intentioned escape or release of gene drive-modified organisms with negative and potentially irreversible ecological consequences [4,27].Subsequently, more attention has been devoted to the potential for negative impacts from gene drive releases conducted intentionally and with the best of intentions ( [43][44][45] provide overviews of such considerations; publications about safely trailing gene drives can be considered part of this category as well, e.g.[22,33,[46][47][48][49][50]).
On the ontological side, gene drives can be regarded as transforming our conceptions of nature and natural phenomena, as a select number of contributions to the gene drive debate have suggested [1,19,[43][44][45].While all biotechnologies intervene into nature in one way or another, gene drives have been positioned as a way of "sculpting evolution" [51,52] due to their inheritance-biasing features.This has led to the suggestion that gene drives transform our understanding of the givenness of this fundamental mechanism of the living world [1,2,19,24,32,44,53].The potential ontological ramifications of gene drive applications also involve understandings of nature as wilderness.While gene drives are often referenced as a way of targeting wild populations, these projected applications simultaneously seem to challenge the very idea of wildness or wilderness as characterized by its relative distance from human intervention [1,3,18,19,24,43,53,54]  4 .Gene drive technologies are thus potentially ontologically disruptive in the sense that they challenge the boundaries and sometimes even the coherence of core concepts for navigating the biological world from "evolution" to "nature" to "wild(er)ness".
The aim of this article is to arrive at a deeper understanding of gene drives by analysing these assumptions in tandem; that is, by analysing the specific ways in which ontological and moral assumptions-about gene drives, nature and human action-are coproduced [55] in the gene drive debate, and by critically 2 Although the latter publication indicated more mixed results; the encountered technical difficulties were grounds for the authors to conclude that 'both the optimism and concerns are likely to be premature' [11, p. 108]. 3Philosophically, the terms "ontology" and "metaphysics" can be differently defined and distinguished.We use the following distinction: whereas metaphysics involves accounts of the fundamental nature of being, reality or nature, ontology tends to be more specifically concerned with providing accounts and categorizations of the entities, processes, and relations of which (certain domains of) reality consist.In this article, for purposes of consistency, we have chosen to adopt the term ontology instead of metaphysics to refer to those assumptions that concern reality and/or nature and the different entities or other phenomena it consists of.We consider metaphysics to be the more appropriate term for referring to (culturally specific) systems of understanding and giving meaning to the world, nature, and the place of humans in it. 4Within the environmental philosophy literature the meaning of wilderness and wildness are contested.While some regard wildness as the characteristic of a natural entity that has never been touched by human hand [152,153], others argue that this is premised on a false dichotomy between human and nature [154] and rather focus on autonomous processes.Vogel [155, p.112] for example interprets wildness as the 'operation of forces in an object or organism that operate unpredictably and beyond the grasp of any human actor'.Others [156,157] distinguish different dimensions of wildness and present it as a more gradual concept.We elaborate this framework of coproduction and intervention in the "Coproduction of Morality and Ontology" section of the paper.The section "The Gene Drive Debate" first explains the literature study and ethnographic material and method that support the article's conceptual-philosophical narrative, and provides a brief account of the gene drive debate.In "Gene Drives as Interventions" we present the analysis of gene drives as interventions and the ontological and moral assumptions this involves.We subsequently discuss the implications of adopting a framework that takes coproduction seriously for our understanding of gene drives and intervention in "A Perspective that Acknowledges Coproduction-Discussion".We conclude the article in the "Conclusion" with a brief summary and reflection on the question of the relevance of our argument for the political and ethical dialogue on the use of gene drives.

Literature Analysis and Ethnography
The following account approaches gene drives by integrating an analysis of the gene drive literature and an ethnographic study of two research groups with a philosophical analysis of coproduction of ontology and morality.We explore a (hypo)thesis about intervention-oriented coproduction through the observational detail of ethnography and framing analysis.The findings from our reading of the literature provide a more general framework for the conceptual analysis, with observations from the ethnography providing more detailed illustrations around particular themes.The literature analysis aims at understanding the gene drive debate in terms of its discourse on the implications of gene drives and how to deal with these.The analysis focused on publications in academic journals about gene drives since 2014-which is when the article by Esvelt et al. [3] was published which kickstarted the recent attention the technology has received; when referring to "the gene drive debate", we thus refer to publications from 2014 onwards-that addressed the wider (e.g.ethical, safety, responsibility) implications of gene drives.Also included was "grey" literature on gene drives from policy advisory organizations."Hard science" publications that focus only on the technical side of developing gene drives were excluded 6 .The aim of the literature analysis was not a systematic, categorizing, descriptive overview, but to arrive at a critical reading of the discourse on gene drives on the basis of noticeable ways of framing the moral and ontological implications of gene drives.
While we do not focus on the technical, state-ofthe-art side of gene drive development (for recent overviews that do, see [57,58] 7 ), we want to point out that recent reviews of the science behind gene drives [57,58] convey two core points: first, that it has proven difficult to design gene drives that would actually be effective and safe in the wild, due particularly to the development of resistance to the drive and reduced fitness as a result of the drive, which hamper the evolutionary stability of the drive construct in the wild; and secondly, that there has been remarkable promise and success in addressing these issues. 8 The ethnographic study focused on two case studies of research groups involved in gene drive development (the second group worked on a particular type of drive, a split drive, that is not formally a full gene drive; when we refer in this article to gene drive technology, we refer to a collection of variants that include full gene drives and split drives-more on the variants in The Current State of the Gene Drive Debate).The studies were conducted by the first author between the end of 2018 and the middle of 2020.The first case study involved a threemonth engagement with MIT's Sculpting Evolution lab in Cambridge, Massachusetts.Data was gathered by attending group meetings, observing lab-work, holding one-on-one conversations with group members and a number of social scientists and humanities scholars who had recently worked with and/or researched the group, studying the group's website, and analysing the literature which has already been published about the group.The lab was selected for their prominence in the gene drive debate.The group's director, Kevin Esvelt, has been a highly visible figure in the gene drive debate, with many academic and popular (science) publications referencing him, profiling him, or being (partly) by his own hand; the PhD researcher from the other research group (introduced below) even suggested he had single-handedly set the course for both the gene drive debate and the technology's research agenda since 2015.Particularly striking has been his vocality concerning the non-scientific side of gene drives.This is unusual for a scientist/biotechnologist, at least to this degree, and as such should not be taken as typical for today's technoscience.The group's projects (at or just prior to the time of visiting) involving the development of or the consideration of using a gene drive to address a societal problem include 9 : studying the evolutionary stability of gene drives using nematodes (roundworms) as a model organism; inventorying and improving upon existing mathematical modelling of gene drive dynamics "in the wild"; (exploring the possibilities for) designing pain-free mice; designing mice for field-release to combat tick-borne diseases on islands in the vicinity of MIT; (exploring the possibilities of) combatting invasive species for conservation purposes, with Australia and New-Zealand as prime locations 10 .Other projects included developing non-gene-drive mammalian population control systems for conservation; developing viral defense models; conducting "directed evolution" using viruses to develop proteins and other microbiological "products"; and developing algorithms for biosecurity purposes.
The second case study involved the Laboratory of Microbiology at Wageningen University in the Netherlands and focused specifically on one PhD researcher working on the development of a 7 The focus in these reviews is particularly on gene drives in insects, although other organisms are also addressed.[158] provides an additional overview of the state of gene drive development for insects specifically.[159] provides an example of one additional recent development in the field: the use of gene drives as tools for laboratory research. 8For mosquitoes, Bier [57, p. 19] remarks on the basis of his extensive review of the research in the field: 'Progress in the gene-drive field has been remarkable over the past 5 years.In this brief period of intensive productivity, nearly all substantive technical barriers have been overcome for drive systems either modifying or suppressing mosquito populations.'He adds that therefore, 'it is now possible to define relatively precise target-product profiles defining detailed characteristics of drive systems that would be suitable for advancing to the next phase of testing in physically or ecologically confined outdoor […] field tests' [57, p. 19].Similarly, Gantz & Bier [58,p. 17] conclude that 'The promise of active genetics at its dawn in 2016 has been fulfilled in nearly all spheres initially envisioned including the development of efficient next-generation genedrives, allelic-drives, drive neutralizing elements, and adaptation of these systems to other organisms including additional insect species, mammals, bacteria, and viruses'.For both case studies, observations, conversations and interviews were written down (permission for audio-recording was not granted) and these transcripts were subsequently analysed for how they involved ontological and morality-related matters.As with the literature analysis, the analysis of the ethnographic material was not aimed at being encompassing or systematic, but at coming to what in our view were the most interesting and remarkable features of the bioengineer's practices and views.

The Current State of the Gene Drive Debate
What image of the discourse on the implications of gene drives and how to deal with them emerges from our reading of the literature?As can be expected for an emerging technology, accounts of gene drives lie somewhere on the spectrum between emphasizing unprecedentedness, radicalness and uniqueness, and claiming continuity and drawing parallels with other developments and technologies, with many accounts acknowledging both sides.The influential NASEMreport from 2016 [1], which has since become a benchmark for subsequent publications, defined gene drives according to two features: the intentional rapid spread in the wild of the genetic trait, and the potential irreversibility of environmental consequences that result from their release.Simon et al. [59] characterize gene drive organisms as between continuity and novelty and stress how, on the side of novelty, the inheritance and spread of a transgene in wild populations becomes the explicit aim with gene drives, as opposed to the danger or risk such spreading represents in considerations of traditional genetically modified organisms.Moreover, the aim to change wild species comes with unpredictability in terms of effects in the target population, their environment, and ecosystems beyond the area of release, since modifications are unlikely to be limitable to the area of release.
The initial spur of excitement about gene drives can be explained by their potential to target difficult-to-control wild species: they are held to provide a solution to a specific set of problems in health, conservation and agriculture for which current solutions are considered lacking or unsatisfactory [1,17,19,60].However, the debate has arguably been characterized more strongly by claims of the technology's potential risks and negative consequences (e.g.[1,15,[18][19][20][21][22][23][24]26]).Theoretically, the release of even a few organisms could lead to the introduction of beneficial or harmful genes in entire wild-type populations, leading commentators to consider gene drives 'highly contentious' [61, p. S203].CRISPR-developer and Nobel Prize winner Jennifer Doudna refers to gene drives as a way 'in which the power to edit the genes of other species could prove to be more dangerous than any changes humans have made to the planet so far' by causing 'unstoppable, cascading chain reaction[s]' [62, p. 147] and a The New Yorker profile of Kevin Esvelt's gene drive efforts states that 'there has never been a more powerful biological tool, or one with more potential to both improve the world and endanger it' [63, n.p.].Claims of gene drives' power potential relate not merely to their possible effects, but also to the easy access to that power 11 [15,27,32].The highly critical Sustainability Council of New Zealand regards gene drives as a technology that 'radically exceeds the existing boundaries of human power over nature' due to it conferring 'the ability to rapidly eliminate or reengineer entire species that have evolved over millions of years' [43, p. 16] and to deliver 'extinction to order or […] permanent reengineering of wild species' [43, p. 62].Part of what makes gene drives such an interesting case is that they were kind of "stumbled upon" in laboratory settings [3,5], after which their potential real world effects-both intended and unintended, both beneficial and undesirable-were quickly realized and became guiding for both the development of the technology and the debate surrounding it.In this context, much of what has triggered attention and subsequently been said about gene drives has been a reaction to the hypothesized variant that "started it all": what has, with some hindsight, been labelled the universal or standard drive, which does not limit itself to local members of the species and has potentially wide population and ecosystem effects [15,21].Combined with the option of using a gene drive to eradicate, not modify, a population, this has led to the (un) popular image of this technology driving entire species extinct [43,64,65].Given the obvious gravity of this ultimate possibility, a lot of attention has since gone out to softening the sharp edges of the (image of the) universal extinction drive.This has been done in a number of ways.
First of all, there were the initial calls for laboratory safety and science self-regulation that emerged together with the first publications on the newly discovered potential of synthetic gene drives [4,15,27,59,[66][67][68] ( [69] provide a more recent reiteration).The aim of these calls was to prevent potentially highly effective gene drives from either escaping the lab or leaving it without due consideration, in the absence of a regulatory system catered specifically to gene drives.The underlying concern for some of these authors was to prevent harm to public trust in gene drive development and science more broadly [15,66,68].Secondly, and in further regard to this regulatory void [70], calls have been made to adjust existing risk assessment and regulatory frameworks to accommodate the novel features presented by gene drives 12 [1,15,33,39,40,42,53,59,61,[71][72][73][74][75][76][77][78][79][80][81].The point here is that existing frameworks are too restrictive to allow for the proper assessment and testing of gene drives, being based on prior biotechnological developments that lacked gene drives' spreading properties.These proposed adjustments to existing frameworks thus support the third category of responses, those that emphasize the need for further experimentation and testing, from laboratory to highly controlled field trail to first field release [1,13,15,16,39,71,77,82].The great potential of gene drives for both benefits and harms is here suggested to vindicate conducting further research under careful conditions in order to understand effects in natural systems and to assess the risks involved.Publications detailing the conditions for safe field trials belong to this category [22,33,[46][47][48][49][50].Fourthly, and responding to the diagnosis of a regulatory void in a different way, various publications have stressed the need for international governance agreements to fit gene drives' disregard for human legislative boundaries [1,14,15,21,24,29,31,36,42,43,46,48,75,[83][84][85][86].Gene drives' spreading properties are here taken to mean that a decision on a gene drive release ought to include governments of all territories where the gene drive modification might spread towards.
Fifthly, a substantial proportion of the more recent gene drive debate has revolved around the advocacy of public inclusion in general and of local community engagement and decision-making regarding gene drive releases in particular [14,25,38,41,42,46,48,50,69,71,84,[87][88][89][90][91][92][93][94][95][96][97][98].While endorsement of community engagement as a principle of decision-making is not specific to gene drives or to technology development, the emphasis it receives in the gene drives debate does appear linked to the technology's idiosyncrasies.The combined features of deliberately causing modification or decimation of a local population and the use of a novel mechanism of deliberate rapid genetic spread appear to strengthen these calls, as does the fact (in principle not gene drive specific) of the colonial history of the parts of the world where currently considered gene drive applications are located (such as New Zealand and Burkina Faso).Finally, and in close relation to the advocacy of community decision-making, more technical responses to gene drives' controversiality have proposed the development of altered or augmented drive systems that are local [15,99].The idea here is to limit the universality or irreversibility of the gene drive spread, or both, with proposals suggesting the development of drives that are self-exhausting [4,100,101], only spread if released above a certain threshold Page 7 of 26 4 Vol.: (0123456789) [102][103][104], reverse the change made by a previous gene drive release [3,105,106] 13 , or are designed to target specific parts of genes (alleles) that occur exclusively in particular local populations [47,48,99,107].There have also been proposals for "split-drives", which separate the core components required for a fully functional gene drive [105,[108][109][110][111][112][113][114]  14 .The connection to the community engagement proposals is that these augmented drives would make local decisions on a gene drive release legitimate decisions: a local decision about a local technology [32,100].
The relevance for our purposes of these six categories of proposals for responsible gene drive development is that despite their obvious specificities, each aims at defusing concerns over gene drives, at reducing the technology's controversiality, at making their future use morally legitimate.Gene drives were stumbled upon, after which their promises and perils were soon realised and strategies to neutralise concerns and legitimize their application were adopted: calls for (self)regulation that would not block further research and testing, calls for community engagement, and the development of technical solutions to control potential negative effects.We return to this legitimizing tendency in the gene drive debate in our analysis of Gene Drives as Interventions.What is important to note is that in its legitimizing and defusing focus, the gene drive debate tends to misrepresent scientific and technological development.The quick and prominent emphasis on addressing the normative, ethical or moral implications of gene drives disregards how these implications are deeply interrelated with fundamental questions about what gene drives are; that coproduction is taking place.As we will explain in the next section, how gene drives are ontologically framed cannot be regarded separately from moral views about gene drives.

Coproduction of Morality and Ontology
The Coproduction Thesis Inspired by contemporary scholarship in Science and Technology Studies and philosophy of technology, we have analysed the empirical (ethnographic and literature) material through a theoretical framework of coproduction that focuses on the interplay between ontological and moral assumptions in the gene drive debate.Frameworks of coproduction have become widely used to explore dynamic processes and feedback loops at the interface of science and society [115][116][117].The basic underlying idea is simple: 'the ways in which we know and represent the world (both nature and society) are inseparable from the ways in which we choose to live in it' [55, p. 2].While in Jasanoff's [55] influential framework the emphasis tends to be on the coproduction of "knowledge and norms"or epistemological and moral assumptions-our particular focus on ontology and morality can be regarded as included in her notion of 'how we know and represent the world'.Coproduction for Jasanoff is not primarily about normativity in relation to how we gain knowledge, but in relation to that which knowledge is about: reality, the world, nature-the domain of ontology.Highlighting specifically the ontological dimension of coproduction allows us to focus on a particular dynamic of the gene drive debate that is crucial for understanding its moral and societal dimensions: its deep entanglement with assumptions about the status, demarcation and properties of nature and human action.
Our reading of the gene drive debate suggests three limitations of the debate in terms of ontologymorality coproduction.First, whereas the debate has been strongly and explicitly concerned with gene drives' moral implications, their ontological implications have received far less explicit attention.As will be illustrated, gene drive development is suffused with often implicit ontological assumptions about nature and human action.Secondly, if explicitly discussed, the ontological implications of and assumptions underlying gene drive development tend to be presented in a rather abstracted fashion-gene drives would change how we understand and give meaning to evolution, nature, or wilderness.This leads to stalemates in the debate between those concerned with specific, problem-solving applications and those 13 Combinations of these types are also suggested [15]. 14These drives are more focussed on safety and controllability than localness per se.In what follows we devote more attention to the idea of "local" drives.

4
Page 8 of 26 concerned with moral and/or ontological questions.Finally, there is the divorce of the moral debate on gene drives from explicit recognition of ontological matters; the issue of misrecognition of their coproduced nature.
A key point of the coproduction framework is that coproduction is a pervasive and unavoidable social phenomenon.The coproduction thesis itself can therefore be regarded as an ontological claim about social reality.As a result, the concept of coproduction gives rise to a number of profound metaphysical questions 15 .We acknowledge these questions and their relevance, but consider addressing them beyond the scope and purpose of this article.Instead, we adopt the relatively uncontroversial version of the thesis that states that ontological and moral claims and assumptions are interconnected in scientific and technological development and debate.Against the backdrop of the wider coproduction thesis, we consider the main question to be not whether gene drives give rise to coproduction but rather what kind of coproduction.The answer to this question comes in two parts.First, gene drive coproduction relates explicitly to the ontology and morality of (parts of) nature; secondly, it relates to the idea of interventions into nature.

Gene Drives and the Coproduction of Nature
While any emerging technoscientific development can be analysed on the basis of ontological implicationswhat does it mean for our conceptions of ourselves, of reality, of the world, of categories used to distinguish between them?-due to their core features gene drives specifically draw the ontology of nature, and related questions of morality, right into the middle of technoscientific development.In a general sense, gene drives can be said to give rise to a coproduction of ontology and morality regarding nature in at least the following respects.First, gene drives make salient the question of the ontological and moral status of evolution (as part of the natural world more broadly), as they have been referred to as an intervention into evolution, as surpassing the rules of evolution, cheating evolution, and sculpting evolution [2,24,43,51,118].Second, gene drives' potential lays primarily with their use to target wild populations, bringing the ontological and moral distinction between the human-controlled and nonhuman world to the fore [1,3,18,19,24,43,53,54]. Third, and closely related, because gene drives are a naturally occurring phenomenon that has been discovered and subsequently noticed for its potential to be "harnessed" to devise human-designed variants [15,119], gene drives foreground questions of the ontological and moral status of artefacts and hybrids ( [17,67,92] illustrate gene drives giving rise to questions of naturalness; explicit ontological questions are raised by [1,19,44]).Fourth, certain potential gene drive applications give rise to the possibility of species eradication or modification, triggering discussion on what defines species and what their moral status is [42-44, 120, 121].Finally, the debate about gene drives has been strongly steered by the suggestion of their highly efficient and irreversible spread and various attempts to downplay that high potential on the basis of the "reality" of natural processes.Because of this, gene drives have given rise to questions of the ontology of nature in terms of the intricacies of its actual functioning-its processes, its mechanisms, its relations, its connectivities, its components, its predictability vs. its unexpectedness-and our morally charged reverence thereof [43,44,122,123].
As we will show in the following section, gene drives give rise to a coproduction of ontology and morality regarding nature in a particular way: they entail interventions into natural systems.The Sculpting Evolution lab defines its mission in part as seeking 'to develop tools capable of precisely intervening in the evolution of ecosystems' [52, n.p., emphasis added].Preston and Wickson [44] argue that gene drives have ethical implications on three levels: their impacts on human and planetary wellbeing, the intentions driving the technology forward, and the character of the interventions they entail.More generally, various policy, regulatory and ethical publications represent the end point of current gene drive development in terms of future gene drive interventions [1, 13, 15, 16, 18, 30, 32, 45, 53, 77, 84, 91-94, 97, 118, 120, 124-131].
We hypothesize that the idea of an intervention, and of an intervention into nature more specifically, comes with particular ways of making the connection between morality and ontology salient and that the understanding of gene drives as interventions into nature matters for the coproduction dynamic of gene Page 9 of 26 4 Vol.: (0123456789) drives.An intervention relation carries assumptions about what is intervened into or upon, but also about the intervener and the means and relation of intervention itself.Thus, an ontological assumption about what gene drives would intervene into is always also, at least implicitly, an assumption about the intervener and the nature of the action relation between the two.Moreover, importantly, the understanding of a gene drive as a specific intervention into a specific natural system or population is itself a powerful ontological representation, as it suggests the very possibility of cutting up reality into an intervention situation on the one hand and "the rest of the world" on the other.Understanding gene drives in terms of an intervention-oriented coproduction dynamic has implications for our ability for and approach to critical engagement with the technology.It highlights how critical engagement with gene drives needs to address empirical, moral and ontological concerns together, and that this in turn requires a perspective that zooms out from a narrow focus on specific intervention context and relation.Such a perspective reveals how a view of gene drives as specific interventions into nature obfuscates the interventionism already present in the bioengineer's lab prior to any field release.A Perspective that Acknowledges Coproduction-Discussion returns to these implications.In the following section Gene Drives as Interventions, we first elaborate the specific ways in which gene drives have been framed in terms of the different dimensions that constitute an intervention.

Gene Drives as Interventions
Adapting the notion of "intervention" from political science [132], we argue that potential gene drive applications have been represented, and can be analysed, in terms of the ontology and morality of intervention along four dimensions [133].First, because interventions occur between two parties or entities, the intervener and the intervened-upon, they come with distinctions between and definitions of who is causing change (the intervener) and who or what is being changed (the intervened-upon).Second, because interventions are cases of deliberate action, the boundary between intentional action and accidents and unintended consequences becomes ontologically and morally important.Thirdly, an intervention aims at realizing a desired change, and this aim is in turn based on regarding the intervenedupon as steerable, guidable, manageable and correctable and as in need of the projected change-all involving ontological and moral assumptions.Finally, interventions are deliberate actions for which the intervening party claims to have legitimating reasons 16 , which creates a space for asking for moral justification.The moral justification relies on the definitions of who intervenes, who is intervened into, intentionality, problem, goal, and control of the other dimensions-on the definitions of the ontology of the gene drive intervention.
We maintain the fourfold typology as much as possible in what follows; however, because the dimensions clearly overlap, the discussion at times takes up two or more dimensions together.

Gene Drives as Externally Initiated Interventions:
The Intervened-Upon One of the key ontological questions to ask about gene drives is: what is the "intervened upon", exactly?The different ways gene drives give rise to coproduction of nature return here, now seen through the intervention lens.Is the intervened-upon the population that is being targeted by the gene drive?Is it the genetic composition of the population/species first and foremost?Or is it the local ecosystem, in which said population has created the problem of invasion or disease it is now being targeted for?Given gene drives' self-propagating feature, can we understand the intervened-upon as something circumscribed in time and space in the way the designation ecosystem suggests?And does "ecosystem" do enough justice to the human presence in those local systems where gene drives are to be deployed 17  questions suggest a tension between acknowledging connectivity in nature and the ability to demarcate a gene drive intervention to convey legitimacy.We return to this tension in the sections below.
As indicated earlier, some commentators take a step further and argue that gene drives intervene into evolution, the evolutionary process, or evolutionary principles [1,24,32,43,44,51,53,118]. Such suggestions raise the question what evolution is, and whether it is best understood as one big process or as many different smaller ones.It also raises questions about its relation to the concept of nature-for example, is evolution the process that gives rise to nature, or does nature encompass evolution?Our case study of Sculpting Evolution provides illuminating reflections on this point.The ethnography revealed at least six understandings or uses of evolution employed by the lab members: (1) natural evolution, the evolution of evolution theory, involving selection pressure, reproduction and fitness; (2) evolution as directed evolution, analogous to dog or flower breeding: changing the genetic composition of a population in response to a chosen and imposed selection pressure; (3) evolutionary stability, or the maintenance or restoration of a population's genetic composition after disturbance; (4) evolution of information, or the view that development of informational patterns and complexity is the most fundamental ontological and moral feature of the universe; (5) evolution as our most important source of inspiration or teacher; (6) evolution as a metaphor for required changes to the scientific enterprise.
But regardless of whether the application of the group's technologies lies in using bacteria to optimize the functions of proteins or in eradicating invasive species with ecological purposes in mind, the actual mechanism of evolution the group works with is evolution on the molecular scale, not the ecological scalewith the latter being the realm of evolution of organisms, populations or species.Therefore, in the group's consideration of using a gene drive mechanism for ecologically informed purposes, two understandings or meanings of evolution are involved: evolution as any change in the genetic composition of a (natural or laboratory) population in response to a (natural or deliberately introduced) selection pressure, and evolution as the grand natural process of biology-Nothing in Biology Makes Sense Except in the Light of Evolution 18-which explains life on earth as we know it today and which helps us understand how and why ecological problems arise in the first place (populations coevolve with their environments).In the case of directed evolution, the group changes the selection pressure in order to stimulate a certain molecular composition to optimization-this is protein breeding 19 , as it were.In the case of using a gene drive in natural contexts, the group would change not the selection pressure-that would be classic conservationism-but the genetic composition of organisms (see also [45] on gene drives in conservation).An answer to the question of what a gene drive intervenes into depends on the understanding of evolution.The bioengineers entertain different uses of the concept, and work in their daily practices with a conception of evolution that only formalistically shares its meaning with what is commonly invoked with the concept of evolution as the grand process of life on earth.
Furthermore, the bioengineers of both groups connect a sense of awe in the face of evolution 20 to taking design inspiration from evolution for new biotechnologies 21 .To them, evolution is not something you intervene into, but something you can use.Evolution thus shifts from being defined as the Page 11 of 26 4 Vol.: (0123456789) "intervened-upon" to being part of the intervention tool, and this shift in the ontological understanding of evolution in the intervention relation has implications for the moral evaluation of intervention and a gene drive intervention specifically.When evolution is not that which you intervene into using a gene drive, but that which you use to intervene, evolution has already been usurped into dominant practice, into the box of tools at our disposal.In this understanding of gene drives and evolution, gene drives would both work with and against evolution at the same time, and therefore cannot so easily be understood as the intervened-upon.

Gene Drives as Externally Initiated Interventions:
The Intervener Moving from the intervened-upon to the identity of the intervener, the gene drive debate shows a further interesting dynamic.In the case of Sculpting Evolution's early gene drive and similar initiatives, the developer of the intervention technology was to an important degree also the intervening party.drives that allow incorporating local views into their design-leads to a transposal of the identity of the "intervener" from the bioengineer to the local community, and with this of the locus of the legitimacy requirement of the intervention.It is local communities that are seen as the legitimate party to intervene using a gene drive, and with this, their justifications become the focus of moral concern.
Moving towards a next dimension of intervention-the specific goal of bringing about a change to address a problem-there is an important further implication of this shift.The bioengineering potential of gene drives was first discovered or realized, and potential domains of application and actual contexts of use were only then envisioned and pursued.For a local community, however, dealing with a dire problem such as local biodiversity threats or malaria, the goal clearly precedes the means of intervention: the gene drive is the tool, one potential option among others.Various Sculpting Evolution members explicitly acknowledged the existence of these other options in conversation; however, their focus is necessarily circumscribed by the kind of contribution they can deliver, given their skills and expertise-if they want to contribute something to these problems, it will be a bioengineering solution.Yet, were the solution to follow the problem, rather than the other way around, gene drives may be more morally legitimate in the eyes of many.

Gene Drives as Forms of Deliberate Action
A lot of the initial energy in the gene drive debate went into drawing boundaries between accidental, unreflective, and malicious releases on the one hand, and deliberate, reflective, and benevolent releases on the other [1,3,27,28] 22 .This initial dynamic is understandable, given the stumbled-upon nature of the contemporary gene drive discoveries [3,5] and the spreading properties of gene drives: the atmosphere upon publication of the first proofs-of-principle in 2014 and 2015 tended towards crisis management, of making sure no disasters would occur, and the involved scientific community accordingly emphasized assuming responsibility and applying self-regulation.
The gene drive community attempted to distinguish interventions (deliberate, well-conceived, legitimated) from accidents and ill-conceived releases, and to distinguish "actual" interventions  [60,134]).But with gene drives, there are reasons to question the outcome of this boundary work.With gene drives, there is a sense in which there is no neat separation between the intentional and unintentional, between the deliberate and the accidental.The very mechanism that explains gene drives' desirability when applied intentionally-rapidly spreading a genetic trait in the wild with certain permanent environmental effects-is also the mechanism that makes them undesirable.Precisely because the gene drive is designed to create a cascading population effect through rapid sexual reproduction, and to do so in a "wild" context, a natural context of interactivitywhich is in terms of application its raison d'être-it can lead to all sorts of unintended consequences.
From the perspective of the coproduction framework, these dynamics of the gene drive debate involve connecting a certain ontological definition of gene drives to their moral evaluation.Attempts are made to associate conceptions of gene drives with the image of carefully considered interventions-although this occurs implicitly to an important degree.Recklessness and malintent are obviously morally wrong, but are not considered inherent to what gene drives are-the wider condition or context of their application determines their moral character.These (implicit) attempts at aligning the ontology of gene drives with that of the carefully considered (and thus morally legitimate) intervention trade on ambivalence in the ontology of gene drives.What is, and what is not, "part of" the gene drive?Should the effects of a gene drive release, both foreseen and unforeseen, both desired and undesired, both to the target population and to the wider ecosystem, be understood as part of what a gene drive is?Or are they outside of the definition of a gene drive, are they just "effects of"?Clearly, the more one includes such potential for harm, risk, uncertainty, and irreversibility into the understanding of what the gene drive is, the more difficult it becomes to argue that the legitimacy of a gene drive application is dependent on the specific application context.We return to these ambivalences in the next sections.

Gene Drives as a Specific Intervention-Steerable, Guidable, Manageable and Correctable
Part of the debate about gene drives has turned on whether that which the gene drive intervenes into can actually be considered steerable, guidable, manageable and correctable 23 .May we assume this for both the gene drive mechanism over successive generations of sexually reproducing organisms and for the local ecosystem in which it is deployed?To what extent?While the power of gene drives to affect natural populations is what triggered the specific trajectory the debate took, the extent to which gene drives would be precise and efficacious is very much dominating the scientific development of the technology at the moment [57,58].In both the Wageningen and MIT group, experimentally attaining a functional (gene) drive proves to be more difficult than expected 24 .Various scientific publications similarly indicate partial success stories [1,11,[135][136][137][138][139][140][141], with stability of the gene drive mechanism over subsequent generations being the main issue.
Page 13 of 26 4 Vol.: (0123456789) While many characterizations of gene drives refer to the effects on the target species, a common concern is the irreversibility of wider ecological consequences following gene drive release: 'no matter how we alter an ecosystem […] we can't guarantee that it will return to its original state, even if we remove the source of the change' [32, pp. 23-24].Similarly, Esvelt & Gemmell [21,p. 2] state that 'the bottom line is that making a standard, selfpropagating CRISPR-based gene drive system is likely equivalent to creating a new, highly invasive species: both will likely spread to any ecosystem in which they are viable, possibly causing ecological change'.Even regarding the proposals for drive systems that would not have these universal (standard, self-propagating) tendencies, Oye et al. [28, p. 1] remark that 'ecological effects would not necessarily be re-versed'.A recurring related topic in the gene drive literature is the complexity of the systems in which gene drives are to be deployed and the lack of knowledge and uncertainty this entails.Here, we recognize the final dimension of the coproduction of nature (see Gene Drives and the Coproduction of Nature) that revolved around the "reality of nature" being one of recalcitrance and unpredictability.When molecular biologists and chemists take their synthetic creation out of the laboratory and into the wild, they enter the realm of ecology, where the regularities of experimental conditions are nowhere to be found: 'our knowledge of ecosystem-level species interactions is limited' [15, p. S51] 25 .
In these contributions to the debate, specification of the uncertainty of engineering complex systems thus amounts to providing an overview of the potential wider risks of gene drives.In this regard, many of the factors commonly referred to under the rubric of "risks" in gene drive publications may be more appropriately understood as uncertainties, given that little or no reasonable estimate of the likelihood of occurrence nor a specific description of the effects can be provided.These characterizations of gene drives challenge the assumption that the involved intervention is based on knowing that the projected effect will actually be realized.The ethnographies revealed an interesting parallel between conversations of the two groups in this regard.In both cases, the bioengineers showed themselves concerned mainly with getting something to work, not with understanding why or how it works the way it does.Moreover, both groups displayed an ambivalent attitude to acknowledging uncertainty and unexpectedness.The role of uncertainty is acknowledged, but to a degree.Reference is often made to biology or nature "always finding a way" and being "unpredictable", while working on "reliable" and "safe" interventions into biology or nature; a vision of the recalcitrance of nature and life is combined with the effort to design reliable tools from and for natural and living systems, as if the two visions were not mutually exclusive.
In navigating this space between gene drive, wider ecosystem, control and uncertainty, there is discursive boundary drawing at play regarding the steerability, guidability, manageability and correctability of the intervened-upon.The development and discursive legitimation of nonuniversal and reversible drives [15,100] accepts the possibility of the universal drive or the priorly released drive not doing what you want it to do-either explicitly or implicitly-and responds by providing an antidote, as it were.But what provides the certainty for the antidote working as planned, when it is designed to counter the lack of certainty of the original gene drive construct?Furthermore, these proposals for augmented drives rely on the narrow circumscription of the intervention situation.They appear intended to bring the geographic and temporal scales implied by the universal gene drive [142] down to a level where steerability, guidability, manageability and correctability seem more plausible.
It is relevant to return here to the early dynamics of the gene drive debate.The recognition of the power of gene drive technology (see The Current State of the Gene Drive Debate) led to calls for limiting the technology's power by designing modified drives [21,27,58,99,100].A relevant question here is the extent to which the "power" of the gene drive mechanism should be assessed depending on how far its effects reach.Would a highly localized, but highly effective gene drive not be a very powerful technology?Would it not, in a way, be more powerful than 25 Some authors have specified such general references to uncertainty and lack of knowledge into overviews of knowledge and research gaps [1,17,18,22,92,128], which converge on mentioning factors such as mutations of the gene drive genetic element, gene transfer outside of the target species, reproduction and mate selection patterns, effects of species decline or eradication on ecosystem functioning, and likelihood of ecological niche of eradicated species becoming filled by other species.

4
Page 14 of 26 one that is less "controlled"?Here, we see an important point of division between those that would challenge gene drives' moral justifiability on the basis of its unforeseen effects and those that would do so on the basis of its intended effects.For the former, improved controllability means reduced power; for the latter, it can also mean increased power of the technology.

Gene Drive Intervention as Coproduction
The previous sections showed how diverging understandings and representations of intervener, intervened-upon, intervention tool, intentionality, problem-solving and manageability shape moral reasoning about the legitimacy of a gene drive intervention.The morality of a gene drive intervention runs through each of these attempts at settling the ontology of the intervention and the domains-human and nature-it connects.In one sentence: the ontological dimensions of the intervention relation are inseparable from its moral dimensions.
The following self-description by MIT's Sculpting Evolution provides as good a summary illustration as any of how ontology and morality come together in the development of gene drives: Evolution gave rise to every living thing and all of human culture, but evolved systems are very different from those designed by humans.They're harder to predict and to design, and exhibit a frustrating tendency to evolve away from engineered behaviors.At the same time, harnessing and directing evolution can generate useful organisms and biomolecular tools that we could never have rationally designed.Our laboratory seeks to understand why systems evolve in the ways that they do, to develop tools capable of precisely intervening in the evolution of ecosystems with the guidance of interested local communities, and to cultivate wisdom sufficient to know whether, when, and how to proceed.[52, n.p.]The passage reveals the ambivalent attitude towards evolution: the source of all of life, including human life; the superior process for designing organisms and molecules; the frustrating, ungraspable phenomenon which we need to understand better; the mechanism we can put to use, even without fully understanding it.It includes explicit reference to intervening in the evolution of ecosystems.It introduces the notion of community guidance in relation to ecosystem intervention.It implicitly refers to the demarcation between lab and outside world in the idea of precise intervention in the field through the tools developed by the group in the lab.And it alludes to wisdom and its cultivation, introducing a (perhaps the) fundamental philosophical and moral concept into the group's main objectives.
The above analysis supports the hypothesis that the idea of an intervention, and of an intervention into nature more specifically, comes with a particular mode of making the connection between ontology and morality salient.Who the intervener is, how we define the intervened upon, how we circumscribe and distinguish the two; when a course of action is deemed an intervention, a form of deliberate action, and when it is something else; who is considered to legitimately speak on behalf of the intervened-upon and its needs-these are all matters bringing ontological and moral concerns together, foregrounded by the idea of intervention.Furthermore, questions about what nature is, how its relations, processes or components should be understood regarding the idea of nature, how fundamental and unchangeable or dynamic and malleable it is, what evolution is and means, and how human beings should act towards, within, or as part of nature are additionally triggered by the idea of a gene drive as an intervention into nature.
However, what also emerges from the analysis is that if a gene drive is an intervention into nature, it is a quite particular one.The explicit aim of causing deliberate, rapid spread of genetic modification through a population (something to be avoided at all cost with regular GMOs [59]); the type of thinking about scales and demarcation such spreading triggers; evolution as that which the gene drive intervenes into, but also that which the technology uses or is based on-these and other dimensions of gene drives give a particular ring to its coproduction of ontology and morality along the intervention axis.Moreover, the analysis shows that a lot of discursive work goes into bringing morality and ontology into alignment in such a way that the image of a legitimate gene drive intervention results-an endeavour that proves difficult, given gene drives' particularities.Gene drive development is imbued with particular, partly Page 15 of 26 4 Vol.: (0123456789) contradictory, conceptions of nature, ranging from biochemistry-based denials of boundaries between artificial and natural, to defenses of gene drives on the basis of their "naturalness" and pleas against the technology for interfering with natural processes.
To what degree gene drive interventions should actually be embraced remains deeply contested (e.g.[43,44]; see also various responses to gene drives by NGOs and activist groups such as the ETC Group [143] and Friends of the Earth [144]).The aforementioned particularities explain why gene drives are a controversial emerging technology, and why so much work has gone and is going into turning the initial image of a universal, irreversible, globally spreading, species-eradicating gene drive into something less controversial, something more politically and socially legitimate(d), something more morally acceptable.In the next section, we reflect on these efforts of gene drive development and debate by proposing what our framework of intervention-oriented coproduction could bring to the way we understand and give meaning to gene drives and by extension to technoscientific development more broadly.

A Perspective that Acknowledges Coproduction-Discussion
We have seen how gene drives can be analysed in terms of their character as interventions, and how this reveals particularities of their coproduction dynamic.One issue that has remained implicit in this discussion, however, is that the representation of gene drives as involving particular, discrete interventions is allowing only certain ethical or moral issues to come into view; that is, only those issues that relate to the particular intervention context and relation.This implicit representation relies on coproduction for its framing: it connects ontology and morality in a certain way and represents them as local phenomena.In what follows, we suggest that adopting a wider analytical perspective that takes the coproduction thesis seriously challenges the ontological and moral assumptions of gene drives as specific, future interventions into nature.
A perspective that acknowledges coproduction provides a crucial challenge for the moral evaluation of gene drives.After the high-profiled 2014 and 2015 publications, the debate turned to case-specific descriptive and normative ethics [18,26,69,91,94,125,127,[145][146][147][148] and proposals for responsible further development [1, 14, 15, 21-23, 27-35, 37] quickly-perhaps too quickly.However, it is difficult to engage in a first-order normative evaluation if normative disagreements partly arise from different ontological perspectives of how to situate gene drives in competing understandings of nature.Furthermore, proposals for responsible further gene drive development tend to cherry-pick, locating ontology and morality in some places but not others, and denying their entanglement.These quick turns to normative ethics and responsible science proposals leave a number of key fundamental questions of coproduction unturned.One of these asks whether human beings are capable of including the nonhuman world into their ethical analyses nonanthropocentrically, and if so, how this can be done.A second asks to what extent and in what way we can ask and attempt to answer ethical questions without talking about ontology at the same time (or at least without accepting that one is grossly oversimplifying the answer).A third, in close relation to the former, asks to what extent the clearly demarcated case of the ethical analysis has truth-value concerning the actual, much more fluid, moral universe.
Asking such questions of coproduction leads to a number of insights about ontological and moral considerations relevant to gene drives that escape the particular intervention context.First of all, if our moral concerns only focus on specific gene drive interventions this would leave concerns about a wider context of interventionism outside of the consideration.Secondly, interventionism is revealed to already start in the lab, prior to any potential "real-world" intervention.These two insights are addressed in what follows.

A Wider Context of Interventionism as the Key Moral Issue
The focus on gene drives as particular interventions has as part of its coproduction dynamic the effect of leaving a broader cultural context of interventions outside of the normative consideration.However, this wider cultural context of human interventions into nature-of interventionism-forms for many the very moral background condition to the question of why gene drives are morally relevant or controversial in the first place.One direction for the future of gene drives would be to embrace their transformative potential as part of a wider interventionist perspective on technology as tools for fundamentally reshaping nature and its underlying mechanisms (as discussed by Sandler [45]).However, interventionism is deeply contested and is not the only choice for the future of human-nature relations, nor more specifically for the coproduction of ontology and morality regarding nature [149].Traditional conservationism, for example, resists the promises of reshaping nature through the power of technologies, diverging from interventionism in its normative assessment and in defending a stricter ontological-moral division of nature and culture.Alternatively, for approaches inspired by (among others) ecofeminism, political ecology, and Indigenous ethics that emphasize reciprocal relationality with and agency of nonhumans [150,151], the ontological-cum-moral objection to interventionism is directed at the hierarchical structure of its humannature relations, in which humans intervene and nonhumans are intervened-upon.
A shift in perspective from intervention to interventionism has implications for claims to legitimacy of particular gene drive interventions, that rely-as we have seen-on an assumption of demarcatability.It raises doubts about whether considerations of specific gene drive interventions can actually exclude a wider cultural context of interventionism.Gene drives can be regarded as reflecting a history of human interventions into nature and a wide range of contemporary technologies and interventions.Moreover, the use of a gene drive to address conservation, disease eradication, and agricultural problems can be seen to reflect the grand contextual issue of human dominance on a planet of anthropogenic climate change and ecological crises.In conservation, for example, gene drives are considered and defended according to an "end justifies the means" argument [45], where the end entails maintaining the state of the system as it has recently existed.However, today's ecological disruption is not locally determined-its root is understandable only as a to an important degree supralocal-global, earth system-phenomenon.To fight the ecosystem disruption using a gene drive under a last resort legitimation is to leave the root problem-a state of heightened disruption due to anthropogenic (climate) change-outside of the definition of the intervention context and outside of the legitimation.
The gene drive debate has tended to steer our imagination of the relevance of gene drives towards the extremes of reckless accidental and deliberate release, and at times malicious release [1,3,27,28].But recklessness and maliciousness are morally unambiguous-we know that they are wrong, and thus we concern ourselves with ways of preventing them 26 .From the perspective of a critical outlook on interventionism, however, the most important realm of morality for gene drives lies in the in-between of these two unambiguous extremes: the realm of deliberate, well-intended release or intervention.Particularly the emphasis on unforeseen consequences, risk and uncertainty [1, 15, 19-21, 23, 27] illustrates how interventionism itself is not what is normatively on the table: we are not challenging the use of interventions to address human-nature problems, we are challenging "reckless" intervention or non-intervention (accidents).If there is reasonable certainty that the desired change to the intervenedupon is to occur as projected, we allow for it.
The Interventionism in the Bioengineer's Lab In this regard, interventionism arguably becomes a stronger presence with every attempt to technically modify gene drives such that they become more localizable and controllable.By moving from the possibility of accidental releases and safety measures to guard against those to deliberate releases, for the purpose of great benefits, and the specific conditions under which those may occur, the gene drive discourse moves further and further towards interventionism.Limiting the gravity of the intervention "in the wild" can only be accomplished by increasing the interventionism "in the lab".A wider coproduction perspective challenges the commonsensicality of the view that gene drives would involve interventions only once they are field-released.Important boundary work occurs here: the moral questions arise only once the gene drive gets into contact with the "outside world".What precedes the real world gene drive application, however, is the bioengineer's development of the technology.The split between outside and inside the laboratory-a powerful ontological Page 17 of 26 4 Vol.: (0123456789) representation-masks how interventionism starts in the lab.
To illustrate the looming presence of interventionism in the lab, consider the Wageningen group's PhD researcher.The project revolved around the creation of a split drive in yeast with further research as its main application.Thus, the precise design is not to be applied to one of the main "real world" problems of human health and environmental and biodiversity challenges.This particular feature of the involved design suggests a distance from the interventionist potential of gene drive technology mentioned earlier.
However, what became clear during the ethnography was that this potential was always present in the background of the daily work of the PhD researcher's project.While getting a functional split drive in the lab was the (at times frustratingly difficult to achieve) objective, the envisaged subsequent step was to work on a drive system to support the work of using them "in the wild" in a hoped-for acceptable manner.Furthermore, in explaining the research to students and interested parties, the morally controversial nature of gene drives and their future use as interventions was a central part of the PhD researcher's narrative.The key take-away here is that in both the experimental work and in the context the PhD researcher tried to navigate, the intervention(ist) potential of gene drives was always present in the background.
More importantly perhaps, questioning the view of nature as that which exists "out there" recognizes that the core of the bioengineer's work is intervention into nature.The idealized image of the scientist is that of someone who attempts to understand nature.Our experience in observing conversations between bioengineers conversely emphasizes the extent to which these scientists do not necessarily know how things work, but that they work and to what effect.Their objective is to realize an effect, to execute a desired change.Sculpting Evolution's aforementioned views on "directed evolution" are a case in point here.What further emerges from both ethnographies is that bioengineering is discursively represented as a domain of ultimate opportunity.The sky is the limit, the only limitation is your imagination is what the Wageningen group's PhD supervisor tells bioengineering graduate students during a lecture on the future of CRISPR; you can engineer (almost) anything is the name of a bioengineering course provided at MIT in which Kevin Esvelt participates.The limitation to what is possible is considered in principle not to exist scientifically, and limitation can therefore only stem from two sources: your own limitedness of imagination and ability, or the outside world that imposes it.
In the latter regard, it is worth mentioning that the Wageningen group's PhD supervisor, during the aforementioned lecture, referred to "ethics" as one of three challenges to further CRISPR development.When asked about this representation, the PhD supervisor explained that we should stop doubting the technical side of CRISPR and that we should move on to addressing the involved questions of ethics-questions "we as scientists cannot provide" and need to be provided by society (personal communication).This underlines how the demarcation of lab from outside world is often also understood as a moral demarcation.For the PhD supervisor, ethics is perceived as something belonging to the domain outside of the lab, to society.More generally, the gene drive debate has tended to embrace similar representations from the outset: once something escapes from the lab, we enter the domain of ethics-the moral world is the outside world, and it invades the lab only by grace of a projected potential mishap, the origin of which can be traced back to the lab.

Conclusion
The debate about gene drives has been explicitly moral from the outset.The concerns about undesirable releases of gene drives that emerged together with their initial promises and the technology's poster child applications in health and conservation pulled the gene drive debate-including, notably, individual scientists and research groups-deeply into discussions about the ethics and politics of science and technology.However, the connection between this normativity on the one hand and underlying ontological assumptions about gene drives, nature, and human-nature relations on the other-so important to understanding the precise nature of the moral claims being made-has tended to remain implicit.
Our focus on intervention has aimed to make these connections more explicit.The intervention concept focuses our attention on the intervener, on what is to be implemented, and on the relation between the two.It emphasizes the importance of projection into the future, a feature central to technoscience in its concern with providing "real world solutions".With gene drives, this temporal dimension is crucial.After gene drives were stumbled upon in laboratory settings, their potential future effects were quickly realized and became guiding for both the development of the technology and the surrounding debate.Thus, in a leap of imagination, as it were, the potential future effects of gene drives were projected and made into a constitutive part of the technology in one move.
Nowhere in this process-that occurred rapidly and outside of formal regulative and policy-making institutions-was there a real world application or experiment present yet.Developing a critical, considered account of gene drives requires recognizing the coproduction of ontological and moral assumptions and recognizing that it is channelled by a framing of gene drives as potential future interventions into nature.Our ethnographic and literature analyses have shown that gene drive development is not a value-free repository of facts from which others develop ontological and moral assessments.Contemporary gene drive development is guided by a specific, intervention-oriented, form of coproduction.These considerations highlight the need for a broadening of the perspective on gene drives in which empirical, moral, and ontological concerns are addressed explicitly and in their interplay rather than in (disciplinary) isolation from each other.
What could such an integrated perspective bring?We conclude with an illustration.A critical reading of an article in Science [39] by a group of key contributors to gene drive development and debate 27 suggests how the advocacy of local community engagement provides only partial recognition of-or "cherry picks"-morality and ontology and their interplay.The authors introduce the relevance of gene drives by suggesting that their transformative potential does not in the first place reside in the actual transformations to the natural environment they would effect, but rather in their potential influence on 'the way societies address a wide range of daunting public health and environmental challenges' [39, p. 1417].In the vocabulary of this article: gene drives' transformative potential lies in embracing a particular variant of intervention(ism) to address certain human-nature problems.This is an astute observation very much in line with a more considered view on technology, intervention, and the coproduction of ontology and morality.However, while the authors frame the stakes of gene drive development in terms of this substantive point, the actual normative content of the article does not reflect this insight 28 .The article is about due (fair, just) process of gene drive field testing.
It is to an important degree appreciable that the authors leave the "substance" to others, to the public, to the local communities where gene drive interventions may be deployed; all they want is a good process that makes sure the public can bring their substantiveness to the table.But there are issues with this position.Given the repeated emphasis on the centrality of local community involvement, it suggests that the substance of gene drives-their moral and ontological implicationsis a local affair.Interestingly, this representation of gene drives implicitly presents a meta-position on coproduction; namely, one which does not so much suggest that certain moral and ontological assumptions come together, but that ontology and morality-as social phenomena-themselves exist and interact locally.In other words, this group of initiates, representative of the gene drive debate, is engaging in coproduction as a framing strategy (though they may not be aware of this).However, while recognition of cultural diversity, democratic decision-making and colonial past rightly supports the attitude of giving local communities the determining voice in these matters, we can question the extent to which ontology and morality and their coproduction can be understood as local.In an important sense, ontological and moral commitments and assumptions are considered to be true non-locally by those espousing them.Acknowledging ontology-morality coproduction in relation to interventionism (as we introduced in A Perspective Page 19 of 26 4 Vol.: (0123456789) that Acknowledges Coproduction-Discussion) further supports viewing gene drives as a non-local technology.
Adopting a more integrated coproduction perspective can thus help bring to light particular tensions in proposals for morally and politically sound gene drive development.It suggests that the global vs. local gene drive juxtaposition of the gene drive debate has fed into a very spatial, scalar conception of technological intervention and its influence: in order to make the decision-making context a manageable, local one, the gene drive ought to be made such that its effects are local as well.But does technological influence work that way?Can we simply trace the lines of the genetic inheritance pyramid to see where a gene drive's effects lead or end?The implicit suggestion in the proposals for local community engagement in combination with local drives is that if the gene drive is not universal/self-propagating, is limited spatially, temporally, or both, it loses much of its controversiality.As long as the decision-making process is ethically sound, the resulting gene drive interventions will be so as well.Philosophically speaking, this is debatable, for it makes morality a matter of decision-making processes-it confuses politics with morality.
The point in raising these considerations is not to challenge the obvious necessity or unavoidability of localizing decision-making-of narrowing-down-in order to make matters concrete and practicable.The point is to reveal the underlying implicit ontological-moral assumption that such a narrowing-down is more than an unfortunately necessary reduction of ontological and moral complexity; that by such narrowing-down and cutting up the world into specific intervention contexts, all factors and considerations considered non-or supralocal are excluded not because they are too difficult to incorporate, but because they actually do not form part of the ontological and moral reality of the local situation.Neither the development nor ethical assessment of gene drives can circumvent the wider challenge of competing modes of coproducing ontology and morality of gene drives and the issues with demarcatability these suggest.ducting this study.The authors have no relevant financial or non-financial interests to disclose.

Declarations
Ethical Approval Keje Boersma's PhD research project has been approved by the Social Sciences Ethics Committee of Wageningen University.
Informed Consent Informed consent was provided by the members of the two research groups who participated in the study.The directors of both participating research groups approved of the final manuscript in terms of references to their group and its members.

4 Page 16 of 26
questioning these framings of gene drives in terms of what they highlight and what they fail to accentuate.The notion of coproduction suggests that ontological and moral claims and assumptions are inseparably connected and mutually define each other in scientific and technological development and debate.As Scott and Marshall[56, p. 531] point out, any 'understanding of the world, or some part of it, must make assumptions (which may be implicit or explicit) about what kinds of things do or can exist in that domain, and what might be their conditions of existence, relations of dependency, and so on.Such an inventory of kinds of being and their relations is an ontology.'An ontological assumption is then an assumption about what kinds of things do or can exist, and what might be their conditions of existence, relations of dependency, and so on Vol.: (0123456789)5.We explore the connection of ontology with morality in light of the recurrent framing of gene drives as potential future interventions into (parts or dimensions of) nature.Our point of departure is that the idea of an intervention, and of an intervention into nature more specifically, relies on connecting morality and ontology in particular ways.By devoting our attention to these particular ontology-morality connections, we provide a new perspective on what is at stake in considerations of the future use of gene drives.Looking through the lens of the coproduction framework reveals that gene drives are framed as demarcatable, local, specific, well-considered and therefore legitimate interventions.However, such a framing of a gene drive intervention into nature as an isolated event into a neatly demarcated piece of the world is inherently incomplete.Our framework further highlights how the gene drive debate's early emphasis on ethical issues is troubled by the fact that it is difficult to engage in such normative evaluation if normative disagreements arise partly from different ontological perspectives of how to situate gene drives in competing understandings of nature.
This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this licence, visit http:// creat iveco mmons.org/ licen ses/ by/4.0/.