Approaching Distributed Epistemic Responsibility
There are various research areas that have provided invaluable insights to crucial aspects of being responsible in knowing within entangled socio-technical epistemic systems. To open up this topic, I will in the following sections briefly introduce crucial insights from three different fields of research: research on epistemic responsibility in (social) epistemology, research on (distributed) moral responsibility in philosophy of computing, and research on distributed or entangled responsibility in feminist theory.
4.1 Epistemic Responsibility: Insights from (Social) Epistemology
Epistemic responsibility can be understood in terms of the duties of knowers in giving and accepting reasons. Within analytic epistemology, for instance, it is discussed whether and to what extent epistemic responsibility is a condition for epistemic justification and knowledge. Some theoreticians focus on very basic questions concerning our duties to revise beliefs in light of new evidence, fundamentally related to the topic of doxastic voluntarism, i.e. the question whether we can voluntary control our beliefs. Others address the question of what being a good informant implies (Craig 1990), focus on concepts of epistemic praiseworthiness and blameworthiness and relate epistemic responsibility to moral responsibility (Corlett 2008), or assess what our responsibilities are in granting authority to sources of information (Origgi 2008). While the topic of epistemic responsibility can be addressed with respect to different sources of knowledge, such as memory or perception, it is most interesting in the context of testimonial knowledge practices, i.e., practices related to receiving knowledge through the spoken or written words of others.
In recent years, testimony has emerged as a central topic within social epistemology, the philosophical discipline addressing the various ways in which knowledge is social. In contrast to the abundance of publications on testimony (e.g. Coady 1992; Fricker 2007; Adler 1994) and related topics such as epistemic trust (e.g. Origgi 2004; Simon 2010), epistemic authority (e.g. Origgi 2008), epistemic injustice (especially Fricker 2007), epistemic responsibility itself has only very recently attracted attention within analytic social epistemology.
Although insights from social epistemology, in particular those addressing epistemic practices in more applied settings are highly crucial for a notion of epistemic responsibility for the 21st century, there are several shortcomings: First and foremost, due to this origin in the debates around the epistemology of testimony, the focus of attention in this discourse of epistemic responsibility is also mostly on epistemic interactions between human agents, i.e. on the responsibilities of speakers and hearers in testimonial exchanges. Yet, taking into account that processes of knowing take place in increasingly entangled systems consisting of human and non-human agents, systems in which content from multiple sources gets processed, accepted, rejected, modified in various ways by these different agents, the notion of epistemic responsibility needs to be modified and expanded to account for such socio-technical epistemic processes. Two issues need to be addressed in more detail than is currently the case in most analytic accounts of epistemic responsibility: (a) the role of technology and (b) the relationship between power and knowledge. To ut technology in general and ICT in particular into the equation, we should turn to philosophy of technology and philosophy of computing. Regarding the relationship between power, knowledge and technology, it has been feminist theoreticians in particular who have provided highly valuable insights. Thoughts from both fields will be briefly introduced in the next two sections.
4.2 Responsibility & ICT: Insights from the Philosophy of Computing
The complexity and entanglement of social and technical compounds in many digital systems has lead to difficulties in locating agency, accountability and responsibility, which various philosophers of computing and computer ethicists aim to tackle. Unsurprisingly, there is a growing amount of research on moral and legal responsibility in computing (cf. Coleman 2004), specific foci being autonomous agents (e.g. Coeckelbergh 2009) and robotics (e.g. Pagallo 2010). With respect to accountability, Nissenbaum's (1997) paper on accountability in a computerized society is surely an early seminal piece, in which different causes for contemporary difficulties in accountability attribution are already worked out: the problem of many hands, the problem of bugs, using the computer as a scapegoat, and ownership without liability.
Of particular importance for the goals of this paper are Floridi and Sander's (2004) early considerations on the morality of artificial agents as well as Floridi's more recent analyses regarding distributed morality (Floridi 2012). According to Floridi and Sanders (2004) something qualifies as an agent if it shows interactivity, autonomy and adaptability, i.e. neither free will nor intentions are deemed necessary for agency. Such a concept of “mind-less morality” (Floridi and Sander 2004,
p. 349) allows addressing the agency of artificial entities (such as algorithms) as well as of collectives, which may form entities of their own (such as companies or organizations). Another merit of their approach lies in the disentanglement of moral agency and moral responsibility: a non-human entity can be held accountable if it qualifies as an agent, i.e. if it acts autonomously, interactively and adaptively. However, it cannot be held responsible, because responsibility requires intentionality. That is, while agency and accountability do not require intentionality, responsibility does. Therefore, it seems that non-human agents—as long as they a) do not exhibit intentionality and b) are considered in separation—cannot be held responsible even if they are accountable for certain actions.
While these considerations on responsibility and accountability in socio-technical systems are highly developed, the specific problem of epistemic responsibility in ICT has not yet been in the focus of attention within philosophy of computing. Hence, it appears worthwhile to take the best from both fields of research to develop a sound notion of epistemic responsibility within entangled socio-technical epistemic environments. Yet, instead of starting from scratch taking a look at feminist theory proves highly illuminating, because different feminist theoreticians have not only focused on the responsibilities of knowers in complex environments. They have also emphasized the important relationship between knowledge and power.
4.3 Epistemic Responsibility in Entangled Socio-Technical Systems: Insights from Feminist Theory
Despite the fact that epistemic responsibility has only very recently attracted attention within analytic epistemology, the term itself has already been used in 1987 as the title of a book by Lorraine Code (Code 1987). In this book, Code addresses the concepts of responsibility and accountability from a decidedly feminist perspective and argues that in understanding epistemic processes in general and epistemic responsibility and accountability in particular; we need to relate epistemology to ethics. Criticizing the unconditioned subject S who knows that p, “the abstract, interchangeable individual, whose monologues have been spoken from nowhere, in particular, to an audience of faceless and usually disembodied onlookers” (Code 1995, p. xiv), Code emphasizes social, i.e. cooperative and interactive aspects of knowing as well as the related “complicity in structures of power and privilege” (Code 1995, p. xiv), “the linkages between power and knowledge, and between stereotyping and testimonial authority” (Code 1995, p. xv).
While Code's work highlights the relationship between knowledge and power, research by Karen Barad and Lucy Suchman adds technology to the equation and therefore appears particularly suited to explore the notion of epistemic responsibility within entangled and distributed socio-technical systems:
Barad's “agential realism” (Barad 1996; Barad 2007) delivers an “[…] epistemological-ontological-ethical framework that provides an understanding of the role of human and nonhuman, material and discursive, and natural and cultural factors in scientific and other social-material practices” (Barad 2007, p. 26).
Barad's approach is theoretically based upon Niels Bohr's unmaking of the Cartesian dualism of object and subject, i.e. on the claim that within the process of physical measurement, the object and the observer, Barad's “agencies of observation”, get constituted by and within the observation process itself and are not pre-defined entities. The results of measurements are thus neither fully constituted by any reality that is independent of its observation, nor by the methods or agents of observation alone. Rather, all of them, the observed, the observer and the practices, methods and instruments of observation are entangled in the process of what we call “reality”. For Barad, reality itself is nothing pre-defined, but something that develops and changes through epistemic practices, through the interactions of objects and agents of observation in the process of observation and measurement. Reality in this sense is a verb and not a noun.
Yet, interaction is a problematic term in so far as it presupposes two separate entities to interact. Thus, to avoid this presupposed dualism, she introduces the neologism of “intra-action”, to denote the processes taking place within the objectobserver-compound, the entanglement of object and observer in the process of observation. This terminological innovation is meant to discursively challenge the prevalent dualisms of subject-object, nature-culture, human-technology, and aims at opening up alternative, non-dichotomous understandings of technoscientific practices.
A crucial concern of Barad is the revaluation of matter. Opposing the excessive focus on discourse in some other feminist theories, Barad emphasizes the relevance of matter and the materiality of our worlds. Taking matter serious and describing it as active, means to allow for non-human or hybrid forms of agency, a step that has been taken already with the principle of generalized symmetry in Actor-NetworkTheory. Yet, if we attribute agency to non-human entities, can and should they be held responsible and accountable? Plus, isn't that an invitation, a carte blanche to shirk responsibility by humans? Do we let ourselves off the hook too easily and throw away any hopes for responsible and accountable actions?
It appears that Barad's view on non-human agency and her stance towards the ontological symmetry between humans and non-humans has changed from earlier articulations (Barad 1996) to later ones (Barad 2007). In 1996, she still underscores the human role in representing, by stating that “[n]ature has agency, but it does not speak itself to the patient, unobtrusive observer listening for its cries—there is an important asymmetry with respect to agency: we do the representing and yet nature is not a passive blank slate awaiting our inscriptions, and to privilege the material or discursive is to forget the inseparability that characterizes phenomena” (Barad 1996, p. 181).
However, it seems that this special treatment of humans and especially the notion of representing does not well match her posthumanist performativity, as depicted some years later (Barad 2003). Finally, in “Meeting the Universe Halfway” Barad offers a more nuanced dissolution of the distinction between human and non-human agency. By stating that “[a]gency is a matter of intra-acting; it is an enactment, not something that someone or something has” (Barad 2007, p. 261), Barad moves the locus of agency from singular entities to entangled material-discursive apparatuses. But even if agency is not tied to individual entities, it is bound with responsibility and accountability, as Barad makes very explicit: “Learning how to intra-act responsibly within and as part of the world means understanding that we are not the only active beings—though this is never justification for deflecting that responsibility onto other entities. The acknowledgment of “nonhuman agency” does not lessen human accountability; on the contrary, it means that accountability requires that much more attentiveness to existing power asymmetries (Barad 2007, p. 218 f). Thus, the possibility to understand agency not essentialist as a (human) characteristic, but as something which is rather attributed to certain phenomena within entangled networks could be regarded as an invitation to shirk of responsibility. But this is clearly not the aim of Barad. When developing her posthumanist ethics, Barad concludes that even if we are not the only ones who are or can be held responsible, our responsibility is even greater than it would be if it were ours alone. She states: “We (but not only “we humans”) are always already responsible to the others with whom or which we are entangled, not through conscious intent but through the various ontological entanglements that materiality entails. What is on the other side of the agential cut is not separate from us—agential separability is not individuation. Ethics is therefore not about right response to a radically exterio/ized (sic!) other, but about responsibility and accountability for the lively relationalities of becoming of which we are a part” (Barad 2007, p. 393).
This focus on responsibility and accountability relates back to Barad's initial framing of agential realism as an “epistemological-ontological-ethical framework”, a term by which she stresses the “[…] fundamental inseparability of epistemological, ontological, and ethical considerations” (Barad 2007, p. 26). Barad insists that we are responsible for what we know, and—as a consequence of her onto-epistemology for what is (Barad 2003, p. 829). Accountability and responsibility must be thought of in terms of what matters and what is excluded from mattering, what is known and what is not, what is and what is not.
This acknowledgement that knowledge always implies responsibility, not only renders issues of ethics and politics of such knowledgeand reality-creating processes indispensable. It also relates directly back to Barad's emphasis on performativity: epistemic practices are productive and different practices produce different phenomena. If our practices of knowing do not merely represent what is there, but shape and create what is and what will be there, talking about the extent to which knowledge is power or entails responsibility gets a whole different flavor.
Lucy Suchman shares many concerns of Karen Barad and her insights promise to be of particular importance for considerations regarding computationally mediated environments due to Suchman's background in Human-Computer Interaction. Acknowledging the relational and entangled nature of the sociomaterial, Suchman claims that agency cannot be localized in individual entities, but rather is distributed within socio-material assemblages. Resonating with Barad, she notes “[…] agencies—and associated accountabilities—reside neither in us nor in our artifacts but in our intra-actions” (Suchman 2009, p. 285).
The question, however, remains how exactly to be responsible, how to hold or to be held accountable if agency is distributed. How can we maintain responsibility and accountability in such a networked, dynamic and relational matrix? Although I think that Suchman goes into the right direction, she remains quite vague about this in her concluding remarks of Human-Machine-Reconfigurations by stating that “responsibility on that view is met neither through control nor abdication but in ongoing practical, critical, and generative acts of engagement. The point in the end is not to assign agency either to persons or to things but to identify the materialization of subjects, objects, and the relations between them as an effect, more and less durable and contestable, of ongoing sociomaterial practices” (Suchman 2009, p. 285).
-  Confer for instance the conference on “Social Epistemology and Epistemic Responsibility”, which took place at Kings College in May 2012. kcl.ac.uk/artshums/depts/philosophy/ events/kclunc2012.aspx (Accessed 18 September 2013).
-  It would be inadequate to argue that the role of technology or the role of power have been entirely neglected in social epistemology. On the one hand, there have been attempts to account for ICT (e.g. some works by Alvin Goldman (2008) and Don Fallis (2006), the special issue of the journal EPISTEME (2009, volume 6, issue 1, on Wikipedia). Moreover, Fricker's book on “Epistemic Injustice” (2007) has also stirred a lot of interest in the relationship between power and knowledge. However, these developments are rather recent and the classical assessment of testimonial processes remains focused on communication between humans often still conceived as an unconditioned and a-social subject S, who knows that p.