Why is science and technology inevitable

Uncertainty as a Challenge for Science

Peter Wehling (Frankfurt)

Responsibility for the inevitable. Scientific ignorance as an object of epistemic self-reflection and political shaping

Abstract: In this chapter, it is argued that, although scientific ignorance (or non-knowledge) is an inevitable implication and corollary of scientific knowledge production, the sciences as well as science politics must nevertheless take responsibility for this ignorance and its potential adverse effects . It is shown that, on closer inspection, "science-based ignorance" is both unavoidable and avoidable, without there being any clear-cut and impermeable "boundary" between these two forms of ignorance. By contrast, as an exemplary analysis of epistemic practices and different epistemic cultures illustrates, there is considerable scope of action for the sciences to become aware of and potentially reduce their self-produced ignorance by reflexively scrutinizing their own background assumptions, standard methods and research routines . This should, however, not lead to adopting the flawed ideal of science being able to produce knowledge without simultaneously generating ignorance, albeit to varying degrees and in different forms. Instead, it is essential to recognize and openly communicate the fact that the co-production of scientific knowledge and ignorance is inescapable while, on the other hand, science politics has to create favorable conditions that foster self-reflexive epistemic practices which aim at potentially avoiding harmful knowledge gaps and blind spots. In the end, however, the question of how to deal with scientific ignorance is a political one; Therefore, stopping large-scale technoscientific experiments such as climate engineering ‘simply’ due to the amount of unknowns (known and unknown) they are likely to produce must imperatively be recognized as a legitimate option.

Keywords: knowledge - ignorance - science - epistemic cultures - epistemic practices - responsibility - scope for action

1 Introduction: The "inseparability" of knowledge and ignorance

In his pioneering study on the “origin and development of a scientific fact” in 1935, Ludwik Fleck formulated a far-reaching scientific-theoretical and sociological insight, the explosive nature of which, however, was hardly noticed for a long time. “To recognize a relationship”, wrote Fleck (1993: 44), “one has to misjudge some other relationship, to lose sight of it 208 → deny, overlook ”. In this way, "the discovery is inseparably intertwined with the so-called error" (ibid.). A more recent formulation of the same facts by Martin Seel (2009: 42) shows that instead of “so-called error” one can just as easily speak of uncertainty, indeterminacy and ignorance: “[W] hen to know something means to know something specific, so it also means leaving other things in the indefinite. ”It is remarkable that Fleck and Seel do not speak here of correctable errors of 'bad' science, which has failed to make their efforts to know comprehensively enough, but of epistemological necessities: in order to recognize something According to Fleck, one has to misjudge and overlook other things; this is precisely why “discovery” and “error”, knowledge and ignorance, are inseparably intertwined. This ignorance turns out to be essentially self-generated, although mostly unknowingly and unintentionally brought about ignorance of science.

Building on such considerations, I would like to justify and explain the initially paradoxical thesis that scientific uncertainty and scientific ignorance on the one hand are 'normal' and inevitable, but that on the other hand science (as well as science policy) is responsible for this ignorance (and its possible consequences ) nevertheless bears and must take responsibility.1 The popular maxim “ignorance does not protect against punishment” already says that ignorance does not per se relieve responsibility for one's own actions and their consequences. As a rule, however, this statement is followed by the restrictive - often extremely complicated - question of whether the persons concerned could have known better at all, whether their ignorance was avoidable or not. Because moral, political or legal responsibility for the unforeseen consequences of their actions or omissions is usually only assigned to social actors if they could have acquired the relevant knowledge.2← 208 | 209 → The suggestion to show oneself responsible for unavoidable ignorance and its consequences may therefore initially appear paradoxical or even absurd, since one usually neither has to nor can assume responsibility for an event that is unavoidable, i.e. outside one's own sphere of influence . Scientific ignorance, however, is both inevitable and avoidable in a very specific way that needs to be clarified more precisely, and for this reason the requirement to take responsibility for unavoidable ignorance proves to be well founded and justifiable.

In the following I would like to explain in more detail to what extent scientific ignorance is both avoidable and inevitable and for what reasons science is not simply released from responsibility for possible undesirable consequences of this ignorance (Chapter 2). Following on from this, I will outline by way of example how science, through its cognitive practices, produces knowledge and ignorance at the same time and to what extent there are nevertheless possibilities for a reflective handling of this problem (Chapter 3). In conclusion, I would like to make it clear that the (in) avoidable co-production of knowledge and ignorance requires both increased self-reflection on the part of science and new (scientific) political frameworks and social scope for action (Chapter 4).

2 The (in) avoidability of scientific ignorance

Already in early epistemological and sociological observations of the so-called ecological crisis, i.e. the increasing, scientifically and technologically generated natural and self-endangerment of modern societies, it was occasionally pointed out that the production and technical application of scientific knowledge leads to ignorance at the same time.3 The science researcher Jerry Ravetz, for example, already stated around 30 years ago that the self-generated “science-based ignorance” (science- ← 209 | 210 → based ignorance) (Ravetz 1990: 1) was even increasing “even more rapidly” (Ravetz 1986: 423) than knowledge. Nevertheless, the reasons for this phenomenon of “co-production” and the mutual increase in knowledge and ignorance have seldom been examined in more detail since then. In such an analysis, two possible approaches can be distinguished:4 On the one hand, one can try to establish the reciprocal constitution of knowledge and ignorance on the horizon of a general theory of knowledge; on the other hand, one can try to reconstruct how knowledge and ignorance are simultaneously generated in concrete, historically and socially situated epistemic practices. The first approach has the advantage of being able to identify the simultaneity of knowledge and non-knowledge production as constitutive for any knowledge and thus fundamentally remove the ground from questionable ideas of secure and complete knowledge without any 'contamination' by ignorance.5 However, such a perspective remains indifferent to the diversity and heterogeneity of epistemic practices in the sciences. Due to its level of abstraction, it tends to assume a static and deterministic relationship between knowledge and ignorance, as if every gain in knowledge were automatically reflected in a complementary increase in ignorance. What is neglected is the different ways in which ignorance is generated, perceived and interpreted within the sciences, as well as the possibilities of reacting reflexively to this problem. Therefore, on the abstraction level of a general theory of “knowledge”, the simultaneous production of ignorance must appear not only to be fundamentally unavoidable, but also ultimately to be unaffected. In this theoretical framework it therefore makes little sense to assign responsibility to science for its ignorance, and every attempt by the sciences to influence the generation of ignorance by reflecting on and modifying their cognitive practices must from the outset be considered futile 211 →

On the other hand, based on the considerations already mentioned by Martin Seels (2009), I would like to explain why it is necessary for empirically and politically oriented science research to examine the simultaneity of knowledge and ignorance in a concrete and differentiated manner in the context of specific epistemic practices. Seel's considerations, like Luhmann's, begin at the level of a general theory of knowledge. Accordingly, everything definite necessarily has a “downside of the indefinite”, because “[we] we cannot think at all, let alone reach definition without restriction” (Seel 2009: 44). Every act of cognition and determination thus includes a complementary indeterminacy. As Seel rightly points out, this is not a deficit in cognition that can be remedied by more comprehensive, more precise observation. Because who

Has any knowing access to the world at all, has limited access to the objects of his knowledge, otherwise he would have no access at all. For knowing beings, the world is determined and indefinite at the same time (Seel 2009: 44).

In Seel (2009: 47), these general considerations lead to the idea of ​​a “constitutive non-knowledge”, which consists in a “horizon of indeterminacy that is connected with all conceptual knowledge but cannot be overlooked by those who know”. Seel (ibid.) Distinguishes from this, however, a “contingent ignorance” that arose by chance or was brought about through one's own fault, emphasizing that the “boundary” between these two forms of ignorance remains “often enough vague”. This can be pointed to a more far-reaching thesis and made more precise: There is no predetermined, objective and unbridgeable difference between the constitutive ignorance, which lies in the 'nature' of knowing, on the one hand, and the contingent ignorance caused by external, situational, and not least social circumstances, on the other. Rather, the ignorance that is 'inseparable' from cognition and co-generated in the cognitive process is both constitutive (inevitable) and contingent (that is, at least potentially avoidable). It is the respective situational conditions and practices of cognition, the specific “means of cognition” (Seel 2009), which bring about and manifest constitutive ignorance in a contingent manner. Even if knowledge cannot be achieved without restriction, without perspective, the following still applies: what the restriction consists of and which perspective is taken depends on the respective situational circumstances. Contingency means more and different than just coincidence and individual “fault”; rather, the given social, cultural and technical prerequisites of the ← 211 | are contingent 212 → Production of knowledge that makes knowledge possible in the first place, but at the same time also limits it. Establish the respective established epistemic practices

represent not only conditions of the possibility of knowledge, but also limits of knowledge. They act as filters of the will to know and the ability to know, in that they “prepare” the attention, the interest in knowledge and the intentionality of the actors […] in research practice (Sandkühler 2009: 169; emphasis in the original).6

The question of the avoidability of and responsibility for scientific ignorance can thus be posed in a more differentiated way: although indeterminacy and ignorance form the inevitable, constitutive condition for the generation of knowledge, it is at the same time dependent on contingent, changeable factors such as well-established research routines and methodical standards or theoretical background assumptions about what is known and what is not known in a given situation. In other words, ignorance of certain events or contexts does not have to be unavoidable in principle, but can under certain circumstances be opened up by reflecting on the blind spots of the cognitive practices involved - even if often 'only' as a possibility, as a hypothetical 'space' of potential, unknown events , not as certain knowledge of what one does not know. Against this background, the expectation should not be hastily dismissed as naive that science can and should deal self-reflexively and responsibly with its self-generated ignorance and the indissoluble connection between knowledge generation on the one hand and the generation of ignorance on the other. How and to what extent this is possible becomes more clearly visible as soon as one examines in more detail how epistemic practices produce not only knowledge but also ignorance at the same time.

3 Epistemic Practices: the co-production of knowledge and ignorance

In order to adequately grasp the possibilities and limits of a self-reflective handling of science with its self-generated ignorance, it is not enough, as we have seen, to remain on the level of a general theory of knowledge and the constitutive, inevitable ignorance. ← 212 | is required 213 → rather to analyze how the constitutive ignorance in the historically and socially situated process of cognition is brought about as contingent ignorance and to what extent the latter is perceived, processed and communicated or remains latent, unrecognized and untapped. The investigation of these questions can be approached in two steps: First, fundamental elements and dimensions of scientific knowledge can be highlighted, which contribute both to the acquisition of knowledge and to the generation of ignorance. At this level of analysis, four general dimensions of scientific knowledge production are particularly relevant: (a) the perspectivity and selectivity of theories, thought models, concepts and metaphors; (b) the isolation and decontextualization of the objects of knowledge in the laboratory or in laboratory-like research settings; (c) the constitution of new, non-anticipated horizons of action through the review and application of knowledge and knowledge-based technical artifacts outside the research context as well as (d) the choice of research questions, priorities and goals that are not only influenced by "external" factors such as available Research funding or expected market opportunities is influenced, but also by internal scientific selection filters such as the (often only seemingly) unproblematic processing with the established research methods.7

In a subsequent, second analysis step, specific epistemic practices can be examined to determine how they produce ignorance on the one hand, but also enable a reflected approach to it on the other. From this perspective, the following six aspects of research practice can be identified, without claiming to be exhaustive, which appear to be particularly informative for an analysis of the connection between knowledge and ignorance:

(1) the spatial and temporal horizons of knowledge generation;

(2) responses to surprises and unexpected results;

(3) Type and extent of the de- and recontextualization of the epistemic objects;

(4) dealing with complexity; ← 213 | 214 →

(5) the explicit perception, processing and communication of ignorance;

(6) the inter- and transdisciplinary openness of a research field or an epistemic culture.8

As already shown by Ludwik Fleck's (1993) analyzes of different “thinking styles” and “thinking collectives” in science, and as confirmed and deepened by later work, especially by Karin Knorr-Cetina (2002) on “epistemic cultures”, the individual scientists shape and Scientists or research groups do not each have completely new and individualized these dimensions of their cognitive practice. Rather, they are more or less oriented towards implicit or explicit background assumptions, standards and routines of their respective disciplines and research areas. When analyzing the six aspects of research practice and their different forms, the respective epistemic cultures or styles of thinking are therefore always taken into account.It shows that epistemic cultures (or cultures of knowledge) are at the same time cultures of knowledge and ignorance, because with knowledge they always produce ignorance in a specific way and perceive, reflect, process and communicate - or do not perceive, reflect, process or communicate.9

(1) A first thematically relevant aspect of scientific knowledge practices are the spatial and temporal (perception) horizons on which the practices of knowledge generation and validation are based or constituted by them: How long and in which spatial sections, in which time intervals and on which specific places must be observed in order to arrive at results and statements that can be considered certain? How long would you have to medically observe users of cell phones (and a control group of “cell phone refusers” who would be hard to find these days) in order to be able to make statements about them ← 214 | 215 → to be able to determine whether mobile telephoning involves health risks?10 And how big would the two groups have to be? Over what period of time and in what spatial area around a so-called iron fertilization experiment in the ocean, measurements and observations would have to be made in order to be able to say whether the supply of the metal has problematic effects on maritime ecosystems and food chains - or whether this is not the case? How often would such an experiment have to be repeated? Obviously, pragmatic research conditions and constraints such as the availability of time and financial resources, of test subjects, control groups and observation instruments also play an important role in the corresponding specifications. At the same time, the choice of spatial and temporal observation horizons is always an expression of the routines, standards and assumptions of normality that have been established in a research field, in an epistemic culture, and which are not checked and adjusted every time, but rather passed on and extrapolated.11 Last but not least, partly implicit, partly explicit expectations with regard to the presumed effects flow into the determination of the examination horizons, e.g. in medical (risk) research more or less reliable assumptions about possible side effects and their latency and incubation periods (see footnote 10).

With the respective chosen temporal and / or spatial observation and expectation horizons on which the generation of knowledge is based, ignorance is inevitably produced at the same time: Everything that happens or could happen 'outside' these horizons is in fact excluded as irrelevant, it remains unobserved and unrecognized or, at best, perceived by chance. At first glance, this may seem almost trivial, ← 215 | 216 → nevertheless represents one of the most important dimensions of the co-production of knowledge and ignorance. Because the temporal and spatial limitation of the research horizons is in a very elementary way constitutive for any gain in knowledge in the sense outlined by Seel (2009), since one has already started pragmatic, but also for epistemological reasons, not everything, not unlimited in time and cannot be observed everywhere. Where exactly the limits of observation are drawn is, however, contingent and largely dependent on the established research routines and assumptions as well as on the available resources. Against this background, epistemic practices and cultures differ not only in how narrow or wide they set their spatial and temporal observation horizons, but especially in the extent to which they remain aware of the contingency and selectivity of the corresponding determinations and to what extent they reflect the reasons for this, and if necessary modify case-specifically.

(2) Dealing with surprises and unexpected test results is a second aspect of scientific knowledge practice, which is of decisive importance for the interplay between knowledge and ignorance.12 In many research areas and epistemic cultures, unforeseen results are ultimately perceived as disturbances in the 'actual' knowledge gain, which should be pragmatically eliminated, for example by varying the experimental setup and the boundary conditions, without systematically investigating the reasons for the surprise. But surprises can also be understood as an important source of knowledge in the sense of the search for “liminal knowledge” and ← 216 | 217 → be actively sought in order to check one's own methodological approaches in a targeted manner and to expand the underlying theoretical perspectives. In this respect, the “persistence” of knowledge systems diagnosed by Fleck (1993: 40 ff.), That is, their tendency to either reinterpret “unsuitable” results, to marginalize them as “outliers” or to silently ignore them, is entirely variable within certain limits. The availability of time and financial resources also plays an essential role in this context in determining which position is and can be adopted (see Kastenhofer 2015: 99 ff.). Above all, however, the diversity of the respective cognitive goals is important: If the research is aimed at the most comprehensive possible understanding of a complex and possibly singular phenomenon, one will usually endeavor to include all conceivable and observable influencing factors - albeit never including them 'Complete' picture will emerge. If, on the other hand, the goal is to uncover, isolate, experimentally stabilize and finally technically reproduce a certain causal relationship for the development of a new drug, it makes sense to design the framework conditions in such a way that possible 'disruptive factors' are neutralized without to have to and want to explore them in detail.

(3) Closely linked to this is a third aspect of scientific knowledge practice that is relevant to the co-production of knowledge and ignorance: the type and degree of decontextualization and possible recontextualization of the research objects. In what way and to what extent the research objects are removed from their respective spatial, temporal and material contexts, and to what extent an attempt is made to `` embed '' them again later in these references in order to make statements about their behavior and their effects in their environments outside of the research context to be able to meet? The decontextualization of the objects of knowledge is a fundamental knowledge strategy of many research areas and epistemic cultures, by no means only in the natural sciences. The systematic neutralization of random, singular, merely local or temporary influences and environmental references should enable generalizable and reproducible results (cf. Bonß et al. 1993b: 181). It is then to be expected, however, that in the process of knowledge generation, “fade-out losses” (Bonß et al. 1993a: 60) occur which can prove to be problematic if the knowledge gained in the experimental 'clean room' is related to the specific relationships outside the laboratory or if artifacts generated in the laboratory are 'released' ← 217 | 218 → (cf. Tetens 2006).13 Epistemic practices and cultures can be differentiated according to the extent to which and with what means they try to “keep such losses as low as possible” (Bonß et al. 1993a: 60) or to compensate for them through forms of recontextualization, i.e. the systematic inclusion of the potential application contexts. As Bonß et al. (1993b: 185) rightly point out that the idea of ​​a complete recontextualization would be naive. Nevertheless, attempts to specifically estimate the “fade-out losses” with regard to possible fields of application can at least result in a “more precise awareness of uncertainty” (Bonß et al. 1993a: 64 f.) And a heightened awareness of the blind spots created by decontextualization.

(4) A fourth important aspect lies in the different ways in which epistemic practices and cultures deal with the complexity of the areas of knowledge. A high level of complexity awareness not only takes into account the incalculable variety of influencing factors and the density of interactions in the examined subject area, but also takes into account other effects that can be associated with complexity. Above all, this includes the fact that minimal, hardly recognizable and possibly not even measurable variations in the initial states of a phenomenon can produce considerable, hardly foreseeable differences in the further course of the event.14 Individual events cannot therefore be adequately anticipated in the “realm of complexity” (Küppers 2009: 141) even if one knows the general regularities of their occurrence. Küppers rightly notes that “in the complex” knowledge and ignorance exist at the same time and side by side: “Although you know everything, you know nothing.” (Küppers 2009: 141) ← 218 | 219 →

In this situation it is therefore crucial to expect the latency and surprising emergence of phenomena as well as widely distributed and / or extremely delayed effects (cf. Ewald 1998). In this regard, cognitive practices and epistemic cultures can be differentiated according to the extent to which they rely on theoretically determined principles and (supposedly) clear and unambiguous empirical findings or perceive them with the awareness that `` below '' or `` behind '' there are unexpected and (still) ) unrecognizable, as could hide effects that have not yet occurred or become manifest. So the question here is not how epistemic cultures deal with manifest surprises; What is of interest is rather what conclusions they draw, as it were, from the absence of such surprises: Does this mean that everything is 'normal' and predictable? Or can it not be ruled out and must therefore be taken into account that unexpected, not yet visible effects could occur and the knowledge gained must therefore be regarded as provisional, incomplete and highly uncertain?

(5) As the problem of complexity makes clear, the different forms of explicit perception and communication of ignorance and the limits of knowledge play a decisive role: which definitions and interpretations of the unknown are in the foreground in different epistemic cultures, are worked on and communicates?15 In order to illustrate the range of possibilities, there are three axes of distinction between ignorance, which I have presented in detail elsewhere (cf. Wehling 2006: 116 ff.): A) the knowledge of ignorance, b) the intentionality of ignorance and c ) its temporality or permanence.

a) Not knowing can first be differentiated under the aspect of whether and to what extent the acting actors know what they do not know - or whether this is also beyond their knowledge. In the former case of the so-called known unknowns, specific questions can be asked and research designs can be drawn up in order to close the knowledge gaps. In the case of unknown or unrecognized ignorance (the unknown unknowns), on the other hand, what they do not know remains hidden from the actors as well as the fact that they do not know something potentially important. Consequently, it is usually also unclear how, when and where one could acquire the knowledge that may be missing; therefore often no need is seen to be, ← 219 | 220 → to research unknown phenomena that may not exist at all (cf. Heidbrink 2003: 29). This problem becomes both scientifically and politically explosive in situations in which there are no empirical clues for a certain event or a certain causal relationship, for example for health risks from mobile phone use. Nevertheless, it remains open whether one knows that such side effects do not exist, or whether the missing evidence may only mean that one has previously searched 'in the wrong place' or that the search has stopped too early. It is inevitably dependent on interpretation whether we have reliable knowledge in such constellations of “negative evidence” (Walton 1996: 140) (“using a cell phone is safe”) or whether we are ignorant and clueless because important clues have escaped our attention (cf. . Walton 1996: 140).

b) Ignorance can also be differentiated according to the extent to which it can be attributed to the action or omission of social actors or is unavoidable, i.e. would have been indissoluble even if all available sources of knowledge had been developed and used. The example of the fatal side effects of the thalidomide sleeping pill makes this distinction clear: Were the severe damage to human fetuses when thalidomide was completely unpredictable when thalidomide was launched on the market and their ignorance was therefore inevitable? Or should the manufacturer of the product have discovered them beforehand with the help of more extensive and more careful tests? As can be seen here, the intentionality of ignorance involves considerably more than 'just' the conscious and deliberate refusal of actors to experience something or to take note of it. Inadequate efforts to know, negligence or limited interest in knowledge can also be considered as socially attributable reasons for not knowing, without necessarily having to be based on an express not wanting to know. It is therefore not surprising that (as in the Contergan case) it is regularly highly controversial to what extent certain actors can be assigned legal, moral or political responsibility for ignorance and its consequences.

c) Finally, ignorance can be differentiated according to its temporal permanence: Is it just a temporary not-yet-knowledge that will soon be replaced by knowledge, or is it a long-lasting, possibly even completely insurmountable non-knowledge -Can do? Even this distinction does not bring any objective characteristics of objects of knowledge or not 221 → knowledge for expression; the ascription of 'knowable' or (in principle) 'non-knowable' is rather made by social actors in a contingent and often extremely controversial manner: Will we ever be able to know for sure (and if so, when) whether attempts are made to reduce the Earth's radiation budget technically influencing (so-called solar radiation management), the hoped-for effects and which undesirable side effects are to be expected?

Against this background, epistemic cultures differ mainly according to whether they perceive and communicate their own ignorance primarily as limited, specified and temporary not-yet-knowledge or whether they allow and take into account the possibility of unrecognized and insurmountable ignorance. The "temporalization" of ignorance (Bauman 1992: 295) to a mere transitional stage on the way to certain knowledge was and is undoubtedly the dominant pattern of perception in modern societies and modern science, which, however, is no longer entirely undisputed and unchallenged (cf. . Beck / Wehling 2012). Relevant differences between epistemic cultures can also be shown in the extent to which they attribute their knowledge gaps as 'unavoidable' to the lack of transparency in the research subjects or understand them as an at least potentially and partially avoidable effect of their own research practices and theoretical assumptions.

(6) As Fleck (1993: 53) made clear, styles of thinking or epistemic cultures are always characterized by a constitutive “harmony of illusions” due to their “tendency to persist”, which they cannot resolve on their own. Therefore, a sixth relevant aspect of epistemic practices is the extent to which they are inter- or transdisciplinary receptive and capable of corrective influences from 'outside'. I have already pointed out above that the empirical knowledge of non-scientific actors often represents an important corrective in order to uncover blind spots of certain scientific perspectives or even to close knowledge gaps (see Frickel et al. 2010 as well as with a view to medicine Wehling et al . 2015). Such a function can also be taken over by other scientific cultures of knowledge or research areas, although it is not immediately possible to 'incorporate' questions, hypotheses or findings from one scientific discipline into another. Nevertheless, according to Sandkühler (2009: 69), the internal homogeneity and mutual incompatibility of epistemic cultures should not be overestimated; Fleck (1993: 142 ff.) also explicitly referred to the possibility of “intercollective thought communication”, that is, contact ← 221 | 222 → and exchanges between different styles of thinking.He saw its "most important epistemological significance" in the reorganization and change of a given style of thinking, which would open up "new possibilities for discovery" and create "new facts" (Fleck 1993: 144). The history of science knows numerous examples of the often very productive interaction between different disciplines and areas of science, for example in the form of theory transfers or method imports. Such effects should be expected above all when different research directions operate in interdisciplinary or transdisciplinary structured fields and are confronted with competing perspectives or contrasting findings, as is the case, for example, in risk research on large-scale climate engineering (cf. Szerszynski / Galarraga 2013). It is precisely in such contexts that epistemic practices and cultures can diverge in terms of the degree to which and in what way they allow themselves to be irritated by other areas of knowledge and use their questions or results to check one's own assumptions, routines and established perceptual horizons.

4 Responsibility for the inevitable: room for maneuver in science, politics and society

A look at these cognitive practices and their respective manifestations in different epistemic cultures reveals two things: On the one hand, the spatial-temporal focus of the observation, the de- and recontextualization of the research objects, the respective handling of surprises and the reduction of complexity do not only knowledge, but inevitably also ignorance. On the other hand, it shows that the form and the extent to which this happens can be influenced: The epistemic practices outlined offer more or less great leeway to react reflexively to the co-production of knowledge and ignorance. However, this cannot and must not tempt one to reintroduce the false, idealizing idea through the back door, as it were, that science can completely avoid the generation of ignorance through a combination of optimal research practices. Whichever cognitive strategies are chosen, it can never be completely prevented that in and through the process of knowledge production, even if in variable forms and degrees, indeterminacies, gaps in knowledge and blind spots are produced at the same time. ← ​​222 | 223 →

It is therefore imperative, not only for science research, but also for science policy to emphasize the inevitability of the co-production of knowledge and ignorance in all clarity and to communicate it publicly (cf. Douglas 2015; Nielsen / Sørensen 2017). Otherwise, problematic social expectations of a supposedly 'omniscient' science as well as unfounded claims to authority and questionable fantasies of omnipotence could establish themselves in science itself or - where they already exist - solidify further. To emphasize this again, the simultaneity of knowledge and ignorance, of certainty and indeterminacy, is not a mark of 'bad' research - or, as is sometimes suggested, only characteristic of an 'antediluvian', inadequately equipped science, as described in operated in the past. The interplay of knowledge and ignorance cannot be resolved by even more complex and comprehensive methods of measurement, data collection and evaluation (big data, etc.), but rather - from the methods described in Chap. 2 explained reasons - constitutive for any knowledge production and therefore unavoidable. And yet this cannot be a reason to release science (and science policy) from its responsibility for this ignorance and its sometimes catastrophic consequences (e.g. in cases such as CFC, Contergan, DDT etc.).

The postulate of responsibility for the inevitable can be based on two important insights:

a) Firstly, it is by no means certain that all knowledge gaps and blind spots that may initially appear to be inevitable were actually inevitable. Under certain circumstances, ignorance could have been avoided or at least reduced if there was a greater reflexive distance to the methods, standards and routines that have been established in one's own research field and which are taken for granted, as well as a self-critical review of one's own preliminary and background assumptions (cf.Kirk 1999 on the Contergan case) . In his considerations on the responsibility of unintended and unanticipated consequences of actions, Ludger Heidbrink argued restrictively that additional, risk-reducing search for knowledge must not only be possible for the actors concerned, but also reasonable in order to be able to legitimately assign responsibility for their ignorance to them (cf. Heidbrink 2013: 126 f.). As a rule, however, there are no clear and objectifiable criteria for the reasonableness of knowledge efforts; In any case, shortage of time, high costs or possible competitive disadvantages are not factors which make such activities per se unreasonable for business enterprises, for example ← 223 | 224 → or research groups would. Because this would mean that it is legitimate to achieve economic profit or advantages in scientific competition by externalizing the risks associated with (possibly avoidable) ignorance.16

With a view to science, it becomes clear in this context that the question of the reasonableness of a reflexive cognitive practice that takes into account the possibility of self-generated and unrecognized ignorance is also a question of the political design of goals and framework conditions of research: As long as the task of Science is seen primarily or exclusively in producing as much new and (according to current standards) as possible knowledge as quickly as possible, and as long as this goal is rewarded with research funding or academic reputation, the demand becomes responsible and self-reflective with what is self-generated Dealing with one's own ignorance does indeed appear to the scientific actors as an unreasonable waste of time. Science policy must react to this with a profound restructuring of research programs, institutional framework conditions and internal scientific reward mechanisms in order to create the temporal, financial and intellectual leeway and leeway that enable a self-reflective design of the co-production of knowledge and ignorance. Science policy must free itself from the questionable modernist expectation that research should provide more and more technologically usable knowledge and economically usable innovations. Rather, what is required is an understanding of the social role and implications of science, which is reflected in a 'societal risk', which obliges it not to concentrate exclusively on the production of positive knowledge, but also on the same level as its dark side, the self-generated ignorance and the potential associated with it dealing with social, technological and ecological risks (cf. Jaeger / Scheringer 2009). For it is no longer primarily the (often only supposed) superiority of scientific knowledge over other forms of knowledge, but rather the "uncertainty of scientific knowledge itself" (Ewald 1998: 20) that is a characteristic of the current social situation 225 →