Verbeek- Do Artifacts Have Morality

beings. Only in this way can justice be done to the observation that the me- dium of ethics is not only the language of subjects but also the materiality of objects. This implies a shift of ethics. In addition to developing lingual frame- works for moral judgment, ethics consists in designing material infrastruc- tures for morality. When matter is morally charged, after all, designing is the moral activity par excellence, albeit “by other means.” Designers materialize morality. Ethics is no longer a matter of only ethereal reflection but also of practical experiment, in which the subjective and the objective, the human and the nonhuman have become interwoven. From this interwoven character two important lines of thought can be discerned in a posthumanist ethics: designing morally mediating teclllnf:l- ogy (designing the human into the nonhuman) and using morally med}atzng technology in deliberate ways (coshaping the roles of the nonhuman in the human). These two lines might seem to reflect the modernist distinction be- tween an actively reflecting subject and a passively designed world. But rather than reinforcing this distinction, a posthumanist ethics aims to think both poles together by focusing on their connections and interrelations. ) Before addressing these lines in the ethics of technology, however, I will explore the implications of introducing the moral significance of technology into ethical theory. In chapter 3 I will articulate what the phenomenon f’f technological mediation implies for the role of the object in ethical fl?eom in chapter 4 I will investigate how the mediated character of mora?l acf‘nons efnd decisions calls for a reconceptualization of the role of the subject in ethical theory. 3 Do Artifacts Have Morality? Petee ~Pil Ver peok 7c“" MGml 12 /rcc hrat. Introduction (')7 How do we come to understand the moral dimension of technology?' Now that we have seen that technologies have moral relevance, and that ethics needs to expand its “humanist focus” to take this into account, the question rises how to conceptualize the morality of technology. What could it imply to say that technologies have a moral dimension? Do the examples that we have seen so far—ultrasound, speed bumps, cell phones—urge us to con- sider technologies to be moral entities, even moral agents? Or are there other ways to conceptualize the morality of technological artifacts? Approaching things in moral terms is not a self-evident enterprise. It goes against the grain of the most basic assumptions in ethical theory. After all, it would be foolish to blame a technology when something immoral happens. It does not make sense to condemn the behaviour ofagun when somebody has been shot; not the gun but the person who fired it needs to be blamed. Tsjalling Swierstra is a good representative of such hesitations regarding “moralizing things.” He discusses how the moral community has been expanded many times since classical antiquity. “Women, slaves, and strangers were largely or entirely devoid of moral rights,” but “over time all these groups have been admitted” (Swierstra 1999, 317).? The current inclination to also grant things access to the moral community, however, goes too far, he argues from the two predominant ethjcal positions: deontology and consequentialism, Consequentialist ethics evaluates actions in terms of the value of their outcomes, When the positive consequences outweigh the negative ones, an action can be called morally correct. From this perspective, Swierstra says, things can indeed be part of a moral practice, since they can incite human beings to behave morally—and from a consequentialist perspective it is only the result that counts. But things can do this only when human beings use

. CHAPTER THREE them for this purpose. Things themselves are not able to balance the positive and negative aspects of their influence on human actions against each other. They can only serve as instruments, not as fully fledged moral agents that are able to render account for their actions. Deontological ethics is directed not at the consequences of actions but at the moral value of the actions themselves. From a Kantian perspective, for in- stance, the morality of an action depends on whether the agent has intended to act in accord with rationally insightful criteria. Artifacts, of course, are not capable of taking up such considerations. Moreover, if they incite human be- ings to act in ways that are morally right from a deontological point of view, these actions are not results of a rationally insightful moral obligation but simply a form of steered behavior. This means that both from a deontological and a consequentialist per- spective, artifacts can only be causally responsible for a given action, not mor- ally. Artifacts do not possess intentions, and therefore they cannot be held responsible for what they “do.” In Swierstra’s words: “Compelling artifacts, therefore, are not moral actors themselves, nor can they make humans act truly morally. Therefore . . .there is no reason to grant artifacts access to the moral community.” (Swierstra 1999). 1 share Swierstra’s hesitations regarding a too radically symmetrical ap- proach to humans and things (cf. Verbeek 2005b, 214-17). Yet the argument that things do not possess intentionality and cannot be held responsible for their “actions” does not justify the conclusion that things cannot be part of the moral community. For even though they don’t do this intentionally, things do mediate the moral actions and decisions of human beings, and as such they provide “material answers” to the moral question of how to act. Ex- cluding things from the moral community would require ignoring their role in answering moral questions—however different the medium and origins of their answers may be from those provided by human beings. The fact that we cannot call technologies to account for the answers they help us to give does not alter the fact that they do play an actively moral role. Take technol- ogy away from our moral actions and decisions and the situation changes dramatically. Things can be seen as part of the moral community in the sense that they help to shape morality. But how to account for this moral role of technology in ethical theory? As stated in chapter 1, to qualify as a moral agent in mainstream ethical theory requires at least the possession of intentionality and some degree of freedom. Both requirements seem problematic with respect to artifacts—at least at first sight. Artifacts do not seem to be able to form intentions, and neither do S DO ARTIFACTS HAVE MORALITY? 43 they possess any form of autonomy. Yet both requirements for moral agency deserve further analysis. From the amodern approach set out in chapter 2, the conc?pt of agency—including its aspects of intentionality and freedom—can be reinterpreted in a direction that makes it possible to investigate the moral relevance of technological artifacts in ethical theory. This will be the main objective of this chapter. First, T will discuss the most prominent existing accounts of the moral significance of technological artifacts. After that, I will develop a new account in which I expand the concept of moral agency in sush away that it can do justice to the active role of technologies in moral actions and decisions. The Moral Significance of Technological Artifacts The question of the moral significance of technological artifacts has popped up every now and then during recent decades. Several accounts have been de- vt‘:loped, all of which approach the morality of technology in different ways. I will discuss the most prominent positions as a starting point for developin, .a philosophical account of the morality of technological artifacts, ¢ LANDON WINNER: THE POLITICS OF ARTIFACTS In 1980 Langdon Winner published his influential article “Do Artifacts Have Politics?” In this text, which was later reprinted in his 1986 book The Whale u'nd the Reactor, Winner analyzed a number of‘ “politically charged” technolo- §l€5.. T:xe most well-known example he elaborated concerns a number of Tacist” overpasses in New York, over the parkways to Jones Beach on Long Island. These overpasses, designed by architect Robert Moses, were delib- erately built so low that only cars could pass beneath them, not buses. This prevented the African American population, at that time largely unable to aff?rd cars, from accessing Jones Beach, Moses apparently had found a ma- terial way to bring forth his political convictions. His bridges are political en.ti(ies. The technical arrangements involved Ppreceded the use of the bridges Px}or to functioning as instruments to allow cars to cross the parkways thesel bridges already “encompass[ed] purposes far beyond their immediat; use” (Winner 1986). Winner’s analysis obtained the status of a “classic” in philosophy of tech- nology and in science and technology studies—even though it became the focus oflcontroversy in 1999, when Bernward Joerges published the article Do Politics Have Artefacts?” (Joerges 1999). In this article he showed that

46 CHAPTER THREE the means” (Latour 2002). Latour proposes instead to understand technology in terms of the notion of fold. In technical action, time, space, and the type of “actants” are folded together. Technologies cross space and time. A ham- mer, for instance, “keeps folded heterogeneous temporalities, one of which has the antiquity of the planet, because of the mineral from which it has been moulded, while another has the age of the oak which provided the handle, while still another has the age of the 10 years since it came out of the German factory which produced it for the market” (ibid., 249). The same holds true for space here: “the humble hammer holds in place . . . the forests of the Ar- dennes, the mines of the Ruhr, the German factory, the tool van which offers discounts on every Wednesday on Bourbonnais streets,” et cetera. By “the type of actants,” the third element that is folded into technical action, Latour means that both human and nonhuman agents are involved and help to shape each other. Technologies should not be understood merely in terms of functionality, for this would limit us to seeing only how human intentions can be realized with the help of nonhuman functionalities serving only as means of extension. Technologies are not simply used by humans— they help to constitute humans. A hammer “provides for my fist a force, a direction and a disposition that a clumsy arm did not know it had” (ibid., 249). In the same way, speed bumps are not simply neutral instruments that fulfill the function of slowing down drivers. “What they exactly do, what they suggest, no one knows, and that is why their introduction in the country- side or in towns, initiated for the innocent sake of function, always ends up inaugurating a complicated history, overflowing with disputes, to the point of ending up either at the State Council or at the hospital” (ibid., 250). Tech- nologies are not intermediaries, helping human intentions to be realized in the material world; they are mediators that actively help to shape realities. Technologies do not merely provide means but also help to form new ends; they do not provide functions but make possible detours. “Without technol- ogies, humans would be contemporaneous with their actions, limited solely to proximal interactions. . . . Without technological detours, the properly human cannot exist” (ibid., 252). The moral significance of technologies, for Latour, is part of this phe- nomenon of folding. Morality is a “regime of mediation” as well (ibid., 254). ‘We usually recognize morality in the form of obligation, but this is not the only form it can take, since it “derives just as much from contract, from re- ligious events, . . . from chains of references, from the law,” et cetera (ibid., 254). Rather than being a merely human affair, morality is to be found in nonhuman entities as well. “Of course, the moral law is in our hearts, but it is also in our apparatuses. To the super-ego of tradition we may well add the DO ARTIFACTS HAVE MORALITY? I under-ego of technologies in order to account for the correctness, the trust- worthiness, the continuity of our actions” (ibid., 253-54). This “under-ego” is present in the speed bumps that tell us how fast to drive, or the coin locks on supermarket carts, demanding that we put the carts back in their rack rather than leaving them beside our parking place. This does not imply, to be sure, that we need to understand technologies as moral agents in themselves. “In themselves” entities are quite meaningless anyway—they are given a character in the relations in which they function, In Latour’s words, “Nothing, not even the human, is for itself or by itself, but always by other things and for other things” (ibid., 256; empbhasis in original). Both morality and technology are “ontological categories” for Latour: “the human comes out of these modes, it is not at their origin” (ibid., 256). Tech- nologies help to constitute humans in specific configurations—including the moral character of our actions and decisions. ALBERT BORGMANN: TECHNOLOGY AND THE GOOD LIFE North American philosopher of technology Albert Borgmann has proposed a third position to describe the moral significance of technology. He has devel- oped a neo-Heideggerian theory of the social and cultural role of technology. In this theory, he elaborates how our culture is ruled by what he calls the “de- vice paradigm.” According to Borgmann, the technological devices that we use call for a quite different way of taking up with reality than did pretechno- logical “things.” While “things”—like water wells, fireplaces, musical instru- ments—evoke practices in which human beings are engaged with reality and with other people, devices primarily evoke disengaged consumption. Borgmann understands devices as material machineries that deliver con- sumable commodities—for example, the boiler and radiators of a heating in- stallation form a machinery that delivers warmth asa commodity. Devices ask foraslittle involvement as possible; they create the availability of commodities bykeeping their machinery in the background as much as they canand putting their commodities in the foreground. Against this, “things” do not separate machinery from commodity. Rather, they engage people. Usinga fireplace, for instance, requires people to collect and chop wood, to clean the hearth regu- larly, to gather around the fireplace to enjoy the warmth it gives, and so on. In his article “The Moral Significance of Material Culture,” Borgmann explains how his theory of the device paradigm makes visible a moral dimen- sion in material objects. He focuses on the role of material culture in human practice, and shows how “material culture constrains and details Ppractice

48 CHAPTER THREE decisively” (Borgmann 1995, 85). In line with his device paradigm, he makes a distinction between two kinds of reality: one commanding, the other dispos- able. While a traditional musical instrument is a commanding thing, one that requires a lot of effort and skill and needs to be “conquered,” a stereo liter- ally puts music at our disposal. The quality of sound can be even better than that of a live performance, but the music’s presence is not commanding like that of the music performed live by a musician. According to Borgmann, the device paradigm increasingly replaces commanding reality with disposable reality. In his book Real American Ethics, Borgmann elaborates the concept of moral commodification to analyze this phenomenon: “a thing or practice gets morally commodified when it is detached from its context of engagement with a time, a place, and a community and it becomes a free-floating object” (Borgmann 2006, 152; italics in orginal). We find the moral significance of the material culture in its role in shap- ing human practices. While commanding reality “calls forth a life of engage- ment that is oriented within the physical and social world,” disposable reality “induces a life of distraction that is isolated from the environment and from other people” (ibid., 92). Human practices take place not in an empty space but in a material environment—and this environment helps to shape the quality of these practices. “If we let virtue ethics with its various traditional and feminist variants stand in for practical ethics, we must recognize that virtue, thought of as a kind of skilled practice, cannot be neutral regarding its real setting. Just as the skill of reading animal tracks will not flourish in a metropolitan setting, so calls for the virtues of courage and care will remain inconsequential in a material culture designed to produce a comfortable and individualist life” (Borgmann 199, 92). Even if we do not entirely follow Borgmann in his rather gloomy ap- proach to technology—I think there is engaging technology as well: see Ver- beek 2005b—his position highlights a significant form of the moral relevance of technology. Material objects, to summarize his position, help to shape human practices. And because the quality of these practices is ultimately a moral affair, material objects have direct moral relevance. Technological devices and nontechnological “things” help to shape the ways we live our lives—and the question of “the good life” is one of the central questions in ethics. Human actions and human life do not take place in a vacuum but in areal world of people and things that help to shape our actions and the ways we live our lives. And therefore, the good life is not formed only on the basis of human intentions and ideas but also on the basis of material artifacts and arrangements. Technologies provide a setting for the good life. DO ARTIFACTS HAVE MORALITY? 49 LUCIANO FLORIDI AND J. W. SANDERS: ARTIFICIAL MORAL AGENCY A radically different but equally interesting approach was elaborated in 2004 by Luciano Floridi and J. W. Sanders in their influential publication “On the Morality of Artificial Agents.” Their article deals with the question to what extent artificial agents can be moral agents. Rather than focusing on the moral significance of technologies in general, they focus on intelligent technologies that could actually qualify as “agents.” Examples of such artificial agents are expert systems that assist people in making decisions, driving assistants that help people to drive their cars, and automatic thermostats in houses. The approach Floridi and Sanders develop is so interesting because they give an account of artificial moral agency in which moral agents do not nec- essarily possess free will or moral responsibility. This way, they take away the obvious objection that technologies, lacking consciousness, can never be ‘moral agents as human beings are. It i crucial to Floridi and Sanders’s analy- sis that they explicitly choose an adequate “level of abstraction” at which it becomes possible and meaningful to attribute morality to artificial agents— such an abstraction is needed in order to avoid the obvious objection that artifacts cannot have agency as humans do. As criteria for agenthood, there- fore, Floridi and Sanders use “interactivity (response to stimulus by change of state), autonomy (ability to change state without stimulus) and adaptabil- ity (ability to change the ‘transition rules’ by which state is changed).” This implies that a system that interacts with its environment but is also able to act without responding to a stimulus and has the ability to learn how to “behave” in different environments could qualify as an agent. They use the ability to cause good or evil as the criterion for morality: “An action is said to be mor- ally qualifiable if and only if it can cause moral good or evil. An agent is said to be a moral agent if and only if it is capable of morally qualifiable action” (Floridi and Sanders 2004, 12). Their approach reveals what Floridi and Sanders call “aresponsible mo- rality” (ibid., 13). They consider intentions—“intentional states,” in the vo- cabulary of the analytic tradition from which they work—as a “nice but un- necessary condition” for moral agency. The only thing that matters for them is whether the agent’s actions are “morally qualifiable”—that is, whether they can cause moral good or evil. However, Floridi and Sanders do not aim to declare the concept of responsibility obsolete. Rather, they separate it from moral agency as such, which opens for them the space needed to clarify the role responsibility actually plays in morality (ibid., 20).

52 CHAPTER THREE be adequate to attribute moral agency to technologies “themselves”—as if “agency” were some intrinsic property of technology. Latour’s claims that nonhumans can be agents as well and that there is morality in technology need to be read in the context of his actor-network theory, in which all enti- ties are understood relationally. From this perspective, technologies do not have moral agency in themselves; rather, when humans use technologies, the resulting moral agency is not exclusively human but incorporates nonhuman elements as well. Contrary to the position of Floridi and Sanders, for Latour technologies only “have” agency and morality in the context of their relations with other agents. MORAL MEDIATION Actually, Latour’s approach occupies a third position with respect to the moral relevance of technology. Rather than moral instruments or moral agents, Latour’s work makes it possible to see technologies as moral media- tors. This position does justice to the active moral role of technologies in moral actions and decisions, without reducing this role entirely to human intentions. At the same time, it avoids characterizing morality as an intrinsic property of the technologies themselves. By mediating human experiences and practices—as elaborated in chapter 1—technologies mediate moral deci- sions and actions. Technologies help us to phrase moral questions and find answers to them, and they guide our actions in certain directions. The notion of “mediator” expresses both the active moral role of tech- nologies and the relational character of this moral role: they mediate, rather than being some kind of neutral “intermediary,” but mediators can func- tion only in the context of an environment for and in which they mediate. The moral significance of Latour’s speed bumps and Winner’s overpasses can be understood best in terms of moral mediation. Understanding them as moral instruments for realizing the racist intentions or safety ambitions of city planners appeared to fall short, because this does not recognize the un- intended roles these artifacts can play. Understanding them as moral agents would go too far, at least in the sense of being moral agents “in themselves,” capable of moral action. Only in the context of the practices in which they function do their moral roles emerge. Sometimes these roles coincide with the intentions of their designers, sometimes they don’t. In all cases, the moral roles of technologies come about in the context of their relations with their users and the environment in which they function. Borgmann’s approach to the moral significance of technology is an inter- DO ARTIFACTS HAVE MORALITY? 53 esting supplement to the notion of moral mediation. He broadens the dis- cussion from action-oriented ethics to the classical ethical question of the good life by focusing on technologies as providing a material setting for the good life. In Borgmann’s approach, the moral role of technologies is not to be found in the ways technologies help to shape human actions but in how they help to answer the classical question of “how to live.” Borgmann’s example of the difference between a stereo set and a musical instrument does not revolve around the different actions involved in operating the two but around their roles in shaping a way of life. By conceptualizing technologies as moral mediators, we can bring the postphenomenological approach to technological mediation into the realm of ethics. As we saw in the example of obstetric ultrasound,technologies-in- use establish a relation between their users and their world. Ultrasound im- aging organizes a specific form of contact between expectant parents and un- born child, in which the parents and the child are constituted in specific ways with specific moral roles, responsibilities, and relevance. Along the same lines, larger-scale technologies mediate moral actions and decisions; energy production systems, for instance, help to organize a way of living in which it becomes ever more normal and necessary to use large quantities of energy, and in doing so they help to shape moral decisions regarding how we deal with environmental issues. To be sure, approaching technologies as moral mediators does not imply that we need to reject Latour’s ideas about nonhuman agency. Indeed the no- tion of moral mediation implies a form of technological agency. Moral medi- ation always involves an intricate relation between humans and nonhumans, and the “mediated agency” that results from this relation therefore always has a hybrid rather than a “purely nonhuman” character. When technologies are used, moral decisions are not made autonomously by human beings, nor are persons forced by technologies to make specific decisions. Rather, moral agency is distributed among humans and nonhumans; moral actions and de- cisions are the products of human-technology associations. The way I use the notion of moral mediation is different from the way Lorenzo Magnani uses it in his book Moralityina Technological World (2007). Magnani lays out an approach to morality and technology that is congenial to the approach set out in this book but reaches different conclusions, Be- cause his approach departs from the perspective of cognitive science rather than phenomenology, it cannot take into account the hermeneutic and prag- matic dimensions of technological mediation that are so central to the ac- count developed here. For Magnani, moral mediators mediate moral ideas.

54 CHAPTER THREE In his definition, “moral mediators . . . are living and nonliving entities and processes—already endowed with intrinsic moral value—that ascribe new value to human beings, nonhuman things, and even to ‘non-things’ like fu- ture people and animals.” Even though he discusses Latour’s work assent- ingly (ibid., 25-26), he does not acknowledge that Latour’s actor-network theory radically differs from his cognitive approach. Magnani’s strong fo- cus on knowledge as the primordial variable in ethics and in moral media- tion is rather remote from Latour’s focuses on practices, interactions, and materiality. For Latour, and for the postphenomenological approach that uses his work, the cognitive approach makes too sharp a distinction between (subjec- tive) minds that have knowledge and the (objective) world that this knowl- edge is about. In the approach I follow in this book, morality should not be understood in terms of cognitive “templates of moral doing” (Magnani 2007, 187-93) but in terms of ways of being-in-the-world which have both cognitive and noncognitive aspects and which are technologically mediated in more-than-cognitive ways. In my postphenomenological approach, tech- nological mediation concerns action and perception rather than cognition; and moral mediation is not only about the mediated character of moral ideas but mostly about the technological mediation of actions, and of perceptions and interpretations on the basis of which we make moral decisions. The concept of moral mediation has important implications for under- standing the status of objects in ethical theory. As indicated in my introduc- tion, in mainstream ethical theory “objects” have no place apart from being mute and neutral instruments that facilitate human action. Now that we have seen that technologies actively help to shape moral actions and decisions, we need to expand this overly simplistic approach. The mediating role of tech- nologies can be seen as a form of moral agency—or better, as an element of the distributed character of moral agency. L will rethink the status of the object in ethical theory in two ways. First, Iwill offer a “nonhumanist” analysis of two criteria that are usually seen as conditions sine qua non for moral agency. An entity can be called a moral agent if it can be morally responsible for its actions, and to be morally respon- sible, it needs at least (1) intentionality—the ability to form intentions—and (2) the freedom to realize its intentions. I will show that these two criteria can be reinterpreted along postphenomenological lines in such a way that they also pertain to nonhuman entities. Second, I will investigate the possibility that the predominant ethical approaches propose to take seriously the moral dimension of technologies. By elaborating what role objects could play in DO ARTIFACTS HAVE MORALITY? 55 deontological ethics, consequentialism, and virtue ethics, I will create the space needed to take the moral significance of technologies seriously. Technological Intentionality The first criterion for moral agency—the possession of intentionality—di- rectly raises a serious problem for anyone who intends to defend some form of moral agency for technology. While agency is not thinkable without inten- tionality, it also seems absurd to claim that artifacts can have intentions. Yet a closer inspection of what the concept of intentionality can mean in relation to what artifacts actually “do” makes it possible to articulate a form of “tech- nological intentionality.” The concept of intentionality actually has a double meaning in philoso- phy. In ethical theory, it primarily expresses the ability to form intentions. In phenomenology, though, the concept of intentionality indicates the di- rectedness of human beings toward reality. Intentionality is the core concept in the phenomenological tradition for understanding the relation between humans and their world. Rather than separating humans and world, the con- cept makes visible the inextricable connections between them. Because of the intentional structure of human experience, human beings can never be un- derstood in isolation from the reality in which they live. They cannot simply “think” but always think something; they cannot simply “see” but always see something; they cannot simply “feel” but always feel something. As experi- encing beings, humans cannot but be directed at the entities that constitute their world. Conversely, it does not make much sense to speak of “the world in itself.” Just as human beings can be understood only through their rela- tion with reality, reality can be understood only through the relation human beings have with it. “The world in itself” is inaccessible by definition, since every attempt to grasp it makes it a “world for us,” as disclosed in terms of our particular ways of understanding and encountering it. In the context of this discussion of the possibility of “artifactual moral agency,” these two meanings of the concept of intentionality augment each other. The ability to form intentions to act in a specific way, after all, can- not exist without being directed at reality and interpreting it in order to act in it. Actually, the two meanings of intentionality have a relation to each other similar to the relation between the two dimensions of technological mediation that I discerned in chapter 1. The “praxical” dimension, concern- ing human actions and practices, cannot exist without the “hermeneutical” dimension, concerning human perceptions and interpretations—and vice

58 CHAPTER THREE has led to a different way of dealing with appointments, especially for young people—making plans far in advance for a night out does not make much sense when everyone can call each other anytime to make an ad hoc plan. This change in behavior was not intended by the designers of the cell phone, even though the phone is being used in precisely the context the designers had envisaged. And nobody foresaw that the introduction of the energy- saving lightbulb would actually cause people to use more rather than less energy. Apparently such bulbs are often used in places previously left unlit, such as in a garden or on the front of a building, thereby canceling out their economizing effect (Steg 1999; Weegink 1996). It seems plausible, then, to attribute a form of intentionality to artifacts— albeit a form that is radically different from human intentionality. The inten- tional “dimension” of artifacts cannot exist without human intentionalities supporting it; only within the relations between human beings and reality can artifacts play the mediating roles in which their “intending” activities are to be found. For example, when expectant parents face a decision about abortion on the basis of technologically mediated knowledge of the chances that the child will suffer from a serious disease, this decision is not “purely” human, but neither is it entirely induced by technology. The very situation of having to make this decision and the very ways in which the decision is made are coshaped by technological artifacts. Without these technologies, either there would not be a situation of choice or the decision would be made on the basis of a different relation to the situation. Yet the technologies involved do not determine human decisions. Moral decision making is a joint effort of human beings and technological artifacts. Technological intentionalities are one component of the eventually resulting intentionality of the “composite agent,” a hybrid of human and technological elements. Strictly speaking, then, there is no such thing as “technological inten- tionality”; intentionality is always a hybrid affair involving both human and nonhuman intentions, or, better, “composite intentions” with intentional- ity distributed among the human and the nonhuman elements in human- technology-world relationships. Rather than being “derived” from human agents, this intentionality comes about in associations between humans and nonhumans. For that reason it could best be called “hybrid intentionality” or “composite intentionality.” Technology and Freedom A second requirement that is often connected to moral agency is the pos- session of freedom. If moral agency entails that an agent can be held mor- DO ARTIFACTS HAVE MORALITY? 59 ally responsible for his or her actions, this requires not only that the agent needs to have the intention to act in a particular way but also that he or she has the freedom to realize this intention. Now that we have concluded that artifacts may have some form of intentionality, can we also say that they have freedom? The answer obviously seems to be no. Again, freedom requires the pos- session of a mind, which artifacts do not have. Technologies cannot be free agents as human beings are. The only degree of freedom that could be as- cribed to them is their “ability” to have unintended and unexpected effects, like the increase in energy use brought on by the energy-saving lightbulb. But this is not freedom, of course, in the sense of the ability to choose and to have arelation to oneself and one’s inclinations, needs, and desires. Still, there are good arguments not to exclude artifacts entirely from the realm of freedom. First of all, even though freedom is obviously required if one is to be ac- countable for one’s actions, the thoroughly technologically mediated char- acter of our daily lives makes it difficult to make freedom an absolute cri- terion for moral agency. This criterion might exist in a radical version of Kantian ethical theory, where freedom is understood in terms of autonomy and where the moral subject needs to be kept pure of polluting external in- fluences. But many other ethical theories take into account the situated and mediated character of moral agency. People do not make moral decisions in avacuum, after all, but in a real world, which inevitably influences them and helps to make them the persons they are. The phenomenon of technological mediation is part of this. Technologies play an important role in virtually every moral decision we make. The decision how fast to drive and therefore how much risk to run of harming other people is always mediated by such things the layout of the road, the power of the car’s engine, the presence or absence of speed bumps and speed cameras. The decision to have surgery or not is most often mediated by all kinds of imaging technologies and blood tests, which help to constitute the body in specific ways and organize specific situations of choice. Moral agency, therefore, does not require complete autonomy. Sorme de- gree of freedom can be enough for one to be held morally accountable for an action. And not all freedom is taken away by technological mediations, as the examples of abortion and driving speed make clear. In these examples, human behavior is not determined by technology but rather coshaped by it, with humans still being able to reflect on their behavior and make decisions aboutit. Nevertheless, we can in no way escape these mediations in our moral decision making. The moral dilemmas of whether to have an abortion and of how fast to drive would not exist in the same way without the technologies