The challenge

Let me begin by explaining how ‘dark innovation’ fits into a broader critical studies of innovation agenda. To date, no one has directly confronted Ben Martin’s (2013, 2016) version of the dark innovation challenge. Martin’s challenge is mentioned briefly by Alf Rehn and Anders Örtenblad (2023) in the introduction to their edited collection Debating Innovation. But the only direct inquiry was a special issue of Industry and Innovation that used ‘dark innovation’ as a euphemism for the negative dimensions of innovation (Coad et al, 2021). This was an important acknowledgement that innovation sometimes has nefarious intentions and/or outcomes. Personally, I have been disturbed by history books on companies like IBM (Black, 2001), DuPont (Ndiaye, 2007), and I. G. Farben (Hughes, 1969). These books describe how such companies gained strength by developing tools for mass murder. And so, I am worried about how ‘bad’ innovation is so often ignored by innovation research. There is an undoubtable pro-innovation bias (Godin and Vinck, 2017a) where scholars, policy makers, and everyday citizens ignore, marginalize, and brush past the innovation that is bad for us and bad for this planet. But I come at this issue sideways. In this book, I try to capture ‘bad’ innovation – and more – under the umbrella of ‘dark innovation’. I take this broader view because innovation studies also neglects many innovation activities that are good for us. I am interested in the many varieties of innovation – and ‘novation’ (Godin and Vinck, 2017a, p 3) – that are cast into the dark. And so, my approach expands on Ben Martin’s (2013, 2016) dark matter metaphor. Dark innovation could be anything absent from our observations.

In this book, I use the example of public innovation in physical goods. Prior research on public innovation (see a review in de Vries et al, 2016) has ignored, dismissed, or minimized innovation in goods. With few exceptions (for example, Bugge et al, 2011; Arundel and Huber, 2013), there is a common rhetorical pattern: first, deploy anecdotes about novel public sector inventions (such as the Global Positioning System [GPS]) to convince readers that public organizations innovate; next, ignore public innovation in goods and either characterize the state as a facilitator of market-based innovation (as in Mazzucato, 2013, 2016) or shift focus to any public innovation aside from physical goods (as in Walker et al, 2002; Halvorsen et al, 2005; and Koch and Hauknes, 2005). For example, a prominent book in the field begins with a chapter that mentions public innovation in medical technologies, medical instruments, and pharmaceutical products (Windrum and Koch, 2008). The same book concludes with the surprisingly definitive counterclaim that ‘technological innovations, especially goods, are the exclusive domain of the private sector’ (Windrum and Koch, 2008, p 239). It has been said that the ‘appropriate’ way to understand product innovation in the public sector is by focusing on services (Walker, 2014, p 23). Public innovation in goods is considered inappropriate. Therefore, it is sidelined from innovation studies.

This prejudice is both political and methodological. Throughout this book, I show how common methodological tools and techniques carry the ‘neoliberal bias’ (Fløysand and Jakobsen, 2011; Cooke, 2016), neoliberal ‘dogma’ (Lundvall, 2016), ‘market bias’ (Gallouj and Zanfei, 2013; Cruz et al, 2015) or ‘market ontology’ (Pfotenhauer and Juhl, 2017) that dominates innovation studies. This is not quite the full-blown ‘phobic reaction to the state’ shared by early proponents of neoliberalism (Peck, 2010a, p xii). However, it is clearly a variation on the neoliberal theme: there is a sense that government should be constrained in favour of private companies. What is interesting is how this theme plays out in innovation research. I do not approach neoliberalism as a coherent concept or use it as ‘a heavy-handed tool of social analysis that knows its answer in advance’ (Phelan, 2014, p 2). But I do think of it as a grand narrative – one that is ‘somewhat ambiguous and situationally specific’ (Peck et al, 2018, p 4). My interest is in specific ways in which mainstream innovation studies constrain observations of public sector innovation.

Along the way, I am making a broader point. I am exploring the neoliberal bias against public innovation because it is one example of the ‘dark matter’ that lies beyond the edges of innovation studies. This is where I depart from Ben Martin’s version of ‘dark innovation’. For him, dark innovation is a result of deficits in technique: ‘the challenge to the next generation of researchers is to conceptualize, define, and come up with improved methods for measuring, analysing and understanding “dark innovation”’ (Martin, 2016, p 434). But my goal is not to produce a set of ‘less biased’ methods that yield ‘more objective’ observations. Different methods only bring forward different understandings. Instead, I reframe ‘dark innovation’ as a call to deconstruct the central assumptions of innovation studies. Following the work of John Law, my goal is to ‘scrape away the self-evident to understand and question how methods structure the world’ (Law, 2016, p 53). This book is not about filling knowledge gaps through improved methods; it is about problematizing the decisions that create dark innovation in the first place – the implicit and explicit choices researchers make about what ‘counts’ as innovation.

Problematization

Of course, some understudied forms of innovation are already being advanced through a gap-finding logic. I could have done the same in this book. Starting from some hint in the literature or a personal hunch, I could have searched for and attempted to observe some previously unreported innovation phenomena. I could have asked: ‘Where might we observe dark innovation?’ (Or, more directly, I could have asked something like: ‘Do public organizations produce innovative goods?’) However, that kind of gap-spotting logic would only yield an incremental contribution (Alvesson and Sandberg, 2011; Sandberg and Alvesson, 2011). The grand challenge would remain untouched. This is because gap-finding maintains a high degree of path dependence in any research field (Palmer, 2006). As Alvesson and Sandberg state, ‘gap-spotting is more likely to reinforce or moderately revise, rather than challenge, already influential theories’ (2011, p 25). Yet, it is the most common approach to framing research questions in the social sciences (Alvesson and Sandberg, 2011). It is ironic that this gap-spotting norm has taken hold in innovation studies – where most researchers applaud novelty. If we aspire towards radical and disruptive insights, we need problematizing research questions (Sandberg and Alvesson, 2011). We need to identify, question, and undermine our social scientific norms. And so, this book asks: why do we struggle to observe dark innovation?

We already know that novel contributions to innovation studies (IS) are experiencing a ‘rough ride’ through peer review because of overly strict adherence to disciplinary conventions (Martin, 2016, p 440). Leading innovation researchers have warned that the field is beginning to struggle with ‘disciplinary sclerosis’ – the rigidity that comes from standardizing and normalizing as a discipline (Fagerberg et al, 2013; Martin, 2013). Innovation researchers have been closing ranks around certain theories, research questions, empirical contexts, and methods. Within the agreed boundaries of ‘the field’, there is disciplinary enforcement of shared values and practices. There is the expectation of a ‘fairly standard form’ of academic writing (Martin, 2016, p 440). The field of innovation studies is paradigmatically stuck.

In response, Jan Fagerberg has argued that it would be better to accept innovation studies as an interdisciplinary ‘mongrel’ (Fagerberg et al, 2013, p 11). Apparently, this means abandoning aspirations towards scientific ‘pedigree’ and instead ‘engaging in fruitful intercourse’ with neighbouring fields, like business history and science and technology studies (STS) (Martin, 2016, p 440). Even setting aside the elitist metaphor about breeding, this suggestion is naïve. We are talking about fields with drastically different social scientific norms – different ‘epistemic cultures’ (Knorr Cetina, 1999). They have not always been so different. In the 1960s and 1970s, notable academics like John Langrish, Donald Marquis, Nathan Rosenberg, Michael Gibbons, and Michel Callon moved between the subjects of innovation, science, and technology. But one of the few recent instances where innovation studies has accepted ‘inputs’ from STS has been the area of innovation for sustainability, and that work ‘was regarded as rather “flaky” by some in IS’ (Martin, 2016, p 435). Culture clash is now inevitable; in fact, I think we should encourage it.

In this book, I use ideas from elsewhere to show that dark innovation is a byproduct of innovation studies norms. I can do this because I am a bit of a scholarly mongrel. During my PhD studies, I learned that my values and assumptions do not fit within mainstream innovation studies. That PhD research became the foundation for this book. But this work does not fit any better into any other field. I am not quite a business historian, although I have training and plenty of colleagues in that area. I have no claim to affiliation with STS, although I have read a great deal of STS work. Nor am I a geographer, although my master’s degree was in that field. I am too critical to be a ‘proper’ business or management scholar. And as I learned in the review process for this book, I am not sufficiently concerned with labour to fit neatly into critical management studies (CMS). In the end, I am not concerned with fitting into a discipline or field; instead, I am concerned with disrupting disciplinary convention. In the following chapters, I will use instrumentalities from STS, CMS, critical organizational history, and critical geography to confront the rigidities of innovation studies.

Instrumentalities

The empirical material for this book comes from an area of science and technology where disciplinary boundaries are very fluid (pardon the pun). I examine innovation in ocean science instruments. In Chapter 6, it will become abundantly clear that ocean science – sometimes called oceanography – is ‘not so much a science as a collection of scientists’ (Bascom, 1988, p xiii). These scientists come from biology, physics, chemistry, mathematics, and other disciplines. Many of them have been united around their own dark matter challenge: we currently have better images of the Martian surface than 85 per cent of our ocean floor. We are ‘in the dark’ about much of what lies below the ocean surface. Throughout this book, I will show that we are also in the dark about the key role of public organizations in developing the necessary new tools and techniques to understand our changing ocean. These are devices like the wave-powered ocean profiler that was developed in the mid-1990s at the Government of Canada’s Bedford Institute of Oceanography (BIO) (Fowler, 1997). That ‘instrument assembly’ (Fowler, 1997, p 1) offers an energy-efficient way to collect time-series observations of ocean conditions. It ‘utilizes ocean wave energy to provide power for repeated ascent and descent’ of a sensor assembly (Fowler, 1997, p 1). Those who would like to ‘geek out’ on the details can read US Patent 5,644,077 (Fowler, 1997) or the plainer language description provided to the Marine Technology Society by Fowler et al (1997). I will share brief examples of other such devices throughout this book: salinometers, hydrophones, variable depth sonar systems, acoustic tracking tags, photosynthetically active radiation sensors, and the like. But while I draw attention to the presence of these gizmos and gadgets, I do not want to get carried away by the details of how they operate. Instead, I want to get carried away by the details of the tools and techniques that help us (fail to) observe such innovations.

Despite their importance, we often fail to notice new scientific instruments and techniques. Nathan Rosenberg once said that ‘the emergence and diffusion of new technologies of instrumentation … are central and neglected consequences of university basic research’ (1992, p 381). Indeed, instruments and techniques likely constitute ‘much of the “technological output” of the university system’ (Salter and Martin, 2001, p 523). Later, we will see that ‘output’ is a crass, linear simplification. But still, ‘surveys of the relationship between science and industry tend not to consider the role of instrumentation and methodologies in any detail and to discount their importance’ (Martin et al, 1996, p 22). Ammon Salter and Ben Martin have argued that this is ‘because of the limited ability of industrial R&D managers to recognize the contributions made by earlier government-funded research’ (2001, p 522). Alternatively, Bernward Joerges and Terry Shinn suggest that ‘since it [research technology] is very much a phenomenon ‘in-between’ and relatively invisible to outside observers, it is not surprising that it has gone largely unnoticed by students of science and technology’ (2001, p 11). And Peter Galison has observed that even within an instrumentation-heavy science like physics, the instruments are easily disregarded as merely ‘engine grease’ that enables the more interesting ‘experimental results and theoretical constructions’ (1997, p xvii). Whatever the rationale, scientific instruments are underestimated and understated.

However, as I have already noted, some of the most important contributions to innovation studies quietly arose from the study of scientific instruments. There might be little acknowledgement of these technologies in innovation studies, but researchers in the history and philosophy of science have given considerable thought to questions of scientific instrumentation (see de Solla Price, 1984; Galison, 1997; Joerges and Shinn, 2001; Baird, 2004; Taub, 2011; Marcacci, 2019). Perhaps the most influential of these was Derek de Solla Price. Submitted only months before his sudden death, de Solla Price’s (1984) paper on ‘the science/technology relationship’ provides a sweeping description of innovation in scientific instrumentalities. Drawing on examples from history, de Solla Price outlined the critical relationships among those who perform science and those who craft scientific techniques. For him, it was important to separate the products of science from the processes of science. The products, or outputs, of science can be described as scientific know-what. Some ancient Greek philosophers called these knowledge products episteme – that is, understandings and beliefs. They used the term techne for the know-how of science – that is, the processes or craft. In lieu of the Greek term, de Solla Price (1984, p 3) wrote about scientific ‘instrumentalities’. Here, he aimed to capture both instruments and techniques. He said that an instrumentality could be any ‘laboratory method for doing something to nature or to the data at hand’ (de Solla Price, 1984, p 13).

I would not confine instrumentalities to laboratories, but I agree with de Solla Price’s (1984) broad and pragmatic definition. I am happy to use the term ‘scientific instrumentalities’ for any materials or techniques that anyone claims to use for scientific ends. This helps me consider both the physical devices of ocean science and the less physical techniques or ‘methods’ deployed in innovation research. In a roundabout way, the word ‘instrumentality’ blurs the problematic distinction between goods and services – the tangible and tacit. More importantly, de Solla Price’s approach avoids closure around the question of ‘what is a scientific instrument?’ (Warner, 1990; Taub, 2019).

Curators of science museums have been especially concerned with this question (Warner, 1990; Taub, 2019). Deborah Warner (1990) of the Smithsonian once pointed out that the term ‘scientific instrument’ only developed in the 19th century and it has always been contested. More recently, Liba Taub has shown that ‘there has not been (and is not) always one universally agreed answer to the question “what is a scientific instrument”’ (2019, p 454). In her study of how scientific instruments have been defined over time, Taub identifies a point at which the Oxford English Dictionary began to make a classist distinction between the words ‘tool’ and ‘instrument’. The former became associated with ‘workman or artisan’ and the latter with ‘more delicate work or for artistic or scientific purpose’ (Oxford English Dictionary, cited in Taub, 2019, p 455). She argues that the label ‘instrument’ thereby came to signify professional or disciplinary status. And so she advises that museum curators ‘need to ask who defined ‘scientific instruments’, why and how?’ (p 453). Here the emphasis is on physical artefacts, but the same questions should be asked of all scientific instrumentalities.

Across the social sciences, we know that certain instrumentalities have prestige, especially standardized survey ‘instruments’. You are more likely to be considered a ‘proper’ social scientist if you use the ‘proper’ methods. Of course, these preferred instrumentalities vary by discipline. Standardized scales and randomized experiments are the norm for some fields. Elsewhere, fieldwork is preferred. In some disciplines, numbers carry all the weight. In other fields, ethnographic storytelling is the preferred way of knowing. This is how scientific instrumentalities make up part of what Karin Knorr Cetina called ‘epistemic cultures’: ‘those amalgams of arrangements and mechanisms – bonded through affinity, necessity, and historical coincidence – which, in a given field, make up how we know what we know’ (Knorr Cetina, 1999, p 1, emphasis in original). Scientific instrumentalities are cultural artefacts and cultural practices.

This idea is reasonably well studied in the physical sciences where there are a great many physical instruments. For example, Peter Galison’s history of instrumentation in microphysics can be read as a set of stories about ‘changing values and meanings as they are read into and out of the knowledge machines we call instruments’ (1997, p 63). This perspective treats scientific instrumentalities as more than methodological algorithms. It sees them as ‘encultured’ or ‘entangled’ (Galison, 1997, p 4) within a ‘complicated patchwork’ (Galison, 1997, p xx) of scientific practices. The instruments become worthy of our attention ‘if they are understood as dense with meaning, not only laden with their direct functions, but also embodying strategies of demonstration, work relationships in the laboratory, and material and symbolic connections to the outside cultures in which these machines have roots’ (Galison, 1997, p 2). In this book, I extend this understanding to the instrumentalities of innovation studies. There are fewer noticeable physical instruments in this field. So, while Galison writes about instruments as ‘material culture of a discipline’ (1997, p 2), I am writing about instrumentalities as key components in the sociomaterial construction of a discipline.

It is not that innovation studies have unique instrumentalities, like the particle colliders of high-energy physics or wave-powered profilers of oceanography. Rather, it is that disciplinary cohesion has emerged around a shared interest in public policy – which has long been tied to instruments of quantification (Alonso and Starr, 1987; Rose, 1991). And so, if there is a coherent ‘innovation studies’ discipline, then its cultural roots are planted in the numbered and neoliberalized soil of public policy. This is not to say that a straight line can be drawn between the policies of Margaret Thatcher or Ronald Reagan and present-day innovation research. It is rather more like how Cosmo Howard describes the neoliberalization of Australian and Canadian official statistics: ‘a complex and only partially coherent assemblage of calculative rationalities, technologies, and practices’ (2016, p 132). This book will confront a variety of scholarly practices that carry neoliberal assumptions – practices that neoliberalize ideas about innovation.

This means that instrumentalities are not merely technical – they also have an epistemic nature. They make knowledge and they contain knowledge. This has been a recurring argument in the philosophy of science since the time of de Solla Price. In his book on scientific instruments, Davis Baird noted that ‘a major epistemological event of the mid twentieth century has been the recognition by the scientific community of the centrality of instruments to the epistemological project of technology and science’ (2004, p 19). He argues that ‘instruments need to be understood epistemologically on par with theory’ and that ‘instruments themselves express knowledge of the world’ (2004, p 31). In so doing, he develops a way of thinking about the ‘thing knowledge’ embodied by scientific instruments. I will visit Baird’s ideas in Chapter 2. For now, it is enough to recognize that scientific instrumentalities sit within a ‘fluid relationship’ (Baird, 2004, p 32) between science (episteme) and technology (techne). In other words, scientific instrumentalities are part of what Bruno Latour called ‘technoscience’ (1987, p 29).

From this perspective, many scientific instrumentalities have ‘interstitiality’ or ‘trans-community positioning’ (Joerges and Shinn, 2001, p 7). They ‘link universities, industry, public and private research or metrology establishments, instrument-making firms, consulting companies, the military, and metrological agencies’ (Joerges and Shinn, 2001, p 3). This means that advancements in research technology can have far-reaching effects on science, industry and government (Joerges and Shinn, 2001).

As we will see in Chapter 2, the theoretical models of innovation studies have struggled to account for this. Scientific instrument innovation has been observed through linear push, market pull, chain-link, and system models. However, all these ‘reductive schemes’ (Galison, 1997, p 15) have been rejected by close studies of scientific instrumentality innovation that used post-positivist methods (see Galison, 1997, p 15; Joerges and Shinn, 2001, p 4). Galison puts it plainly when he says, ‘the dispersion of instrument knowledge follows no univocal pattern’ (1997, p xxi). For this reason, scientific instrumentalities are especially helpful empirical materials for disrupting the patterns of innovation studies.

Overview of methods and chapters

Throughout this book, I use ocean science instrumentality innovation as an empirical motif for exploring the instrumental biases in innovation research. This is a double entendre. I am writing about the need for innovation in social science instrumentalities. I am also writing about innovation in ocean science instruments. I examine archival records and analyse structured interview data from a regional concentration of ocean science and technology organizations where I live on Canada’s Atlantic coast (see Figure 1). Using a variety of post-positivist methods from outside innovation studies, I show how public organizations have developed novel technological goods, while interacting symbiotically with private companies that are ‘quartermasters’ for this scientific enterprise.

View Full Size

Map of Canada’s Atlantic coast (including major cities in the US for reference)

Source: Author’s adaptation from the Atlas of Canada (Canada with Names, 2022) under the Government of Canada’s Open Government License.

• Download Figure
• Download figure as PowerPoint slide

It will become clear that my work was originally conceived as a snapshot study – albeit one with substantial historical background material. When I presented this work as a monograph PhD thesis, I focused on the question of public innovation in goods and strayed only slightly from innovation studies norms. The ‘messiness’ (Law, 2004) was constrained. But as I have said, this book is concerned with how disciplinary norms constrain the mess. To give this work an ‘after method’ (Law, 2004) sensibility, I reframed it in the spirit of the ‘biographies of artifacts and practices’ (BOAP) approach from STS (Hyysalo et al, 2019). Sampsa Hyysalo, Neil Pollock, and Robin Williams recently argued that ‘if STS is to continue to provide insight around innovation this will require a reconceptualization of research design, to move from simple “snap shot” studies to the linking together of a string of studies’ (Hyysalo et al, 2019, p 4). I agree. Across the next seven chapters, I do as they suggest and ‘knit together different kinds of evidence – that includes historical studies, ethnographic research, qualitative studies of local, and broader development’ (Hyysalo et al, 2019, p 16). I make inquiries into my theoretical and contextual points of departure. I question different technological framings by allowing ‘ocean science instrumentalities’ to float within the categories of ‘scientific instrument’ and ‘ocean technology’. Most importantly, I play with methods, because these are the real research subjects for this book.

Law challenged us ‘to imagine what research methods might be if they were adapted to a world that included and knew itself as tide, flux, and general unpredictability’ (2004, p 7). And as we will see, this is the world of dark innovation. The pages that follow include wave after wave of concepts, theories, technologies, research areas, references, and writers. There will be calm moments, but I will also be asking a lot of readers. By the end of this book, we must have at least breached the instrumental ‘sea wall’ that has been keeping innovation studies dry.

In Chapters 2, 3 and 4, I will show how critical organizational historiography (Wadhwani and Bucheli, 2014; Durepos et al, 2021) can help reveal dark innovation. Schumpeter once argued that many of the important questions about innovation are best tackled with the tools of history rather than statistics (Godin, 2017, p 64). Over seventy years later, research on innovation remains dominated by econometrics. It tends towards the ‘don’t ask, we won’t tell’ (Gaddis, 2004, p 92) stance on historical methodology that permeates organization studies (Wadhwani and Bucheli, 2014).

First, in Chapter 2, I build on the excellent genealogies of innovation theory developed by Benoît Godin (2006, 2017, 2011, 2012; Godin and Lane, 2013). I follow those working in the historiography of management knowledge (such as Cooke, 1999; Dye et al, 2005; Kelley et al, 2006; Genoe McLaren et al, 2009) and argue that literature reviews are scholarly histories. Here, I am interested in showing how some ideas are excluded from our sense of the theoretical past. Since the extant literature is our epistemic point of departure, our research is always cognitively constrained by our sense of the past. I will show that scientific instruments are important missing characters in the stories we tell about major innovation theories. When those theories are cited and discussed today, they are black boxed in such a way that the scientific instruments are easily forgotten. In response, Chapter 2 is a history of innovation theory that centres scientific instrument innovation. I consider why scientific instruments might be missing from most stories of past innovation research and I quote personal correspondence with Eric von Hippel on how his colleagues were initially dismissive of scientific instruments as an overly ‘abnormal’ field of study. Thus begins my exploration of some of the scholarly processes that produce dark innovation, and some of the instrumentalities that can uncover hidden or marginalized subject matter.

In Chapter 3, I problematize the taken-for-granted nature of ‘context’ and demonstrate an alternative way of ‘practicing context’ (McLaren and Durepos, 2019, p 74). It is quite normal for research in innovation studies to include some discussion about the history of a regional-industrial context before engaging with primary data. But readers are typically asked to take the authors’ expert knowledge of the historical context for granted. Instead, I use an ANTi-History approach (Durepos and Mills, 2012) to (re)assemble three histories of one ocean science and technology sector in Nova Scotia, Canada. I present three incompatible newspaper and magazine accounts of this sectors’ emergence – from 1960, 1980, and 2012. The earliest account of the sectoral history positions scientists, scientific instruments, science organizations, and geopolitics as key actors. But in the latest account, scientific instruments are not present; the main actors are private companies and science is lauded as the knowledge base that spawned these companies. I argue that these three different histories are traces of efforts to define a sector/cluster/industry identity and to rhetorically impose that identity on various actors. I argue that these are ‘rhetorical histories’ (Suddaby et al, 2010) that aim to ‘assemble’ a cluster as historical fact, thereby establishing a regional competitive advantage. By treating the industrial history in this way, I demonstrate the need to take historical method (that is, historiography) seriously in research on innovation.

In Chapter 4, I consider the narrative tools that shape our understandings of innovation. The apolitical treatment of science and technology (Fagerberg et al, 2012b; Martin et al, 2012) – including the apolitical treatment of the state (Pfotenhauer and Juhl, 2017) – has already been flagged as an issue in innovation studies. I advance those discussions by directly surfacing the political nature of ocean science and technology in Nova Scotia. I argue that when we turn to the past from a present-day neoliberal standpoint, we always end up writing stories about market-dominant evolutionary processes. In contrast, this chapter presents the stories of three public research organizations (PROs) and the politics around their establishment. In these organizational settings, I find private industry becoming enrolled in political missions of military defence, Canadian sovereignty, and scientific one-upmanship. Analysing these short histories as a set, I find that the public organizations can be characterized as active political agents. Meanwhile, the private companies around them can be characterized as ‘quartermasters’ – like the individuals responsible for providing supplies to units in an army (or the Q Branch in James Bond). They were producing the scientific instrumentalities needed for multiple ‘cold wars’. This relationship was more nuanced than the simple provisioning of equipment and services – it was often a close two-way partnership. The technical expertise provided by scientific instrument companies helps to set the course for science, and vice versa. Telling the past in this way makes the boundary between public and private organizations messier than it appears in neoliberal ideology.

At the halfway point of this book, I shift focus from historiographic methods to modern day acts of counting and classification. First, in Chapter 5, I explore the puzzle of innovation taxonomies. Anecdotally, I consider how strangely normal it is for people in Atlantic Canada to speak of an ‘ocean technology sector’. Does these mean there are only two other types of technology: land and aerospace? Ocean technology is very clearly a ‘folk taxonomy’. It is akin to the way in which many people speak of spiders as ‘bugs’. This makes it a useful category for demonstrating that sectoral boundaries are not as ‘natural’ as innovation research assumes.

Product-based industrial classifications (Standard Industrial Classification, North American Industrial Classification System, etc.) are embedded within innovation theory through the methods that were used by Keith Pavitt (1984) and his heirs (Archibugi, 2001; Castellacci, 2008). The resulting ‘taxonomies’ of innovation have been widely used in research and public policy (de Jong and Marsili, 2006). I argue that all this activity is driven by one of the many biological metaphors that permeate innovation studies. The idea of taxonomic classification is one of those that we have explicitly borrowed from biology (see Archibugi, 2001; de Jong and Marsili, 2006). I use analogies from taxonomic biology to reconsider three major methodological problems already described in the innovation taxonomies literature. This includes, but is not limited to, the taxonomic separation of public and private organizations. I then turn to the deeper problems that arise from the implicit use of an organizations-as-organisms analogy. Following the work of Gareth Morgan (1980, 1986), I consider the metatheoretical implications of biological metaphors and the corresponding political assumptions that are inscribed in sectoral classification tools. I argue that we will need other metaphors if we hope to observe other innovation patterns.

I take the question of metaphors one step further in Chapter 6. There, I approach ‘region’ as a spatial metaphor that both enables and constrains the systems of innovation literature. Thinking about innovation as a regional phenomenon allows for the surveyable, measurable, Euclidean spaces we call innovation systems. But this makes it difficult to observe innovation processes that fold people, things, and places together in new ways. Taking ‘object lessons’ from John Law and Vickey Singleton (2005), I explore the different possibilities that come from framing innovation as a region, network, fire, or fluid object. First, I unpack the regional metaphor by describing the boundary choices I made while preparing a survey of ocean science instrumentality innovation. Then, I share ‘excess’ network data that extended beyond my region and other observations that would have ‘flooded’ or ‘set fire’ to the boundaries I had established. On these grounds, I question the ‘hegemony’ of regionalism (Sepp, 2012, p 47) within innovation studies. Many innovation scholars would say that we need the regional metaphor so we can pin things down, survey them, and quantify them – otherwise they do not count. But I suggest a turn towards critical geography, where other topological metaphors provide ways ‘of understanding space and time when the numbers no longer quite add up to anything significant’ (Allen, 2011a, p 316).

In Chapter 7, I dig deeper into the idea that good numbers can be meaningless. We know that there are problems embedded in the construction of standardized innovation measures (see Godin, 2002, 2005; Gault, 2018; Gault, 2020). To address some other problems of statistical practice, I develop and deploy autoethnostatistics. This is a fusion of autoethnography (Ellis, 2004) and ethnostatistics (Gephart, 1988, 1997, 2006). Autoethnography is cultural research that uses narrative inquiry into one’s personal experiences (see Ellis, 2004; Prasad, 2019). It has seen limited use in innovation research (for example, Rehn, 2023), but it is a well-known approach elsewhere. Meanwhile, ethnostatistics is ‘the empirical study of how professional scholars construct and use statistics and numerals in scholarly research’ (Gephart, 2006, p 417). It is generally underutilized (Gephart, 2006; Helms Mills et al, 2006) and has also seen limited use in innovation research (see Kilduff and Oh, 2006). Rather than producing a distanced critique of how other people do statistics, I combine these methods to examine my own experiences inside the cult of numbers that dominates mainstream innovation studies.

My work in Chapter 7 might challenge both critical and mainstream scholars. Following the norms for reporting statistical work, I present four analyses of my data on ocean science instrumentality innovation in Nova Scotia. From the outset, simple descriptive statistics provide evidence to falsify any claim against the existence of public innovation in goods. But these numbers also lack any measure of statistical significance. And so, I follow the descriptive statistics with three sets of statistically significant results. These results follow the canonical progression of innovation theory discussed (and debunked) in Chapter 2. After each analysis, I break into an autoethnographic discussion on the meaning(lessness) of those results. Using Albert Camus’ The Myth of Sisyphus (1955), I build directly upon autoethnography’s ‘existential convictions’ (Bochner, 2012, p 209). I suggest that following convention is like being condemned by the statistical gods to push numbers up a hill, hoping to successfully reach the summit, only to realize that the effort was meaningless and alienating.

In the final chapter, some readers might expect me to lay out a research agenda for finding dark innovation. I do suggest some opportunities arising from this book. But my focus is on how we might further deconstruct the instrumentalities of innovation research. To this end, I deconstruct some of my own instrumentalities. I finally notice the most influential oceans-related innovation to ever emerge from my part of the world and I consider how the ‘epistemic culture’ of innovation studies pointed me and my tools away from it. Returning to the dark matter analogy, and with insights from both Karin Knorr Cetina (1999) and Karen Barad (2007), I explain why some physicists might think their observations are meaningless and why this leads them to obsess over their instrumentalities. I argue that we need a similar humility and obsession in research on innovation.