PROBLEMS OF RESPONSIBILITY AND INFORMATION SYSTEMS

We now have all of the theoretical background to understand what responsibility can mean in the field of information systems and at the same time recognise the problems that this poses. In order to understand what the different problems of responsibility and IS are, we should look at some brief scenarios that revolve about different applications of responsibility and IS:

• The CIO (Chief Information Officer) is responsible for the development and functioning of the company s information systems.

• Programmer Smith assumes (is ascribed) responsibility for not eliminating the bug that resulted in the loss of vital data.

• Electronic commerce is responsible for a huge potential in savings due to the decrease in transaction costs.

• Computers and computerised trading are responsible for the globalisation of financial transactions.

• The central computer of the air defence system was responsible for the decision to fire the rocket and shoot down the airliner.

We see that responsibility and IS can be combined in a number of different ways that are not necessarily related to one another. Computers can play the role of the subject as well as the object of responsibility. They can have different temporal directions, they can be descriptive, normative, reflexive, transitive, and so on. In the following subsections we will discuss most of these aspects in so far as they are relevant to the use of computers and IT in business. In the first step we will concentrate on the conditions of responsibility.

The purpose of the next few sections is to recapitulate the literature on responsibility and demonstrate the difficulties it entails. It will become clear that it is almost impossible to construct a case or an example of responsibility that fulfils all of the conditions, has an acceptable subject and object, that follows the reasoning of moral philosophers , and at the same time has tangible results. Not only are some of the commonly named ingredients of responsibility contradictory, it is even doubtful whether the notion holds any promises for those cases that it was designed for. The great moral questions stemming from social change set off by technological developments mostly seem resistant to the notion of responsibility. This is why Apel (1980, p. 229) can say: A universal, i.e., intersubjectively valid ethics of collective responsibility thus seems both necessary and impossible.

Briefly , what we intend to do in the following pages is demonstrate the limits of responsibility in order to lay the foundations for the subsequent development of a theory of reflective responsibility.

Conditions of Responsibility

Conditions of responsibility are those circumstances, traits, or properties that have to be fulfilled in order to render the ascription of responsibility viable and acceptable. The bulk of these conditions refer to the subject. The subject is the person, group , or other entity that is responsible. Generally it is understood to be the grown-up human in her capacity as a person. We will look into the questions concerning the subject in the next section after discussing what conditions it is supposed to fulfil.

It is without a doubt one of the central tasks of any theory of responsibility to determine what the conditions of responsibility are. They are necessary to determine who or what can act as subject, object, or instance. They are also the foundation of the discovery of excusing circumstances and a successful ascription. Depending on the theory of responsibility one talks about and the corresponding objectives of the ascription, the conditions that have to be fulfilled vary. There is, however, a group of circumstances often cited as conditions that seem to be necessary for most theories of responsibility. In a first step one can distinguish between objective conditions that are independent of specific subjects and the personal conditions that have to be fulfilled by a subject in order for it to play its role. One of the most commonly named objective conditions is that of causality .

Causality

We have already mentioned that one of the meanings of responsibility can be causality. The limitation of responsibility to purely causal relationships is too narrow to be of any real interest. Nevertheless, there is a close connection of responsibility between responsibility and causality that can be expressed by naming causality as a central condition of the ascription of responsibility. Causality is never the only condition, but it is generally recognised as a conditio sine qua non of responsibility. A necessary condition an agent must satisfy to be held morally responsible for a decision is that the agent must figure in the causal chain resulting in the decision or be capable of such figuring in the causal chain (Bechtel, 1985, p. 298). This necessity of the subject to figure in the causal chain is true for moral (cf. Jonas, 1984, p. 172; Ulrich, 1997, p. 74) as well as legal responsibility (cf. Moore, 1999).

Causality, however, is a problematic concept. In our everyday life we may not be bothered by its intricacies, but if we want to realise valid ascriptions of responsibility, we find that it is difficult. The first problem is that the nature of causality is unclear. Is it something that can be found in the objects or is it just in our perception? Is it maybe just one of our own habits, forced onto us by evolution as a programme with life-enhancing properties (Riedl, 2001, p. 76)? Another problem is that it cannot be limited. If we look at the physical image of cause and effect, we find that every effect turns out to be a cause for the next generation of effects. Whenever we talk about causality and look at a certain effect and look for its causes, we have to be aware that this distinction is always temporal and may change quickly. On the other hand it is also impossible to find the first cause for any given effect because that in turn was caused by something else. When De George (De George, 1999, p. 111) says that the causal chain sometimesisalongone, hepalliatesthesituation.Itisnotonlylong,itisinfinite. We therefore have to be aware that every causal relationship that we use as the basis of a responsibility ascription is itself just that, an ascription. Being the start of a causal chain as a prerequisite of responsibility, as R tzer (1998, p. 18) puts it, is only a way of describing that we have to posit a point at which we start the ascription.

But even if we accept that causality is nothing but a construct with the purpose of explaining things and ultimately imposing judgments and sanctions, we still have the problem that there are no clear rules to follow in this construction. There are different sorts of causality. For example, it can be direct or indirect (Ladd, 1992, p. 293). Direct causality is what is intended, obvious to the agent and the observer (even though what is obvious to me does not have to be obvious to you), and follows in a straight line. Cases where this sort of causality exists are generally cases where responsibility ascriptions do not pose too difficult a problem. However, normally this is not the case. Most ethically relevant developments today are not intended, not caused by a single person or group, and do not follow a straight line. The obvious example in this work is computer networks, especially the Internet. The Internet has a pervasive influence on our daily lives, on the way we do business, on the way we interact. Causalities in the Internet are hard to come by, however. There is no clear answer to questions like, Who built the Internet? or Who is in charge of the development so far and who determines how it continues? The Internet is a good example of an artefact that was built by a collective that was never constituted formally . Many people have participated for many different reasons, and the result that we see today is cumulative and did not causally result from anything or anybody that could be easily described in hindsight. Collective and cumulative action is thus one example for a loss of causality that leads to difficulties with the ascription of responsibility.

Apart from the fact that there is no single person or group to whom one could ascribe causality, another problem with the Internet is that it was not explicitly intended the way it is used now by anyone . When we look at the Internet today from an economic point of view, we see a universal platform used for e-commerce and e-business that promises huge improvements in the way we do business and equally huge cost savings by facilitating information and interaction, and lowering transaction costs. This was not intended by the founders of the mainly military ARPA net, nor by Berners-Lee when he thought up the World Wide Web. One can see that even if we suppose that the philosophically fundamental problems of causality are overcome by using a constructivist approach, causality is still a problem because it may be next to impossible to construct causality for cumulative action or unintended results. These problems will return later on in the discussion of the subject and object of responsibility. For now we will leave these questions and proceed to the next condition of responsibility that is often seen as contradictory to causality, to freedom.

Freedom

It is generally agreed that responsibility presupposes freedom. For a subject to be responsible, it must have (had) alternatives to the actions it did or will do. Without such a margin we cannot speak of responsibility (cf. Picht, 1991; Frankfurt, 1997).

Traditionally the most influential view about the sort of freedom necessary and sufficient for moral responsibility posits that this sort of freedom involves the availability of genuinely open alternative possibilities at certain key points in one s life. Without this sort of alternative possibility, it is alleged, one is compelled to do as one actually does. ( Fischer, 1999, p. 99)

Responsibility and freedom are in a relationship of mutual reference. Responsibility is based on the supposition of freedom, and freedom leads to responsibilities (Bayertz, 1995a, p. 12). Freedom can even become an object of responsibility because the subject has to ensure its freedom. If I give up my freedom to choose between alternatives without a good reason, then I may be held responsible for the loss of liberty (Wunenburger, 1993, p. 174). The relationship between freedom and responsibility may be even clearer when we look at the negative side. It is not plausible to ascribe responsibility for something that the subject had no freedom to decide, that happened as an act of fate (Birnbacher, 1995, p. 169; S nger, 1991, p. 76).

This apparent unanimity of authors on a basic point is probably based on common sense. In cases where there is no choice, it does not make sense to hold someone responsible. If, for whatever reason, we could agree that it would be a moral duty to float one foot above the ground, then we still could not hold people responsible for not doing so because it is simply impossible for humans . Similarly we only hold people in IS responsible for things that they are free to change. The programmer is not responsible if the hardware does not work correctly, and the CIO is not responsible if an earthquake destroys the company s information system. On the other hand these examples lead us to the question of the limits to which freedom goes and how people are supposed to evaluate their choices. The programmer might have had the freedom to check the hardware first and the CIO might have known that in California earthquakes are a frequent occurrence.

There are other more philosophical problems of freedom. One is the distinction between freedom of action and freedom of will. Even if we follow H ffe (1995, p. 23) in saying that freedom of action is not problematic because it results from mankind s anthropological constants such as lack of instincts and intelligence, which necessitate free action, it is much less clear whether humans also have a freedom of will. In our context this means that even if an external observer might discern several alternatives that the subject could choose and thus conclude that it is possible to ascribe responsibility to him or her, it is not clear whether the subject knows about these alternatives and perceives the possibility of such a choice. If the subject does not see the alternative, then it is doubtful whether he or she can really be described as free and therefore whether responsibility is appropriate. Usually, when freedom is mentioned as a prerequisite of responsibility, both sorts of freedom ”freedom of will and of action ”are implied (cf. Koller, 1996; Bayertz, 1993, p. 31; Voland, 1998, p. 47). On the other hand the subject may feel perfectly free to decide what she does, but an outside observer may be capable to completely predict what she will to. This might suggest a lack of freedom.

In other words, subjective and objective meanings of free will need not, and in general do not, coincide. The feeling of freedom can perfectly well coexist with perfect predictability and perfect determinism. (Staddon, 1999, p. 148)

This leads us to the most frequently discussed part of the question of freedom, to the relationship of freedom and determinism. In the world of physics, there is neither freedom nor coincidence . Everything can be reduced to cause and effect. This has led the French philosopher and mathematician Laplace to stipulate that if there was a demon that knew the location and momentum of every particle, then this demon could perfectly predict the future. While modern physics and Heisenberg s uncertainty principle indicate that it is impossible to gather the required information, even in theory, this does not change the principle. Human beings live in this world of physics, and a strong theory of determinism assumes that we as physical beings are thus completely determined by the laws of cause and effect. Even our thoughts and in this case my writing about them are nothing but predetermined. This may contradict my feeling that I am writing because I decided to do it and could just as well do something else, but that feeling is of course itself an effect caused by the movement of electrons, atoms , or molecules in my brain, and those movements are subject to predetermined paths. Fisher calls this causal determinism, which is defined as:

the claim that a complete statement of the laws of nature and a complete description of the (temporally non-relational or ˜ genuine ) facts about the world at some time T entail every truth about the world after T. (Fischer, 1999, p. 99f)

There are also weaker forms of determinism such as social determinism. In these theories we find statements to the effect that we are much less free than we think and that even those things that we may perceive as intimate personal decisions are in fact determined by our social background. Bourdieu (1998) spends a considerable effort showing that matters of taste ”of our likes and dislikes in music, art, etc. ”are to a great extend predetermined by the social status of our parents, our formal education, etc.

The clash between determinism and responsibility is not a new one. A perennial subject of dispute in the Western philosophical tradition is whether human agents can be responsible for their actions even if determinism is true (Meyer, 1999, p. 250). Many authors hold that if a strong determinism is true, we lack the freedom of will and action necessary for an ascription of responsibility. But the alternative is no better. If we move away from determinism, the solution seems to be to give humans the opportunity to do things that are not caused by anything outside of them. Apart from the fact that this would ascribe attributes of the divine first mover to humans (Chisholm, 1997, p. 49f), it also leads us to problems of causality. The freedom to do things that are not caused by anything or anybody would tear apart causal chains. This would then in turn lead to problems with the first condition of responsibility, with causality. This ends in a dilemma. If determinism is true, we cannot make a difference in the world and can therefore not be responsible. If it is not true, then causality does not work any more and we may make a difference, but nobody will know about the reasons, and therefore responsibility does not make sense either.

The position that sees responsibility and determinism as mutually exclusive is often called ˜incompatibilism . There are many good philosophical texts on this topic that we will not be able to discuss here. ^[9] One solution to the dilemma is to move away from the strong determinism towards a weaker one. Responsibility depends on a degree of determinism, namely, predictable response to contingencies of reward and punishment (reinforcement contingencies) (Staddon, 1999, p. 173). Wallace (1996) also argues for the compatibility of determinism and responsibility on the basis of his concept of responsibility as a moral sentiment. To ascribe responsibility, which means to have distinctive moral feelings towards someone, does not require that person to be free.

The problem of freedom and determinism is a very old one, and we will not be able to solve it here. If we want to keep on talking about responsibility, that of course means that the philosophical position of this paper must be compatibilist. For us, a constructivist position seems most appropriate. Starting from Kant s observation that man s freedom can be proven as well as his lack of freedom (cf. Kant, 1995b, p. B566ff), one can ask how that is possible. A constructivist solution would be to say that we can never observe the world itself and determine the possibility of freedom within it. The only thing we can do is analyse our models of the world. Differing models can lead to different solutions. In this case, the physical model of the world does not leave any room for freedom. Our own experience, however, is another model that is no worse than physics, and some of us do occasionally feel free in certain respects. If we follow the constructivist credo and give up hope that we can ever see the world as it is and that we will always have to work with self- constructed models of differing adequacy for different problems, then we may be able to combine two apparently contradictory world views. The conclusion for this text is that one can recognise the problem of determinism and still believe that the ascription of responsibility makes sense. This may be done by Wallace s strategy of interpreting responsibility as a moral sentiment or by another strategy, for example by concentrating on the presumed social utility of responsibility ascriptions.

One particular aspect of determinism with important implications for our topic is technological determinism. This is the doctrine that technological development follows laws that are akin to natural laws in that they are unchangeable and can at best be described but never changed. This is a necessary consequence of a strong determinist theory. One can also hear this kind of thought from others who would not subscribe to a strong determinism. Indeed, the development of computers and information technology appear more similar to natural growth than to controlled and intentional developments. There are even so-called laws that describe this development. The most important one of these laws is probably Moore s law, which says that the number of transistors on a computer chip doubles every 18 to 24 months. The better known result of this law is that the speed of micro-processors doubles every 18 to 24 month at a constant cost (Pitt, 1999, p. 119). Another law is Grosch s law, which says that the computational power of computers rise with the square of the cost (Zerdick et al., 2001, p. 116).

While especially Moore s law has worked for more than 30 years despite regularly recurring objections that we would soon come to a natural end of the development, one should nevertheless be careful to compare these computer laws to natural laws. The danger in calling them laws is that they may be accepted as unchangeable and thus cease to be subject to ethical discussions. Another case in point is the development of the Internet. There are many ethical pros and cons to this artefact, some of which will be discussed later on in the book. These ethical issues should give rise to a discussion about the way we want the Internet to be developed. Do we want it mainly as an economic means, do we want to use it for educational or political purposes, do we want central control or complete freedom? All of these are valid questions, and their answers determine how societies go about developing technical infrastructure, rights of access, etc. However, if we accept that the development of the Internet follows some kind of economic, technical, or other law, that it is thus predetermined, then these ethical discussions will not take place and this is in itself a moral fact that would need to be discussed.

We, as a society and as humankind, have to realise that we can control technical developments and that we have to take responsibility for them. Apart from the argument of technological determinism, we can also find the thesis of technological constraints. This is a weaker thesis in that it says that, originally, technological development followed human intentions, but the original development now requires certain technological advances. This argument has more validity than the pure determinism because it can point to some facts that are hard to ignore. Following a majority of philosophers of technology, we should keep in mind that technology is made and shaped by humans, and that humans therefore keep responsibility for it (cf. Rohbeck, 1993, p. 204; Guggenberger, 1992, p. 37; Murswik, 1991, p. 160). Even admitting that not all choices are open to us, that we are for example unable to return to an agrarian society without great losses, one still has to admit that the use of technology for good or bad is up to human beings (Walther, 1992, p. 85).

Freedom, as we have seen, implies that there is the freedom of choice and the possibility to act. Consequently, one can find this ability to act as another possible precondition for responsibility. The ability to act, in order to be wholly convincing, should be based on an theory of action. A theory of action in turn will generally have to make the same assumptions as a theory of responsibility. The agent is usually imagined as possessing a certain degree of freedom of will and action, intentions, imagination , means to achieve objectives, control over her actions, and the ability to correct mistakes (Zimmerli, 1991, p. 84). The ability to act can thus be seen as a criterion of responsibility that is similar to but weaker than the criterion of being a person that will be discussed later on (Maring, 1989, p. 38).

While the questions of causality and freedom are probably the most complex ones with the longest philosophical history, they are not the only conditions of responsibility. The next one to be discussed here is the question of power.

Power

Power as a precondition of responsibility is a mix of the first two conditions, of causality and freedom. When we say that power is a condition of responsibility, that usually means that the subject had the power to do whatever responsibility is ascribed for at any particular moment. This includes some sort of causality, and it also implies that the subject did it voluntarily, that she did it according to her free will. Another way of saying this is that having control over something is the condition for being held responsible for it (Nida-R ¼melin, 1998, p. 31).

Power as a precondition of responsibility is a wide notion including capabilities, personal qualities, and so forth. It is not restricted to political power in the sense of being able to make others do one s bidding. However, this sense is also included. One can find that power in the narrower sense also has an influence and that most authors agree that more power leads to more responsibility (cf. Lenk & Maring, 1990, p. 50; Birnbacher, 1995, p. 152). The correlation of power and responsibility corresponds to what was said earlier on about the correlation of responsibility and technology. The fact that technology increases our ability to do things, to cause changes, and also to know about these changes, makes it a prime reason for an increase in responsibility (cf. Lenk, 1997, p. 77; K ¼ng, 1997, p. 99; Ricoeur, 1995c, p. 64).

It is of course hard to practically determine who had the power to do what and whose doing the result was. Therefore one can often find the condition of power expressed negatively, that is to say that a lack of power is accepted as an excuse from responsibility (cf. De George, 1999, p. 121). you cannot be held responsible for something over which you have no control (Staddon, 1999, p. 146). Fain suggests that for someone to be held solely responsible, that person not only brought about the result, but it was in his power to prevent the situation from occurring (Fain, 1972, p. 29).

Apart from causality and freedom, power is usually presumed to be linked to personal attributes of the subject that would enable him or her to make use of this power. There is a host of these attributes that will be discussed in the next section.

Personal Qualities

We all know that responsibility ascriptions differ with personal qualities of the subject. If I shoot my neighbour, I will most probably be held accountable for that, whereas this would probably be different if my one-year-old son did the same thing. The reason for this is that an adult and a child have different personality traits that are relevant for being a subject of responsibility. There are different abilities that a person needs in order to become responsible, and according to Kaufman (1995, p. 88), these can be grouped in moral, cognitive, and communicative abilities . Without wanting to rank these attributes, we will just briefly describe them, starting with emotions.

Emotions play a complex role in responsibility depending on the prevailing theory. For theorists like Strawson or Wallace (cf. Wallace, 1996, p. 8) who see responsibility as a moral sentiment, and according to Stocker, also for Aristotle it is constituted by, and in this sense secondary to, certain emotions and emotional reactions (Stocker, 1999, p. 196). Other theories of responsibility also have a link to emotions. Humans often need to cross some personal threshold before they start acting responsibly. That threshold can be a personal concern that changes the perception of a social dilemma to a personal one (Bierhoff, 1995, p. 225) and it is linked to emotion. As long as we do not feel that we are concerned , we will not be willing to assume responsibility.

Another point where emotions are of relevance is the social utility of responsibility. Going back to the purpose of responsibility, one finds that a frequently named reason for invoking it is social change. This social change is to be engendered by affecting the motivation of agents, which in turn is done by giving incentives to do or refrain from doing certain things. This intellectual exercise can only be fruitful if there is some sort of transmission that brings it to action. The transmission mechanism in question can be emotion. There is no use in the agent knowing that a certain punishment awaits her for some behaviour if this knowledge cannot impress her and make her act accordingly . The following quote that concerns emotion s role in ethics in general can be transferred to responsibility as well:

Indeed, emotions serve specific roles in the moral life. As expressive capacities , they function as a medium by which we signal our moral interest to others, such as when we express compassion when a friend suffers, or grief at her loss; as sensitivities, they are the receptors or antennae by which we, in turn, pick up signals of moral salience or interest. They are the modes by which we track moral relevance. (Sherman, 1999, p. 297)

Furthermore, one can be held responsible for one s emotions. There are emotions that are viewed as positive, such as love, mercy, or compassion, and emotions that are regarded as negative, such as hate, anger, or ingratitude. ^[10]

The subject must dispose of more than just emotions. Another important property is intention. Nature and degree of intention influence the outcome of responsibility ascriptions. Sanctions will differ depending on whether I shot my neighbour on purpose or whether I did it accidentally while cleaning my gun. Intention relates back to the other conditions already discussed. It implies causality, knowledge of this causality, and the freedom to change things. The discourse concerning intentions is well developed in the legal tradition where determining and judging intentions are an important part of the process of imposing sanctions. It is therefore little wonder that a philosopher of law like Hart uses an elaborate theory of intentions:

Intention is to be divided into three related parts , to which I shall give what I hope are three self-explanatory names . The first I shall call ˜intentionally doing something , the second ˜doing something with a further intention , and the third ˜bare intention because it is the case of intending to do something in the future without doing anything to execute this intention now. (Hart, 1968, p. 117)

The legal aspect of intentions is often discussed under the heading of mens rea (a guilty mind), which in Anglo-American law seems to be a condition of criminal proceedings . According to Hart (1968, p. 114), the mens rea requirement includes all of the mental or intellectual elements that the perpetrator must fulfil, but intention is the central one among them. Long (1999, p. 129) sees the difference between responsible and non-responsible wrongdoers in the presence of malicious thoughts in the former. Intention, however, is not limited to the legal concept of responsibility but applies to all other sorts as well. More generally, We say that some person P is responsible for the outcome O of an action A, when P has intentionally caused A in order to achieve O (Collste, 2000a, p. 125). According to R tzer (1998, p. 18), responsibility, in a strong sense of the word, means that in addition to being causally responsible, one must have intended something knowing the potential consequences.

Intention as a condition of responsibility may be a plausible thought, and the examples given above explain why this is so. On the other hand intention is also a serious problem for the ascription of responsibility because it is by definition unknown. All those cases in which responsibility is ascribed to the subject by someone other than the subject itself have to deal with the problem that they have to rely on the subject s description of her intentions or on her behaviour. Both may be misleading. The subject can lie in order to escape prosecution , she may be unaware of the real reasons for her actions or her motivations, and intentions may be too complex to be adequately understood. Behaviour can be a good guide for the judgment of intention, but it can also be misleading. When I threaten you by pointing my gun at you and saying that I will now shoot you, this may express my intention to murder you, but it can also mean that I am kidding with you and that I think the gun is not loaded and you probably know that. Any responsibility ascription that takes the intention of the subject into account is therefore also an act of judgment and it is not least of all for that reason fallible.

Another problem of intention as a necessary condition of responsibility is that in many cases where responsibility might be called for, there is no discernible intention to be found. As we have seen, the rise of the concept of responsibility is closely linked to the technological development and the moral problems resulting from that. Neither these developments in general, nor the spectacular occurrences that stirred up discussion about technology, are clearly intended by anybody. If intention is perceived as necessary for responsibility, then we can disregard the concept for technological catastrophes such as Chernobyl, Bhopal, Three Mile Island, or the Challenger explosion. It would make even less sense to search for responsibility for the underlying development of technology and resulting problems such as the depletion of the ozone layer or the greenhouse effect. Intention is thus another condition that has to deal with theoretical and practical problems and whose stipulation can endanger the ascription of responsibility.

The condition of intention implies the next condition, which is knowledge. It makes no sense to speak of the intention of a subject if he of she has no knowledge of the expected outcome of an action (cf. Hubig, 1995b, p. 62; Wehowsky, 1999, p. 94; McGary, 1991, p. 83; Hassemer, 1990, p. 64). To hold an individual responsible for an action requires that he be aware of the nature of the action, in the sense that he is not doing A in the belief that he is doing B (Held, 1991, p. 90). The knowledge required for responsibility, however, goes beyond knowing what one is doing at a given moment. Knowing about the results of an action requires foresight, judgment, and again contains an irreducible aspect of contingency. Another sort of knowledge required is the knowledge of the relevant rules. If the subject is not aware of moral, legal, or other standards, then she may not realise that something she is doing can result in responsibility ascriptions which may endanger the validity of these ascriptions. On the other hand a lack of awareness does not always preclude responsibility from being ascribed. Someone coming from a cannibalistic society might not realise that in Western democracies, it is generally regarded as bad practice to eat the dead. He would still have to suffer consequences from such an action even if these consequences might be of a different quality from the ones a native citizen of such a society would have to put up with.

While knowledge is a condition of responsibility, it is also an object of responsibility. We not only need knowledge to be able to act responsibly, it is also a duty arising from responsibility to acquire the necessary knowledge to be able to act responsibly. We can thus also be held responsible for a lack of knowledge. This goes for individuals who can be required to acquire certain kinds of knowledge for certain roles, and it can also be true for societies that may feel an obligation to fund research to gain knowledge.

The inverse relationship between knowledge and responsibility is valid as well. Ignorance may be an excusing condition for responsibility, but it can also be the object of responsibility. Some sorts of responsibility that are linked to certain roles require the holder of the position to keep up with current knowledge, and a lack to do so can lead to the ascription of responsibility. This is true for all of the positions that carry an expert status. One area where this status is of great importance is the area of technology in general and thus also the area of computing and IT.

Special knowledge coupled with the power of position gives computer professionals efficacy; they can do things in the world which others cannot. Thus, they have greater responsibility than others. (Johnson, 2001, p. 73)

Ignorance, finally, can be divided into several groups. First of all there is the fundamental ignorance defined by proofs of impossibility to know something. Secondly, there is ignorance concerning the results that can be overcome by learning of the pertaining facts and relationships. The third form of ignorance is that where the results are known in principle, but unknown to the subject in question. The fourth form is that there is knowledge of results, but not all of the causes may be known (cf. Zimmerli, 1994a, p. 184 ). These different forms of ignorance might be judged differently and lead to different results in responsibility ascriptions.

Another condition the subject has to fulfil is that it needs to have the faculty of reason. Reason can be understood in the Kantian sense as enabling us to see the world as a world of freedom and an infinite task (Schwartl nder, 1991, p. 19). Reason in this sense is what enables us to conceive of the categorical imperative and to follow it. Other concepts of ethics and morality also build upon reason. It can be understood as the faculty that allows us to develop our idea of the good life, which we then follow. Reason is one of a multitude of character traits and capacities that a subject of responsibility is often understood to need. In Wallace s view of responsibility, it is the powers of reflective self-control (Wallace, 1996, p. 190) that the agent needs in order to be a subject. For H ffe (1995, p. 293), it is a combination of reason, sensitivity, and conscientiousness, without which responsibility remains a wish.

Many of the traits that the subject should have are usually ascribed to persons. For this reason some authors say that only persons can be subjects of responsibility. We can assign responsibility only to persons (Baier, 1972, p. 50). All of the other conditions of responsibility discussed so far can be subsumed under the idea that they are necessary to make us persons. Freedom, for example, is not an end in itself, but it is freedom that makes us persons or morally accountable agents, which is the reason why we long for it and why it is necessary for responsibility (cf. Wallace, 1996, p. 3). Apart from the conditions already mentioned, there are others that need to be fulfilled to allow us to speak of persons and to consider moral responsibility.

One ability a subject must have is that it can render the ascription of responsibility useful in the sense that the purpose of the ascription can be fulfilled. What this is depends on the definition of the purpose but, as we have seen, responsibility often stands for the attempt to bring about social change by administering sanctions. If those sanctions or punishment are to fulfil their role, then the subject must dispose of the mental capacity to understand them and react accordingly. Wallace describes these rational powers that he calls powers of reflective self-control as follows: the power to grasp and apply the principles that support the moral obligations we accept, and to control one s behaviour by the light of such principles (Wallace, 1996, p. 188). This requirement relates back to the faculty of reason, but also to questions of freedom and causality. The subject must be able to learn from the consequences of her behaviour, which in turn means that she must be aware of the causal relationships (Staddon, 1999, p. 150). To stay within the terms introduced earlier on, the subject must be able to understand the economy of threats and to act accordingly.

This comprises two aspects that lead to consequences, the mental aspect and the physical aspect. In order to be able to act accordingly, the subject not only needs to meet the conditions mentioned so far, most of which are part of her nature, but she also needs a personality, developed through the process of socialisation, education, and formation. The development of the personality includes pedagogical aims, among them the training or instilling of a sense of responsibility (Hager, 1990, p. 60f; Dehner, 1998, p. 95; B hret, 1987, p. 6). Education holds a central point in the enabling of individuals to be moral or ethical agents (Severson, 1997, p. 26). It is thus also necessary to engender humans with the ability to act responsibly. Since this development is part of the long process of socialisation of an individual, it is part of the responsibility of the group or society in question, but we will return to this later.

Most of what was said so far refers to the mental capacities of the subject, to the mens rea requirement of responsibility. There is, however, also another side to it. Mens rea in itself is not punishable by law, and in moral philosophy it is contentious whether a guilty mind in itself should and could be the basis of a responsibility ascription. Usually there also needs to be a guilty deed that needs to have been done; it is this what is also called the actus reus requirement. In relation to the condition that the subject must fulfil, this can be interpreted to mean that only the mental and physical unity that characterises human beings can be the basis of responsibility (Velasquez, 1991, p. 114). Only a being with both of these aspects can act freely knowing about the limits to its freedom and can therefore be responsible.

This should suffice as a discussion of the conditions of responsibility. One thing should have become clear by now, that is that these conditions are hard if not impossible to meet. They require a huge amount of abilities on the part of the subject. It is questionable whether anyone can fulfil all of them, but it seems clear that they are too high an aim to be attained even by the majority of persons who so far have been regarded as the typical subjects. The next step will therefore be an analysis of who or what is generally understood to be possible subjects. This discussion will underline the problems of responsibility. It will become clear that the traditional subjects of responsibility are often unable to fulfil the conditions, while other entities that might be able to fulfil the conditions are not recognised as subjects.

The Subject of Responsibility

Who is the subject of responsibility? There are many potential answers to this question, but one of them is clearly the traditional and the most widely accepted one: it is the human individual. One also often hears that it is the person. Let us start with the human individual. Many authors define responsibility and human individuals as mutually dependent. Responsibility belongs essentially to the individual (Lewis, 1991, p. 17; similar: Rapp, 1991, p. 23). On the other hand the human individual is characterised by his or her capacity to be responsible (Lenk, 1994, p. 43; Neuberg, 1997b, p. 2). From here it only seems to be a short step to say: The subject of responsibility is the person. However, the concept of the person is a complex one and has led to a number of problems for theories of responsibility. It is therefore worthwhile to spend a little time on the analysis of the concept of the person and personhood.

The Person as Subject of Responsibility

One can find an interdependency between person and responsibility similar to the one between individual and responsibility. We can assign responsibility only to persons only persons who have failed in this responsibility can be responsible for something that has happened (Baier, 1972, p. 50). While this sounds convincing it remains to be seen what exactly a person is. The term is derived from the Latin persona and originally stood for the dramatis personae . It then became a term used in Roman law where it referred to anything that could act on either side of a dispute (cf. French, 1979). The concept developed other meanings and today can be interpreted to stand for unity and duration, rational agency, and self-awareness (French, 1979). The person is one of the possible functions of the ego (next to ontological unifier and ultimate agent) and according to Flynn (1984, p. 10) is the centre of moral ascriptions. This, it seems, is the most important function of the notion of a person. It is a construct that fulfils the conditions of responsibility discussed earlier. Responsibility is only possible if past actions and their results can be ascribed to the subject. This requires that successive acts are linked and that they can be understood as acts of the same actor. The term person is supposed to achieve this goal. It is a legal and moral construct that is necessary in order to identify the substance that remains throughout the sequence of acts which is necessary to hold a subject responsible (Bayertz, 1995a, p. 13). The construct of the person also stands for the property of freedom. It is defined as free in the sense that it can decide against the given alternatives, that it can transcend the given reality (Priddat, 1994, p. 272).

Persons are created or constructed for certain ends, and according to these ends there are different sorts of persons, the most important ones being the legal person and the moral person. The most important one in terms of factual relevance is the legal person. While there may be discussion about the metaphysical status of legal persons, they are clearly defined by law. The legal person is an entity that is subject to legal rights or obligation. Legal personhood is always something conferred, never merely the result of the act or acts of parties (French, 1992, p. 134). Legal persons are not necessarily natural persons. Companies, organisations, or states can be subjects of legal ascriptions, which is generally recognised and does not produce philosophical problems. A more controversial point is the moral person, which can be defined analogous to the legal person but which runs into philosophical objections as we will show later on. Before discussing the philosophical problems of moral and legal persons, we will take a brief look at two examples of individual moral responsibility that are relevant to this discussion. The examples are the scientist and the engineer/technician and their personal responsibility. For questions relating to responsibility in IS these, two can be considered important because they embody the groups of people who are usually considered responsible for IT, the computer professionals. Computer professionals tend to have a scientific background and work in a technical environment. Furthermore, many of the traditional approaches to ethical problems of computers and information technology specifically address this group and hope to be able to solve the problems by guiding the individuals who work with computers according to ethical principles.

Since technology is often understood to be the application of science, it makes sense in the case of responsibility to look at the responsibility of the scientist first and then at the responsibility of the engineer. An ethics of science cannot be reduced to questions of the responsibility of the individual scientist. However, for now we will limit ourselves to this aspect because the aim of this analysisistoshowtheshortcomingsofindividualresponsibility.Concerningthe responsibility of the scientist, there are two possible extreme positions. Either the scientist is completely responsible for the results of his or her research, or he or she is not responsible at all. The latter position is based on the idea of the amorality of science. Science is understood to be a good thing, to further progress and welfare, and thus not to be in need of further justification. A scientist working in this generally positive environment does not have to worry about ethical implication of research. Prof. L. F. Fieser, one of the codevelopers of Napalm, is a good example of this position. He stated in the New York Times (27 December, 1967, quoted in Lenk, 1991b, p. 7) that he regarded his job as completely technical and that he did not consider ethical implications. In his opinion it is the distinction between development and use of scientific results that counts. The scientists therefore bear no responsibility for what happens with their research, only those who use these results. The example of a military researcher dealing with weapons of mass destruction was of course chosen because it is not convincing. It is easy to construct even more drastic examples of scientists doing nothing but research and still being morally responsible. A physicist helping a dictator to build nuclear bombs or a chemist helping a terrorist organisation to develop chemical weapons would probably be considered cases in point even by the staunchest defenders of a moral freedom of science. Once we recognise, however, that scientists can be responsible for the results of their work, we run into the problem of determining who is responsible for what. If the physicist working for the dictator is morally responsible for a potential use of these weapons, how about the physicist doing the same sort of work in a democracy? And what about Otto Hahn, who discovered nuclear fission? One possible way out that some scientists use is the distinction between basic research and applied research. One can never know what the results of basic research will be and therefore there can be no responsibility for the results of basic research. Applied research on the other hand has a clear goal, and at least the intended effects of this research can be attributed to the involved scientists. However, the boundary between basic and applied research is blurring. In fields like genetics or computer science, the two are hard to distinguish. Meyer-Abich (in Lenk, 1991b, p. 26), for example, argues that the nuclear bomb was a direct result of basic research and that therefore those scientists who collaborated on the Manhattan Project are responsible for the dead of Hiroshima and Nagasaki.

Therefore, the other possible extreme solution would be to hold scientists responsible for everything that results from their work. Carl Friedrich von Weizs cker, one of the German scientists working in nuclear physics under Hitler and arguably one of those who were responsible for Germany s failure to develop a nuclear bomb, believes that scientists are responsible for the results of their research, if not in legal terms, then at least in moral terms (von Weizs cker, 1991, p. 95). The reason for this responsibility is that the scientist is never just a scientist. He or she is always a human being and part of society, and as such he or she has the normal moral duty to consider the result of actions with regard to their impacts on other people s lives (von Weizs cker, 1964, p. 8).

While this argument appears convincing to many contemporaries, who, in the light of the problems and destruction caused by science, believe that there should be someone responsible, it is at the same time limited. The fundamental fact that the future is unknown, in combination with the limits of individual knowledge, preclude the individual scientist from being responsible for all of the results of his or her work. It is not plausible that Otto Hahn is quintessentially responsible for not only Hiroshima and Nagasaki, but also for the subsequent cold war and any future use of nuclear weapons. Neither would anyone hold Berthold Schwarz, according to legend the inventor of gunpowder in Europe, responsible for the millions of victims of guns over the last few centuries.

The answer to the question of the responsibility of the scientist is somewhere in between the two extremes of complete responsibility and no responsibility at all. Most authors agree on this fact. However, they do not at all agree where in the spectrum between the two extremes the true solution is supposed to be (cf. Lenk, 1984, p. 102; Hegselmann, 1991, p. 218; Mohr, 1995, p. 105; Rapp, 1994, p. 157). An interesting fact concerning scientific responsibility is that the code of scientific conduct in fact seems to be the only sort of morality that has worldwide validity. Scientists can therefore be said at least to be responsible for the adherence to this code (Eibl-Eibesfeld, 1997, p. 149). Also, this existence of a moral code should make scientists prime contenders for an overarching responsibility. If scientists are recognised to be unable to fulfil responsibilities, then what other groups should be able to do so?

Another group that might be considered a prime example of responsibility are those people whose work builds on science, the technicians or engineers. Among them one can find some of the potential subjects of responsibility in or for IS. For this group the same sort of arguments counts as for scientists, with the exception that they do not have the potential excuse of doing basic research. It is generally recognised that engineers have some sort of responsibility for the results of their work. At least in their role as citizens and participants of public life, they share the same responsibilities as everybody else. This general sense of responsibility also extends to their professional role. On the other hand they have to deal with more severe limitations of their freedom of action. Engineers tend to be employed by large organisations, and their degrees of freedom within these organisations are usually fewer than those of scientists. Therefore they may lack the ability to assume responsibility even in those cases in which they have the necessary knowledge (Ropohl, 1987, p. 162; 1996, p. 114). Again, as in the case of the scientist, one can conclude that the engineer is neither completely responsible for the results of his work nor is he absolved of every responsibility. In order to better realise responsibility, many authors have therefore suggested to stop concentrating exclusively on the individual and to look instead at the organisation that the individual is working in.

We not only need moral people. Even more importantly we need moral structures and organizations. Only by paying more attention to these can we adequately resolve the questions of the ethical responsibility of engineers in large organizations. (De George, 1991, p. 164)

Others, however, fear that this may lead to the decline of the idea of individual responsibility towards an ethics of institutions, and that this development would miss the nature of ethics and morality (Zimmerli, 1991, p. 79). In order to understand the two positions, we will use the next section describing the strengths and weaknesses of individual responsibility and the exclusive definition of the person as subject.

Strengths and Weaknesses of Individual Responsibility

The reason why individuals are generally seen as subjects of responsibility is that only they seem to fulfil most of the conditions discussed above. Individuals have the personal qualities required such as emotions, intentions, the faculty of reason, self-control, a propensity to react to punishment, and so forth. Many of these attributes can be summarised as having a conscience, which is one of the most frequently named reasons for individual responsibility (cf. Wehowsky, 1999, p. 21; Etchegoyen, 1999, p. 55). Closely related to the conscience is the ability that is the root of responsibility, the ability to answer. Only humans act morally because they recognise the other as a moral subject, and only they feel the need to answer and thus to assume responsibility. Furthermore, individual human beings are generally seen to be free in their will and action. This is not to deny the factual constraints under which everybody lives, but it implies that everybody can at least imagine alternatives to his or her actions that might enable them to act differently.

The strongest argument, however, for a limitation of responsibility to natural persons or individuals is of a metaphysical nature and pertains to the definition of the term. The argument is that only human beings have the metaphysical status to be able to assume or be ascribed responsibility. Therefore the only definition of the subject of responsibility that makes sense is the one defining the individual as subject (cf. Zimmerli, 1987, p. 107; 1991, p. 86; 1994a, p. 183; 1994b, 12; Breil, 1993, p. 230; Downy, 1972, p. 69f). ^[11] If this argument is true, then it does not make any sense to even discuss the possibility of other potential subjects. The problem with this point of view is that the individual does not fulfil all of the conditions. In fact, life in modern societies seems to contradict the idea that individuals are responsible for what happens. Individuals also do not have the capacity to satisfactorily deal with the consequences of responsibility ascriptions.

The classical model of responsibility that admits only individuals runs into severe problems in a world where organisations, institutions, or other collectives become the relevant actors (cf. Bayertz, 1995a, p. 53). Individuals are not able to deal with the most important problems of responsibility that we face today because they lack the knowledge of causal chains and the power to do anything about them. Speaking of global problems such as hunger relief or environmental protection, French (1992, p. 79) says: Individuals are virtually impotent with respect to any of these problems, despite the best of intentions. Irrespective of questions of responsibility, many authors agree that the important agents are no longer individual human beings. Human beings still act within organisations, their responsibilities, however are usually limited to their tasks within these organisations. Especially with regard to the design and development of complex technologies, individuals can no longer know the results of their action and thus lack the prerequisite of being held responsible. This is certainly true for the area of the design of computer software and hardware. The single programmer who works in a large company is not able to assess the functionality of her source code in the context of the entire program. It is this development that Hubig (1993, p. 282; 1995, p. 98; 1995b, p. 61) calls the loss of the subject.

In some ways this development may even appear to increase the possibility of responsibility. Human beings in modern industrial societies are freer than they have ever been. Economically and sociologically they dispose of degrees of freedom unknown before. Since freedom is a condition of responsibility, one might think that this development should be positive for the practice of responsibility ascriptions. On the other hand this new freedom also signals a loss of individual importance. Human beings can be replaced in their functional roles in organisations. Individuals are but cogwheels in the greater gearboxes of social life (Wehowsky, 1999, p. 20). Anyone can be replaced at any time. The positions endure (French, 1992, p. 143).

The loss of the subject, the decrease of apparent importance of individual human beings in societal decision making, is frequently described with the model of systems. Systems are understood to be complex entities that consist of human beings, but develop their own logic and are relatively independent of single individuals. If individual responsibility is possible at all with respect to systems, then this refers to responsibility within the system (Hubig, 1995, p. 98). The software engineer can thus be responsible for the reliability of the software she is programming, but she cannot assume responsibility for the social system constructed around computer use.

The systems in society are not only an expression of complexity; they also contribute to it. Different systems are interrelated and influence one another in ways that are hard to understand for the individual. This is another reason why institutions and organisations become the prevalent decision makers (Lenk & Maring, 1995, p. 243). Unlike individuals, they can increase their capacity of acquiring and dealing with knowledge.

This description of modern society is not as new as it may seem. The question whether the individual can be responsible is probably as old as moral philosophy. Furthermore, the idea of social systems taking over and leaving humankind powerless is also not confined to our days. For Marx (1998) the economy was the sole reason for social developments, and the individual, whatever he or she did, could not change the flow of events caused by economic laws.

There are several possible ways to react to the dilemma of individuals losing their role as agents. The first solution is to recognise that collectives are the relevant agents, while still insisting on individuals as subjects of responsibility. Not surprisingly, this solution is chosen by those authors who see responsibility confined by definition to the individual (Sinn & Zimmerli, 1986, p. 35). This approach, however, not only seems to be too narrow to be useful, it can even be seen as explicitly dangerous (Hengsbach, 1993, p. 19). Responsibility ascriptions looking exclusively at the individual necessarily overlook most of the important moral and social problems we face today. If we want to hold on to the idea of responsibility as a moral notion, then we should enable it to deal with structural questions and problems of collective actors(Str ker,1996,p.19).

Apart from these practical problems of individual responsibility, there are theoretical reasons why the use of the term person as the sole possible subject of responsibility is not satisfactory. The first group of objections against the concept of the person is that it is opaque . The person, as we have seen, is the main competitor of the individual in terms of being a subject of responsibility. However, one should only use the term person with highest caution. The problem is that the term is ambiguous, multifunctional, and leads to a tendency to mix questions of facts and values. Locke s legal understanding of the person is that the human being can embody different persons within a lifetime that are responsible for different actions. Human beings thus only exist in certain phases as persons: neither the newborn nor the old person in a state of dementia is a person. Collste (1998, p. 86) criticises Fletcher s criteria of personhood that would have the same effect of excluding some human beings that do not meet criteria of rationality or self-awareness. On the other hand, some Catholic moral theologians tend to see the fertilised egg as a person. It is thus enough to have the characteristic human genes. Another example of problematic definitions of the persons that no longer seem viable is given by Feinberg (1991, p. 63), who reminds us that the fiction of conjugal unity was only given up by American law in the 1840s. Before that, women did not have a separate legal identity from their husbands.

The flip side of this uncertainty regarding the definition of the person is that it is often defined as the being that is capable of being ascribed moral actions (cf. Kant, 1990). The person and responsibility are conditions of one another; they depend on their mutual existence (Ilting, 1994b, p. 96). According to Habermas (1998a, p. 107), culture, society, and person each presuppose the others. In effect, we have to admit that the person is nothing but a social construct with the explicit (if maybe not sole) aim of ascribing responsibility. Therefore, saying that persons are potential subjects of responsibility is true, but it is also tautological. It does not help us determine who persons are and which sort of entity can be considered a person. In the next section we shall see how this uncertainty concerning the concept of the person has led to the development of a highly controversial theory of moral personhood that might also include collective subjects.

Collective Responsibility

The problem with collective responsibility is that the two sides ”those who favour it and those who are against it ”are opposed in a way akin to what one can usually find between the factions of a religious war (cf. Ropohl, 1994a, p. 187). The idea of collective responsibility is not new at all. The biblical idea of the original sin is certainly an example of early concepts developed in this direction. Collective responsibility in this sense can be seen as the norm rather than the exception in less developed cultures. One could therefore argue that the discussion about collective responsibility is a step backward in our cultural development. However, the classical ideas and paradigms of collective responsibility, as embodied in the doctrine of the original sin, are not helpful for our modern problems (cf. Neuberg, 1997c, p. 254). We therefore see new developments in this field that are based on the hope of establishing a new sort of ascription that would include collectives as subjects.

The reasons for this development are the same as those discussed in the last section as the problems of individual responsibility. The central problem is the loss of the subject. The individual can neither change the course of the world, nor can he or she generally even change the course of a company. Engineers do not decide the development of IT and users do not decide the development of the Internet or infrastructure. Even those developments that can be traced back to a single decision have to deal with the problem of who made the decision. Teams, groups, commissions, and so forth make most important decisions. Reducing these decisions to individual acts is in most cases neither possible nor plausible. The problem becomes worse when one looks at developments that cannot be traced back to clear decisions. Neither environmental problems nor political developments nor technological progress are subject to any one decision that might have been altered . Nobody who bought a refrigerator in the 1970s did this deciding that the ozone layer should be destroyed by the chlorofluorocarbon. Nobody working on nuclear physics intended the cold war and the overkill capacities eventually produced by the two sides. And nobody using a computer intends the sweeping social changes that are a result of the general use. This is the reason why French (1992, p. 97) can claim that no sensible responsibility relationship can exist for this sort of problem.

Corporate Responsibility

The way out of this dilemma seems to be the expansion of the notion of responsibility to collective subjects. Collective responsibility is the kind of answerability incumbent either on social wholes or on individuals by virtue of membership in such ensembles (Flynn, 1984, p. 124f). This expansion alone, however, does not solve the problem either. As the quote indicates, there are different sorts of collective subjects and thus of collective responsibility, each with their own theories and difficulties. The sort of collective responsibility that is most widely discussed in the literature and that has the greatest impact on most of the questions relating to businesses or information technology is usually called corporate responsibility. The idea behind it, which may seem natural to many of us, is that there is a type of collective entity, embodied by the form of organisation that is usually called a corporation, that can be held responsible by itself. While this idea is anything but new in the legal sphere, it is still a controversial issue when it comes to moral responsibility.

In order to be able to talk about corporate responsibility, it is imperative to define what a corporation is. A first approach might be to see it as systems that first of all try to insure its own survival (cf. Apel, 1988, p. 208). According to Velasquez (1991, p. 123), there are three things we have in mind when we speak of corporations: (a) the fictitious legal entity, (b) the organization as a structured set of relationships, and (c) the organization as a set of human beings. For Werhane (1985, p. 31) a corporation is an association given legal status by a state charter to operate as a single unit with limited liability over an indefinite period of time. A corporation is originally created by a group of individuals for a specific purpose or purposes.

Corporations can also be defined as purposeful systems with a governance structure (Maring, 2001, p. 112). It is clear that corporations consist of human beings and without these human beings they would cease to exist. On the other hand corporations seem to exhibit signs of their own will and in many cases they keep on being even after all of their members have changed. The same collection of individuals that collaborate within the corporation would not produce the same outcome if they worked together randomly , without the umbrella of the corporation (Maring, 1989, p. 37). Many authors agree that corporations, even though they consist of human beings, can still be said to have the ability to act (Werhane, 1985, p. 50; Maring, 2001, p. 133f).

We have tried to show elsewhere (Stahl, 2000) that the fundamental question that decides whether we admit corporations as responsibility subjects is in fact a metaphysical one. Authors who admit any sort of metaphysical existence of corporations independent of their members tend to hold them responsible. Those theorists that propose a reductionist view, that is to say, believe every aspect of a corporation can be reduced without remainder to their individual members, generally do not believe that they can be held responsible. In view of the factual power of corporations and their apparent agency, many philosophers go a middle way and concede that corporations have some sort of moral responsibility, but that this is not the same as individual human responsibility (cf. Lenk, 1998, p. 281; De George, 1999, p. 196).

There are in fact some good reasons why corporations may count as subjects of responsibility in cases where individuals are no longer able to act in that role. Corporations are frequently better prepared to fulfil the conditions of responsibility discussed earlier. In many cases they have the necessary power and they make the relevant decisions. Depending on one s point of view, one can discern in corporations the preconditions such as intention, projection, the ability to decide and to act (Bayertz, 1995b, p. 55). They not only have more power and clearer decision algorithms than individual humans, they also have a longer life span (B hret, 1987, p. 9). Political institutions, just like economic organisations, may survive their members by centuries. They may therefore be able to assume responsibility that would not have a human counterpart left. One example of this are the payments made by German companies and the German government to former slave labourers from the Second World War, 60 years after the war. It would be hard if not impossible to find out which human might be responsible, but it is comparatively easy to see which company benefited from the slave labour, and most of the companies accepted the responsibility and made the payments, even though none of the former decision makers are still left in power.

A counterargument against corporate or organisational responsibility is that corporations cannot be moral because they do not have a conscience and therefore do not have the ability to deal with moral facts and arguments. The opposite is true, however. First of all, corporations can react to moral facts and interpretations, as many of the scandals in recent years show. Shell s initiative of publishing a triple-bottom-line report, dealing with the financial, social, and ecological balance sheet, shows that the moral pressure that was brought to bear upon Shell did influence the decision process deeply. Institutions and organisations can at least fear moral discourses (Hubig, 1995, p. 108). Corporations can even be better in fulfilling moral obligations because they have more options in dealing with realities and their abstractions (cf.Wiesenthal, 1990, p. 72). Therefore Goodpaster and Matthews (1982, p. 133) can say: A corporation can and should have a conscience. The language of ethics does have a place in the vocabulary of an organization.

Another condition of responsibility that corporations fulfil is the ability to answer. Corporations can communicate, and for many of them communication is even the main line of business. If they can communicate, they can answer to specific questions of specific stakeholders and these questions include moral ones. This answerability is the basis for some concepts of corporate responsibility (cf. Homann & Blome-Drees, 1992, p. 171).

Answering is a particular sort of action and if the answerability of the corporation is admitted, then it is only a short step to the admittance of general agency of corporations. As already mentioned, the assumptions of agency by corporations is closely linked to a tacit underlying metaphysical theory of corporations. Many authors, however, prefer not to discuss the metaphysical side and concentrate on the more easily observed aspect of acting on the part of corporations. Werhane (1985) or Lenk and Maring (1995) base their theories of corporate responsibility on the corporation s ability to act. This solution allows the introduction of corporate responsibility without having to discuss the problems of metaphysics and personhood. Since agency also implies intention, it fulfils another one of the central preconditions. Agency as the foundation of corporate responsibility also solves another problem, namely the question of the underlying morality. Since corporations may exhibit intentions but their motives are complex and hard to discern, it is easiest to subject them to some sort of consequentialist moral norm as background of the responsibility ascription.

their actions, not their motives, are the proper object of moral evaluation. (De George, 1999, p. 196)

The criterion of being an agent is weaker than the one of being a person, and it allows the use of responsibility ascriptions (Maring, 1989, p. 38).

Some authors, notably Peter French, do not confine their concept of the corporation to the aspect of agency, but go further and describe it as a moral person. The starting point of French s and similar ideas is the legal person. The law treats corporations as persons because they have the relevant features of personhood, including a method for forming intentions and a capacity to act (May & Hoffman, 1991, p. 3). Legal persons are entities that have rights or are recognised in law. In the case of legal persons, it is clear that they are not something natural but social constructs.

Legal personhood is always something conferred, never merely the result of the act or acts of parties. (French, 1992, p. 134)

Legal persons are the subject of legal responsibility. Accordingly moral persons should be the subject of moral responsibility.

French compares legal and moral personhood, and concludes that both can be applied to corporations. Corporations can be said to have a personality that is not identical with their employees personalities (French, 1992, p. 140). Corporations can show intentions, make decisions, act responsibly, enter into relationships, and they can be punished. The emphasis in French s theory lies on the fact that corporations have something akin to intention. The mechanism by which they form intentions and realise them is called the Corporate Internal Decision (CID) structure. It is this CID that determines the personality of the corporation (French, 1992, p. 139). French speaks of persons as noneliminable subjects of responsibility ascriptions, and the corporation fits the description. In short, corporations can be full-fledged moral persons and have whatever privileges, rights, and duties as are, in the normal course of affairs, accorded to moral persons (French, 1979, p. 207).

Many philosophers have attacked this extreme view of the corporation as a moral person as being comparable to the human being. French himself no longer advocates it. The idea of corporations being morally responsible, however, has gained credibility and is now quite widespread. De George uses a Kantian argument, saying that corporations are not ends in themselves and therefore not moral persons. Nevertheless he concedes that they can act intentionally and thus be held morally responsible for their actions (De George, 1999, p. 195f).

Summarising the problem of corporate responsibility, one can say that the only really contentious issue is whether they can be held responsible in a moral sense. This question is basically a metaphysical one and is hard to solve by arguments because it deeply impacts on what we perceive as real and as fiction. However, corporate responsibility is the sort of collective responsibility that is easiest to handle. Firstly, we can find definitions of corporations; secondly, at least corporate legal responsibility is clearly established; and thirdly, corporate responsibility is part of our everyday way of dealing with responsibility issues. The statement that Microsoft is responsible for the functioning or non-functioning of their new system software does not need to be explained. In the area of software we may bemoan the fact that corporations do not live up to their perceived responsibilities, but that does not mean that we doubt their responsibilities. In fact, the opposite is true. We can only talk about the shortcoming of corporate responsibility because we collectively accept that the idea makes sense. Unfortunately, there are sorts of collective responsibility that are much harder to handle and that have a huge impact on questions of responsibility in business in general and with regard to the use of IT in particular.

Other Forms of Collective Responsibility

Corporate responsibility ”contentious as it may be ”has the advantage of being part of our everyday communication and can thus be seen as generally recognised. Furthermore, corporations not only have a social reality, but are also defined by law and therefore subject to clearly defined rules of legal responsibility. The same cannot be said for other forms of responsibility that result from collective action or are caused by collectives. The problem we are driving at here is the constitution of a collective subject that is usually called a group. Among groups one can distinguish between groups that have something like a collective consciousness or collective purpose and those that do not, that are thus purely random. This is a similar distinction to the one Held (1972) suggests between organised and unorganised groups. An organised group:

has a method for deciding to act: it has officials who can act in its name , or a voting procedure to arrive at its decisions, or customary procedures to guide its actions. The possession of such a decision method by a collection of individuals, is, we might say, that which transforms a collection of persons into an organised group or collectivity. (Held, 1972, p. 105)

An organised group that is subject according to this definition can be said to have an internal decision structure and could therefore be described using French s CID. A group collectively internalises responsibility when, but only when, members willingly take responsibility for themselves as a group (Schmidtz, 1998, p. 8). It can then be subjected to the same arguments as a corporation. Examples in our field of interest might be the members of a computer club, a citizens action committee on data privacy, or any special interest group with an appropriate internal structure.

The most important questions in the case of such group responsibility are:

1. How is the group constituted?

2. How far does the behaviour of group members differ from the individual behaviour of these persons?

3. What is the relationship between the responsibility of the group and the individualresponsibility?

Question 1 is important because, unlike in the case of corporations, this is generally unclear for groups. A group can have an enduring existence, but it can also be constituted for certain tasks. Most groups are less formal, so it may be difficult to determine whether they really have a decision structure. If groups have an internal decision structure, then they are able to act and thereby fulfil at least the requirement for being held responsible.

When a decision structure is not clearly visible, there may be other facts that hint at the possibility of a group s responsibility such as a common way of life (Cooper, 1972, p. 90). Another clue could be a shared culture that is produced by the group and accepted by its members (May, 1991, p. 246). This sort of group consists of individuals who voluntarily come together for some purpose, and who willingly and knowingly constitute the group. The problem with regard to responsibility that this type of group faces is similar to the problem of corporate responsibility. It is a question of metaphysics, whether one wants to recognise the independent existence of this sort of group or not. Many philosophers struggle with this question and try to find a way out that allows them to preserve their metaphysical views while at the same time allowing for the plausible claim that this sort of group can be responsible. L ¼bbe (1998, p. 127), for example, emphasises that collectives are nothing but the sum of their members, but at the same time she has to admit that there are cases where the result of group action is not the same as the action of the group s constituents. This is the reason why collective responsibility seems not only possible but even necessary. Individuals in groups act differently from those acting alone. There is, for example, a tendency to collectively accept risks that an individual would be loath to accept (cf. Lenk, 1994, p. 130). The reason for this increased acceptance of risks may in fact be the question of the distribution of responsibility. As members of a group, people feel that their actions have to be accounted for by all of the members, not only by them alone. The distribution of responsibility is therefore one of the criteria that Held offers for distinguishing organised from unorganised groups (Held, 1991, p. 97).

A collective responsibility in this sense is one that is distributed to each and every member of the group of individuals sharing it: the same responsibility falls to each and every one of them. (Goodin, 1998, p. 146f)

This sort of organised group can be considered a candidate of collective responsibility. Examples for such groups are all sorts of assemblies of people with similar interests who voluntarily come together and view themselves as a group.

More difficult to handle are groups that lack an internal structure. Examples in this area might be the group of the users of the Internet, users of home computers, etc. The members of these groups usually neither know the other members, nor do they care a great deal about them. The decisive difference between the two sorts of groups is that in organised groups, members are aware of their membership and consider themselves members, whereas in unorganised groups, membership is usually not voluntary and members do not consider themselves to be members. In organised groups the description of the member from the point of view of the environment corresponds to the members description of themselves, while this is not so for unorganised groups. Unorganised groups are usually defined from the outside, and that definition is controversial because it constitutes the group without taking into account the potential members views. Since this process is more or less random ” everybody is always part of an infinity of virtual groups that only come to existence by being named ”the entire process of assigning responsibility becomes apparently random and thus loses its moral validity (cf. Bates, 1991, p. 107). Group responsibility poses several serious problems that do not exist in the case of corporate responsibility. A central problem is that a causal ascription is hard, if not impossible, to realise if the results at which the ascription aims are cumulative (Bayertz, 1995b, p. 54). Examples could be the environmental degradation caused by millions of automobiles or the changes in society caused by the Internet users. Every single one of them can rightly say that his or her contribution to the problem is minimal and that without him or her the problem would remain the same. Briefly, in the case of group responsibility, it is hard to establish a convincing relationship between individual agents and the object of responsibility.Wenowhaveasufficientlydetailedinsightintothedifferentsorts of collective responsibility to summarise the problems it produces.

Problems of Collective Responsibility

The central problem for those who accept the possibility of collective responsibility is how to determine the relationship between individual and collective responsibility. Most of the authors who admit collective responsibility do so with the more or less explicit supposition that it is not the same thing as individual responsibility and that it should not replace individual responsibility either. Both aspects are generally seen to have their strengths and weaknesses, and they should complement one another (Mai, 1994, p. 166; Hubig, 1994, p. 157; Spinello, 1997, p. 26). If one agrees to this view and allows for collective responsibility while at the same time insisting on the irreducible individual responsibility that every citizen in a democratic state has to assume (Kettner, 1995, p. 321f), then the central question becomes: How is responsibility distributed between the individual and the collective?

Several models are offered in this regard. According to Cooper (1972, p. 86), there are three ways in which group responsibility can be related to individual responsibility. First, there is the case where the group is responsible but none of the individual members are. Second, some of the individuals can be responsible but this does not exhaust the responsibilities of the group. Finally, there is the tricky case where one does hold each member of a group individually responsible, yet still feels that the responsibility of the group is something more than the sum of all these individual responsibilities (Cooper, 1972, p. 86). Similar ideas can be found in other texts (cf. May & Hoffman, 1991, p. 4; Lenk, 1987, p. 127; Seebass, 2001, p. 86). This sort of idea makes sense because it reflects our everyday experience. When a company does something bad, it can be held responsible. At the same time there usually is someone within the company who is also personally responsible for the misdeed. On the other hand it is usually not plausible to condemn the entire company and every single one of its members personally . A distribution of responsibility between collectives and individuals is therefore a convincing idea.

The downside is that this theory does not provide us with clear instructions on how to deal with the distribution of responsibility. First, there is the possibility that the individual, even though she might have knowingly and willingly caused something, can hide behind the responsibility of the collective. Viable collective responsibility would have to guard against this possibility (B hret, 1987, p. 9; Maring, 1989, p. 32; Lenk & Maring, 1996, p. 240). Second, there is the pragmatic objection that it will be impossible to determine what part of the responsibility should be ascribed to the individual and what part to the collective. To this there does not seem to be an easy answer because any answer would have to rely on a complete theory of responsibility which, if it is possible at all, could only be applicable to a specific case because the relevant dimensions change from case to case.

Summarising the problem of collective responsibility, one can say that it is confronted with several difficulties. Originally, the concept of responsibility was developed with the individual person in mind (Homann, 1998, p. IX; Breil, 1993, p. 230). Overcoming collective responsibility was in fact a progress in moral theory and social practice. The notion of responsibility indicates the individual should be the subject. At the same time, the social developments which led to the importance of responsibility seem to marginalise the individual. Relevant actors today are increasingly collective actors. But even if one admits the concept of collective responsibility, there are no easy answers. Collectives as subjects of responsibility can again be divided into different groups. Among the collectives that are candidates for the role of a subject of responsibility, the most prominent one is the corporation. Firstly, corporations and organisations are the most visible actors in today s societies, and secondly corporations have a precedent in responsibility; they are established as subjects in legal responsibility. This is the reason why most literature on the subject refers to this case. However, there are other collectives that also act, whose acts have results but that are not usually recognised as subjects. In the case of these more or less organised groups, the theory of responsibility runs into even worse problems than in the case of corporations. The most important questions here are: Is collective responsibility just an excuse for the individual who does not want to assume her responsibility? and Even if collective responsibility is admitted in these cases, how should we model the relationship of individual and collective responsibility?

Where does this leave the question of responsibility and business information-technology? It seems that the entire question of trying to find a subject that is responsible for IS is highly contentious. On the one hand we have individuals who might be responsible for IS. These might include programmers, hardware specialists, systems analysts and management from IT management, and the CIO (chief information officer) to general management and the CEO.

However, none of these individual candidates fulfils the conditions. They usually do not have the necessary knowledge about what consequences their actions will have. They lack the power to stop what is happening and they are frequently changed so that there is no personal continuity to the roles. The only ones for whom there might be exceptions to this problem are the highest level of management, who may have power and the opportunity to know about and change things. These few individuals at the top, however, have the problem that they have to deal with a multitude of facts and aspects, not only from the IT area, so that they quickly come to the limits of their capacities.

On the other hand there are collective actors that can overcome these problems. First of all there are companies ” producers of software, hardware, and services ”that make their living in the area of IS. These without a doubt have a great influence on what is happening in the market, and they also have good intelligence concerning the results of their actions and other future developments. While these corporations seem the ideal candidates for being subjects, they lack some of the qualities that are often considered necessary. It is unclear whether they can really act or whether collective actions are in fact nothing but the aggregate of individual actions. Also, they lack the personal qualities usually associated with responsibility. Corporations have no personality, no conscience, and no feelings, which is why many scholars do not want to accept them as subjects.

Finally we have potential collective subjects that are even harder to grasp than corporations. There are groups such as the group of computer users, hackers, data workers, or information professionals that might be described as being responsible for IS. Among these groups there are some with a higher degree of organisation such as the computer professionals and some that have no discernible organisation at all such as the group of computer users. These groups too can be said to be responsible for some aspects of IS, but in their case it is even harder to see what they are responsible for and which consequences such an ascription might have.

There seems to be a great number of candidates for responsibility in IS, but not a single one of them is able to fulfil all of the conditions. Using the classic theory of responsibility developed so far, responsibility seems to be a rather useless concept in IS because there does not appear to be anyone or anything that could rightly and bindingly be held responsible in the sense that this ascription would be valid and viable.

For the sake of completeness, we will dedicate the next section to a last candidate of being as subject, a candidate that obviously lacks some of the conditions but that nevertheless has been discussed in the literature and that may develop with technological progress.

Computers and Information Systems as Subjects of Responsibility

The case for computers as subjects of responsibility could be made similar to the case for collective subjects. The most important aspect is that computers appear to make decisions that might be considered objects of responsibility. Some of the most convincing arguments along this line come from the military. The modern military is highly dependent on information systems for reconnaissance, assessment of the situation, and correct reaction. Nuclear forces everywhere have an extremely short warning time, and reaction to threats must be immediate. Human beings are not able to react as speedily and are not able to process the necessary information in the same amount of time. So, in fact computers make decisions or they at least prepare the information that serves as the basis of decisions. While the military is the most obvious example of this, many other information systems serve similar tasks in more mundane environments. Expert systems, decision support systems, or executive support systems are all examples of information systems that either make decisions or at least strongly influence decisions.

Apart from this fact, the next argument that may be used to ascribe responsibility to computers is that it is often impossible for humans to understand how and why the decisions made by computers came about. This is due to the size and complexity of modern computer systems. The attempt to reduce computer decisions to the acts of individual human beings is doomed to failure. Even if it were possible to identify the line of source code that caused a certain decision, it is impossible to forecast how exactly any piece of software would interact with other programmes. Thus, Weizenbaum (1976, p. 239) can say:

no human is responsible at all for the computer s output. The enormous computer systems in the Pentagon and their counterparts elsewhere in our culture have, in a very real sense, no authors.

Johnson (2001, p. 188) identifies the same factors, scale, and complexity of computers, the many hands involved in their development, and the mediation of human decision making as factors contributing to a diffusion of responsibility.

From the weaknesses of traditional models of responsibility and the facts just described, Bechtel has drawn the conclusion that one might consider computers and information systems as potential subjects of responsibility with whom humans might share responsibility (Bechtel, 1985, p. 297). The main thrust of Bechtel s argument is that computers are flexible and might be constructed to be adaptable. If they were they might be able to learn and to act appropriately to their environment. If computers ever reach this stage of development, it would appear appropriate to Bechtel to assign responsibility to them. Other authors go even further in their assumptions of what computers can or cannot do. For Stewart (1997) it is conceivable that computers take over the function of judges in criminal courts. Computers and expert systems might be used to evaluate evidence and to come to conclusions as to who the perpetrator was and what the fitting punishment is. These arguments for information systems as responsible agents hinge on whether they can fulfil roles that were designed to be filled by humans. This in turn leads us back to metaphysical and anthropological questions, and the question of what it is that makes humans subjects of responsibility. For diehard behaviourists this question is easily answered . For them it is a function of input and output. If something reacts in a certain way to given stimuli, then this is the base of judging the qualities of that entity. That means that if a computer passes the Turing test, which means that it is impossible to determine whether answers given originate from a human being or from a computer, then the two are functionally equal, and consequentially, if humans can be responsible, the computer should be so as well (cf. Lenk, 1994, p. 51).

There are also good reasons against the admittance of computers as subjects of responsibility. The first one is the definition of responsibility as pertaining only to humans. The second one is based on that and could be described as a general unease resulting from the elevation of computers to subject status. While the sentence the computer is responsible for is one that can frequently be heard , few people would go so far as to describe information systems as morally responsible. In most cases this sort of exclamation stands for the exasperation resulting from a malfunction of computers and at most means causal responsibility. Finally, computers are even less able to fulfil the manifold conditions of responsibility than human beings. Their knowledge base is limited (even if it may be vast) and in most cases they are not able to learn (yet). They may have some power in the sense that they influence events and they are part of causal chains, but they cannot intentionally change the course of events. Clearly computers lack freedom and personal characteristics. While freedom is a contentious issue for humans, it is clear that it is lacking in computers. A computer s reactions are determined. Furthermore computers do not have the necessary feelings, fear of punishment, empathy, or reason, all of which have been cited as conditions of responsibility. For all of these reasons, it seems ludicrous to speak of (moral) responsibility of computers. However, there are aspects to the topic that might warrant attention. Firstly, the development of computers is progressing at such a speed that it seems hard to predict where exactly it is heading. Who is to say that computers never will be able to develop self-awareness, consciousness, and maybe even a conscience? ^[12] It is a recurring theme of science fiction literature that technical creations develop a life of their own and that humans have to deal with it despite their initial lack of willingness to do so. From the world-dominating computer systems in the Terminator movies, to the friendly if somewhat unworldly android Data in the Star Trek series, we find examples of the topic. We so far lack any sort of categorical tools for dealing with this which in itself may turn out to be a problem of responsibility one day.

There is one conclusion one can draw from the question of responsibility and computers at this point. The development and design of computers and information systems should reflect the need for responsibility. This refers to responsibility within the design process, but even more to questions of clarity, accountability, and the distribution of responsibilities between users and computers (cf. Bechtel, 1985, p. 297; Lenk, 1998, p. 469).

Having talked at length about the subject of responsibility, the intermediate conclusion we can draw at this point is that responsibility and IS do not seem to give clear instructions or hints concerning concrete behaviour. The combination of conditions and problems, and the different candidates for the status of being subjects all seem to be more or less contradictory. However, the subject is only one of the central dimensions of responsibility. In the next section we will take a closer look at the second most important dimension ”the object of responsibility.

The Object of Responsibility

The object is that which the subject is responsible for. As diverse as the different candidates for the status of being a subject were, they are outdone by the complexity that one finds when looking at the object. Object and subject are closely related. Depending on who or what one admits is the subject, the potential objects vary. The candidates for being an object of responsibility in IS are legion. The functioning of a computer program, the accuracy of data, the organisational change within a company, and questions of the quality of life are potential candidates, as well as the global information infrastructure or the use of potentially faulty information systems in nuclear defence systems.

In this section we will start by recalling some of the conditions that a potential object has to fulfil in order to be admissible as an object. After that we will discuss a few typical objects as they can be found in the literature, and finally we will take a look at some of the fundamental difficulties that are generated by responsibility ascriptions on the side of the object.

Conditions for Being an Object of Responsibility

These conditions mirror the general conditions of responsibility as discussed in Section 5.2.1. Some of the properties that the subject should have must be reflected in the object in order for an ascription of responsibility to become viable. Sinn and Zimmerli (1986, p. 35) develop a list of criteria that an object of responsibility would have to fulfil in order for an ascription to make sense. These criteria limit what is admissible as an object to things that are within the power of a subject. Among this list, one finds a temporal limitation to objects that do not exceed the human life span. Also the order of magnitude of an object and its intensity must be manageable. Furthermore the foreseeable results of the object in temporal, technical, social, and all other respects must be assessable and limited. In other words, the subject must have the object more or less well under control.

But what is it that responsibility is eventually ascribed for? Usually it is the result of an action. To begin with, the action itself can be seen as the object of responsibility. Following Apel (1988, p. 132), an action can only be described as such by looking at the results, be they intended or unintended. In this sense the term action implies responsibility. On the other hand not all of the results of human actions are always thought to be objects of responsibility. In fact many authors try to limit responsibility to those actions and their consequences that were intended.

Certain conditions have to be fulfilled for being responsible for an action. An agent has an intention and through acting[,] he/she wants to realise the intention. The consequences of the act are supposed to realise the intention. (Collste. 2000b, p. 190; cf. Spaemann, 1975, p. 324)

De George speaks of doing an action knowingly and willingly, which means that the subject had a choice, knew what it was doing, and did it deliberately (De George, 1999, p. 111). Similarly, Wallace emphasises the aspect of choice as crucial for being an object. This choice does not have to be a rational or conscious one, it is only necessary that the agent has the power of self-control that would have facilitated this choice as a conscious one (Wallace, 1996, p. 190).

What one can see reflected in these conditions is the realisation of the general conditions discussed earlier. The aspects of action, choice, and intention point towards the questions of causality, freedom, and power, as well as the human agent s ability to deal with them. As will be shown later, these conditions are as controversial on the level of the object of responsibility as they were on the general level. In order to show how these problems manifest themselves, the next section will be dedicated to a discussion of some of the most important and frequently discussed objects.

Some Objects of Responsibility

What is the object of responsibility? In the most general sense, it is the other. Responsibility, whatever the subject, whatever the circumstances or the specific objective, always aims at improving social life, and by doing so aims to improve life for all of the members of society or at least the majority of those individualsthataredirectlyinvolvedintheascription.Inthissenseresponsibility relates directly to the French tradition of moral philosophy and its emphasis on the other as the underlying reason of morality. Responsibility for the other can mean that no harm be done to her, that her dignity be respected, that her utility be maximised, etc. Starting from these general observations, one can also try to distil more specific objectives that relate to morality, such as advancement of peace , environment, or the chances of future generations (Zimmerli, 1994a, p. 7). One particular example of the other that one can also be responsible for is the self. The first object that the subject can be held responsible for is its own capacity of being a subject. This is a thought that is prevalent in the French moral philosophy and that has consequences for any theory of responsibility.

According to Weil (1998; 1960, p. 152), I am responsible for who or what I am at present. If I am not who I think I ought to be, then I should have decided to change; if I am lacking knowledge necessary for being a responsible agent, then I would have had to acquire it. I am responsible for my own conduct. In a more fundamental way, Sartre (1997, p. 105) sees us as responsible because we are sentenced to freedom, and we are responsible not only for us but because we create the world and we are thus responsible for everything. While Sartre s view is somewhat extreme because it endangers the concept of responsibility by overburdening it, the idea of responsibility for oneself is certainly part of our everyday understanding of responsibility. We hold people responsible for what they are and that includes not only their action but even more so their character and their personality. We praise individuals for their compassion, think less of them for their ingratitude or hatred, reproach self-righteousness and unjust anger (Sherman, 1999, p. 294).

While the first object of responsibility is the individual human being, the person, and particularly the person who is the subject, another frequently named object is the aggregation of all of these human objects, the human race of the existence of humans. This object is especially relevant because it is closely linked to technology and the development of the concept of responsibility. Human survival is one of the newer objects of responsibility. In most of the classical philosophical approaches to morality, one finds considerations of the individual, the person, or the self. None of them deals with the question of humanity in general because it was unalterably given. The conceivable end of the human race used to be in the hands of God and therefore out of the reach of philosophers. Since that situation has changed with the advent of modern technology, especially with the nuclear threat and in the light of World War II (Tugendhat, 1990b, p. 6), moral philosophy has had to add it to the list of responsibility objects.

Modern theories of responsibility tend to give a considerable amount of attention to human existence as an object of responsibility. Hans Jonas (1984), for example, who started the responsibility debate in the German speaking area, saw it as the fundamental problem that a morality of responsibility would have to address. Many scholars and maybe even more laypeople agree that the preservation of humanity is a high, if not the highest, ethical goal. It is thus our responsibility to ensure that human existence is not endangered. Plausible as this may seem, however, it is hard to justify from a philosophical point of view. Human existence is a fact, and it is always a problem to draw conclusions about norms from facts. One of the solutions to this dilemma offered by Jonas is that the highest norm is that there should be the possibility of norms. Thus the first responsibility is that there be responsibility.

Ignoring the subtleties of this debate, one can certainly count on broad support when stating that human existence must be protected. In the 1970s and 80s, questions relating to this were usually asked in the frame of the cold war and the possible nuclear annihilation of humanity. While this is technically still as much of a possibility today as it was fifteen years ago, the threat seems less probable from a political point of view and the situation is not perceived as threatening any more by the large majority of the population.

Today, when we speak of threats to humanity, we think more of potential changes of the nature of mankind, of environmental problems, and subsequently of threats to future generations. The first threat is that to the nature of mankind, and it is posed by progress in science, especially in genetics. Having deciphered or at least spelled out the genetic code of humans, it is only a question of time until we will be able to modify our genes according to our wishes. Where this will lead eventually is completely open, but there is no doubt that this development has the potential to change humanity at its roots. Genetic control of humanity is therefore one of the main reasons for Jonas demand for a new responsibility (Jonas, 1987, p. 162). While the consequences of this are not yet clear, it seems obvious that humanity has recognised its own nature as an object of responsibility. This corresponds to the responsibility for the self on the individuallevel.

Another point where humankind slowly sees that it must take responsibility is the area of the environment. Environmental degradation, global warming, the depletion of the ozone layer, problems of the provision of drinking water, etc., make themselves felt all over the world. Even though ecocide is an old phenomenon that even Plato complained about (cf. Mohr, 1995, p. 32) and even though it seems to be part of human nature (Eibl-Eibesfeld, 1997, p. 38), overpopulation and shrinking resources make the problem more urgent. Again, the protection of the environment, just like human existence itself, is not a very contentious object of responsibility. The exact content and justification of environmental protection, however, are very much subject of debate. The arguments can be divided into anthropocentric and physiocentric ones. Anthropocentric arguments say that we must protect the environment because failing to do so will hurt or even wipe out ourselves. It is thus a matter of pure selfinterest to make sure the environment is in order (Breil, 1993, p. 202; Lenk, 1994, p. 38; Lenk & Maring, 1995, p. 245). Another anthropocentric defence of environmental protection is a perceived Kantian duty that we have towards ourselves in view of nature (Birnbacher, 1980, p. 111f; Tribe, 1980, p. 36; Hubig, 1995, p. 154).

This sort of argument is based on a dichotomy between man and nature, between culture and nature, that does not convince everybody. Firstly, our nature and environment are by now so much shaped by mankind that the discrepancy between culture and nature does not seem relevant (Hastedt, 1994, p. 202). Secondly, humans are themselves part of nature and therefore everything we do can be considered natural. Furthermore, many people are unhappy with the anthropocentric defence of environmental protection because it treats the environment as a means to our ends, while they have a perception of nature as having intrinsic value. This remoralised view of nature (van den Daele, 1993, p. 173) contradicts the classical description of nature as merciless and indifferent. It is usually expressed as part of a theological or transcendent worldview where creation is holy and mankind has no right to interfere with the process of creation (cf. Schweitzer, 1991, p. 120; Devall, 1997, p. 23; Ehrenfeld, 1997, p. 173). Such a transcendent foundation of environmental protection may sound convincing to many of us, but it has the disadvantage of being binding only to those who share the transcendent presuppositions .

Depending on the justification, the idea environmental protection leads to vastly different implementations . If the environment only needs to be protected for the sake of mankind, then trade-offs are possible between the environment and other human needs or wants. A holy nature as an expression of God s creation would have to be treated differently.

Another object of responsibility that is accepted as unanimously as nature and that leads to difficulties at least as serious is the future, more exactly future human generations. ^[13] Future generations are linked to the environment because they are often invoked as a justification of environmental protection. Traditional ethics never had to worry much about future humans because it was certain that they would come and there was nothing one could do for them. Future generations are thus a relatively new object of responsibility comparable to the environment. And as in the case of the environment, the argument for the protection of future humans is deeply plausible. Human beings will procreate as all species do. Since there is no reason to think that today s humans are better or more privileged than yesterday s or tomorrow s, we must assume that they will have the same rights and obligations. Among them is the right to a happy and healthy existence, which presupposes that we leave them a world in which a dignified existence is possible.

When one looks at the argument, however, one finds that that it is hard to specify what exactly is meant by future generations and how they can be an object of responsibility. It starts with the fact that it is impossible to give reasons for a right to live for someone who has not even been conceived yet. If such an individual right cannot be found, then how do we justify the right of a collective? But even if one concedes that right, the next problem is that one of the central features of morality, reciprocity, cannot exist with future generations. In the case of the immediately succeeding generations, there is the reciprocity that is evident in the classical roles. The parents raise their children until they are independent, and later on the children care for the parents when they are in need. An analogous relationship to humans more than two or three generations away is not possible. So if there should be responsibility for future generations, it would have to be iterative, from generation to generation.

Another problem is that we do not know what the needs and wants of future generations will be and how we can cater to them. The assumption is usually that their nature will be like ours and that they will thus want to live in a world similar to ours. However, it is obvious that this argument is not valid. If it were true, then we should never have evolved from cavemen, and yet most of us prefer living in a world with hospitals and telephones. So, purely conservative arguments using future generations as reasons to slow or stop development are misleading because they make assumptions that cannot be verified . These problems explain why French (1992, p. 98) can say: It is very difficult to see how any relationship deep enough to support responsibility assignments can exist between any persons existing today and distant generations.

The environment as well as future generations combine epistemic and normative uncertainty when used as responsibility objects. They are thus two examples for a typical property of responsibility ascriptions in modern society, the property of risk. During the introduction of the concept of responsibility, it was pointed out that the term responsibility can be understood as a reaction to the impotence of traditional moralities when faced with risk. Since risk played a part in the ascension of the concept of responsibility in public discourse, it is not surprising that one sometimes hears the suggestion that risk is an object of responsibility or even the prototypical object where the concept of responsibility can show its powers (Bayertz, 1995a, p. 46). The idea of responsibility for risks, however, refers to a lot of the principal problems of responsibility: Who is responsible for risks, which actions or results must be considered risky, to what point do we consider risks, is it possible to take responsibility for risks that others must take, how can the ascription of responsibility for social risks be realised ? These questions relate back to the objects just discussed. Can we ascribe responsibility for risking the well-being of future generations or even for risking human existence? To better understand these questions, one should reflect on the problems that a potential object of responsibility can face.

Problems of the Object of Responsibility

The object of responsibility is relatively uncontroversial in all those cases where the subject intended something and this intended result is what responsibility is ascribed for. It becomes problematic if there is either no discernible subject for a clearly defined object, or if the object itself is either unclear or not intended. These cases are related to the conditions of responsibility, especially intention and knowledge. As we have seen, the growing complexity of social life is one of the main reasons for the success of the idea of responsibility. At the same time complexity threatens the very idea because it undermines the applicability of the conditions. If we say that a situation is complex, then this implies that causalities are unclear, that knowledge about it is uncertain , that outcomes cannot be assessed, and that consequentially a potential subject cannot know what will result from its actions. While the growing complexity leads to an increase of topics and areas that humans can be held responsible for (Krawietz, 1995, p. 185), it diminishes at the same time the ability to assume responsibility. In the risk society it becomes impossible to ascribe damages to any particular person because they lose their spatial-temporal limits. Potential damages reach proportions where it is impossible to compensate for them, and therefore there are no longer even plans to do so (Beck, 1998a, p. 79). This leads to the organised irresponsibility where individuals are no longer expected to assume responsibility because they would not be able to live up to it anyway (Wehowsky, 1999, p. 20). Three factors can be identified as the prime causes for this development: cumulative effects, side effects, and the problem of acting and omitting.

The cumulative effects have already been mentioned. Most of the facts that are seen as necessitating responsibility are of a collective nature and come about through collective action. The discussion of collective subjects of responsibility reflects this fact. A particularly difficult case is the problem of environmental protection. For our topic of business information technology, there are economic as well as technological problems that also originate from the collective nature of these processes. Our collective way of doing business, for example, leaves a considerable part of humanity in poverty. Some 30,000 children a day starve or die of curable diseases. On the other hand it is hard to point to any one individual whose action could be considered responsible for that. The use of information technology seems to promise a remedy for some of the questions of international distribution, but at the same time it produces a digital divide that deepens the distance between rich and poor. Again, it is not possible to say that any one user of IT should be held responsible for that. So, collective responsibility, especially responsibility without even a discernible collective subject, that is to say group responsibility, aggravates all of the problems of the object mentioned so far. It impedes the discovery of causalities, it hides the individual, it can mask problems of risk and uncertainty. However, it is not the only problem of the object.

Another problem that is at least as serious are side effects. It is clear that the subject is responsible for intended actions and their results (Ropohl, 1994b, p. 113). But how does an ethics of responsibility deal with unintended results? To answer this question it is imperative to first define these unintended results or side effects. They are presumably called side effects because they do not appear in the straight line from intention to final end, but materialise in some other area that can be defined as being beside the main point. Such a definition would recur exclusively on the subject s intention for determining which causes are effects and thus objects of responsibility, and which are side effects with an unclear status of responsibility. However, the exclusive limitation to the subject s intention is not sufficient. First, one may not know the true intentions of the subject and the subject may lie about them or be unaware of them. Second, the quality of the side effects can make them the primary result, whether intended or not. A nuclear power plant produces electricity that is used in a moment for purposes that tend to be forgotten in a minute, and at the same time it produces nuclear waste that will be radioactive and fatal for tens of thousands of years. Saying that the electricity is the main result and the nuclear waste a mere side effect seems euphemistic. Spaemann (1975, p. 323) points out that the dichotomy of intended and unintended results is not sufficient either. Every action can have at least three kinds of results: the intended ones, the unintended but known and accepted ones, and the unforeseen ones. In this sense the nuclear waste would be unintended but known, whereas the effect of the CFCs on the ozone layer was not only unintended, it was also completely unforeseen.

The question of which effects to ascribe to an action and thus to a subject goes back to other problems discussed earlier. Which are the causal chains that have to be taken into account, how can we determine them, and how can we have knowledge of them? Causal chains do not end, and even the most irrelevant action taken today may have serious repercussions later on.

Therefore, if responsibility meant being answerable for every consequence of an action, then this would sentence us to complete inactivity because the mere attempt to find out about all of the results is by definition an infinite endeavour. On the other hand it is easy to see that complete inaction would also produce results (at least we would all starve to death), which would also required the ascription of responsibility. This then leads us to the next big problem of the object of responsibility ”the question of doing and omitting.

It is clear that omitting to do something can have consequences that may in turn lead to responsibility ascriptions and to sanctions. If I have the opportunity to save someone from certain death without too much risk to myself , then I am expected to do so. If I omit doing so, then I am liable for punishment. However, there are problems with this model. The first one is that it is hard to determine who omitted to do what. Actions ascribed to an actor can usually be determined quite clearly through a description of physical acts that then lead to consequences. Omissions are more virtual. Depending on the point of view, an actor s activity might easily be described as typing on his computer. At the same time he can be said to omit making love, saving the environment, repairing his motorcycle, or reading a book. Omitting is virtual and thus potentially infinite. Another problem is that of causality. While, despite the theoretical problems with causality, it is clear that certain actions lead to given results, the same can often not be said about omissions. Omissions as negative entities seem to lack a relationship to causality (Picht, 1991, p. 31). Another problem is the question of intention in omissions. While intentional omissions are similar to actions in many ways, the inclusion of everything one does not do and does not even think about in the potential objects of responsibility again leads to an overstretching of the notion. However, many of the contemporary problems of responsibility result from such unconscious omissions and lack of attention. Cumulative effects threatening the environment, for example, are often omissions. We omit to think about the potential effects of using a car or we omit giving to charity, thereby producing results that we are unaware of.

This example shows that doing and omitting are a problem of definition. When is something an action, when is it an omission? Whenever I do something, I omit to do the opposite. So, what is the definition of omissions? Birnbacher (1995b, p. 24) suggests that action and omission should be defined as contrasting concepts; that means that an action cannot be an omission at the same time and vice versa. However, actions can contain omissions, and one can omit by acting. When I see someone drown, I can walk the other way, which is an action, and thereby omit to save the person. Since the consequences of omissions are often the same as those of actions, at least from a purely consequentialist viewpoint, it seems to make no difference whether the subject s actions or omissions led to the result (Bayertz, 1995b, p. 46). Some authors have drawn the conclusion that for responsibility purposes, actions and omissions are therefore equal (cf. Rohbeck, 1993, p. 278; Rachels, 1997, p. 195). On the other hand, the history of western ethics shows that there has always been a clear distinction between the two. The prohibition to do damage has always been stronger than the commandment to do good. This is, for example, one of the pillars of medical ethics since the times of Hippocrates (H ffe, 1995, p. 80).

While it may be more difficult to prove responsibility for an inaction or omission than for an action (Etchegoyen, 1999, p. 94), the most clearly defined example of responsibility, legal responsibility, shows that there is a difference in social evaluation between the two. Omissions are punishable by law, for example in the case of negligence, but they are punished differently from actions. But does this help us with questions of responsibility and business informationtechnology?

Information Technology as an Object of Responsibility

Technology in general and information technology in particular is an object that reflects all of the problems discussed. Technology has always been developed as a means, but it often develops results that were not intended and not foreseen (Rohbeck, 1993, p. 16). It is typical for modern technologies that their development is neither planned nor linear. Instead, typical characteristics are anonymity, inherent dynamism , and complexity (Mohr, 1995, p. 105). Technology produces results, but few of them are intended. In fact, the side effects are often the key reason for thinking about technology at all. The most important single problem that technology produces is probably the degradation of the environment. A considerable part of the literature dealing with philosophy of technology and technology assessment addresses this point. Technology was intended to improve our circumstances, make life easier and more comfortable, a task that it achieved in many respects (cf. Wild, 1991, p. 45). At the same time technology threatens the very basis of our existence. To make matters worse, our growing knowledge and power have led to situations where it is highly plausible that omissions can be more dangerous than actions and the threats we have to assess can come in equal parts from doing and omitting (L ¼bbe, 1998, p. 183). All of this has led to the call for responsibility in the area of technology and subsequent attempts to come to grips with the problems, to institutionalise some form or other of responsibility for technology. The entire area of technology assessment can be understood as such an attempt to realise responsibility for technology.

Information technology forms a part of this problem, but a part with growing importance. On the one hand IT is a necessity for most modern forms of production, and therefore the problems may not be caused by IT but they are aggravated by it. At the same time IT contributes to the epistemological side of the problem by allowing new forms of information gathering, modelling, and predictions . It thus helps transform side effects to known effects that responsibility has to be assumed for. The increasing use of IT also poses environmental problems of its own due to sometimes poisonous materials needed to produce it and the increasing energy consumption necessary to keep the computers running. Finally, the effects of IT on the social environment are probably the most serious ones and the least researched ones. What does it mean that people are available 24 hours a day, that the limits of work and leisure are blurring, that we communicate increasingly using electronic means, that networks promise us infinite information which at the same time exceeds our capacity to understand?

We will return to the question of information technology as an object of responsibility later on. Right now we are still in the process of demonstrating the difficulties of the classical concept of responsibility. It should have become clear in this section that determining an object is not only difficult in the sense that the participants of the process of ascription have to agree, but that there are also inherent problems that complicate the whole process. Before introducing a possible solution to this problem, the theory of reflective responsibility, we need to continue the discussion of the relevant aspects of responsibility, the next of which is the instance.

The Instance of Responsibility

The instance is the third classical dimension of responsibility, next to the subject and the object. It is more frequently mentioned in German literature about responsibility than in English texts. The idea of an instance, however, plays a role in most theories of responsibility. The instance is that entity which is external to the relationship between subject and object, and which at the same time has a fundamental influence on it in that it decides about the ascription. It might therefore also be called the authority responsibility. The subject is ascribed responsibility for the object before the instance. Since responsibility is a social process involving many potential parties, it makes sense that there be someone or something with the authority to finalise the ascription and the resulting sanctions. Some authors go so far as to suggest that responsibility can only make sense if it produces tangible results, which in turn seems to presuppose the existence of a powerful instance (L ¼bbe, 1985, p. 59).

The paradigm of responsibility that makes the instance most clearly visible is a legal court . The accused is the subject, the object is that which the subject is accused of and the instance is the judge. In most other types of responsibility, it is less clear who or what the instance might be. One can distinguish between internal and external instances (cf. Birnbacher, 1980, p. 111f; van Luijk, 1990, p. 41). The most frequently named external instance next to the judge is God. For a secular theory of responsibility, however, God as an instance produces the problem that He can only be effective for those who believe in him, a group which in modern societies does not constitute a sufficiently large majority to justify this sort of ascription. Furthermore, as soon as one speaks of God, a whole host of theological problems arise. A secular theory of responsibility should recognise God as a motivation for many agents to act in specific ways, but it cannot rely on him as a binding instance.

Since in most cases no external instance is clearly discernible, many theories of responsibility rely on an internal one, which in most cases is the conscience. While the conscience can be seen as an expression of the human ability to reflect upon oneself (Ropohl, 1994b, p. 113), most authors agree that it is also an internalised version of exterior instances. The human conscience can perform the function of a court of law; it is the internal acceptance of social rules and standards (Walzer, 1994, p. 396). The conscience has something to do with moral and legal rules (Feinberg, 1980, p. 141) and the way they are realised in a society. It can also be seen as God s representation in humans or, in philosophical terms, as a representation of something higher or transcendent (Lenk, 1998, p. 161). Unlike the external instances, however, the conscience is not completely rational. It contains a measure of emotions and of intuitive recognition of right and wrong. This intuitional knowledge of the right or good is then supposed to motivate the subject to act right (Kohlberg, 1995, p. 289). The conscience also disposes of the most important property that an instance of responsibility must possess; it can attribute sanctions. The so-called pangs of conscience or remorse about something one did are the internalised sanctions that are triggered by the conscience (Mill, 1976, p. 50).

The conscience as an instance of responsibility overcomes some of the problems of external instances, but it also runs into some problems of its own. First of all it is too uncertain a phenomenon to base a general theory on. Even though the structure of the conscience seems to be an anthropological fact, the content of that structure depends on circumstances. It can contain all sorts of normative contents, even the worst kind (Tugendhat, 1990a, p. 12). The interpersonal differences between consciences are too great to guarantee a halfway equal distribution of responsibilities. Even though responsibility is recognised, the conscience as instance cuts off public debate and relegates responsibility to the status of something purely private (H ffe, 1995, p. 26). This contradicts the social nature of responsibility. Wuneburger (1993, p. 58) points out that the conscience as the internalised form of social roles is also a kind of psychological group pressure that can be used to ensure the individual s reliance on the group.

Furthermore, we do not know whether all of the subjects of responsibility discussed so far in fact have a conscience. It is an open question whether, for example, corporations can have a conscience. If one says that corporations are morally responsible and that the conscience is the instance of responsibility, then one may be driven to the doubtful conclusion that a corporation can and should have a conscience (Goodpaster & Matthews, 1982, p. 133). Most philosophically minded authors, however, tend to locate the conscience in the individual human being (Etchegoyen, 1993, p. 177). Finally, despite all of the recognition and personal experience of the existence of our conscience, there is no more a proof of its existence than there is a proof of God (Wehowsky, 1999, p. 86).

Since neither external nor internal instance are without serious problems, another solution to the problem of the instance is some form of process. The idea here is that the effectiveness of responsibility ascriptions should not be made dependent on any entities, but on the process of ascription itself. One version of this idea is offered by Ulrich (1997), who builds on discourse ethics and tries to see how responsibility can be realised in social settings. In this model one acts responsibly by confronting all the considerations and claims by all of those who are affected and by recognising the legitimate claims. In this case it would be explicitly immoral to escape to the excuse of the individual conscience. Discourse ethics is seen as an expression of the universal moral point of view. Realising a discourse by discussing all of the relevant validity claims with everybody concerned in the absence of power, and under conditions as close to the ideal discourse as possible, will lead to the morally best outcome. While this model seems plausible and might even be convincing to other moral theories apart from discourse ethics, it has to deal with the problem of theory and practice. Real discourses will never be held under the legitimising conditions of the ideal discourse, and in the absence of these ideal conditions it is doubtful whether and in what way the results of discourses really can lay a claim to moral validity.

A morally less ambitious process for the realisation of responsibility is therefore necessary. One suggestion, put forward by H. Schmidt (1992, p. 170), aims specifically at the problem of responsibility in the technological age. Schmidt refers to the complexity, the rapidity, and the risks of technological development. His suggestion is to start a fundamental democratic process that would bring together the drivers of the development and its victims in order to facilitate a mediation between the two groups. This idea is based on morality s task of avoiding violence and arbitrating between conflicting parties that we have seen as the cornerstone of the French tradition of moral philosophy. In practice it would probably look very similar to Ulrich s discourses, but the theoretical goals are more modest and therefore more realistic. These different processual approaches are easily combined with the idea of the future as an instance. If there are neither transcendent nor internal instances, what is left as a basis for the ascription of responsibility? One possible answer is that it is the future (Kissling, 1995, p. 424), something we have already encountered as an object of responsibility as well. But how are we supposed to cater for future needs? According to which rules should we design ascription processes and what are the criteria for a successful ascription?

These questions are addressed by another group of theories which diverge from the approach chosen so far in that they do not posit the necessity of an instance at all but concentrate instead on the normative background of responsibility. Instead of the three basic dimensions discussed so far, one can also find the description of responsibility as a relation between an object, the qualification that makes it an object, and the good that was affected (Trigeaud, 1999, p. 92). This directs the attention away from the entities that play a role in the ascription toward the normative background. In fact, some of the more traditional approaches to responsibility move away from the instance as a necessary ingredient and replace it with a system of standards of judgment (Bayertz, 1995b, p. 15f).

It is also important to note that responsibility as a procedural notion is devoid of a clear content if the normative background is not defined.

Responsibility does not produce values or moral statements, it can only transport them. It is based on moral valuations that it cannot justify by itself. This is metaethically relevant because it shows that responsibility cannot be identical with a moral theory and is necessarily subordinated to a moral theory (Bayertz, 1995b, p. 65f). Therefore subjects can only be responsible if they have rules or norms to which they can refer (Ilting, 1994b, p. 96). In this sense it may be sufficient for a responsibility ascription if the moral principles that it relies upon are known (Hartmann, 1991, p. 99) and the question of whether there must be an instance may be irrelevant.

But even if a theory of responsibility refrains from relying on an instance, the next question would be whether an instance would not be necessary to determine which moral norms and rules are acceptable. Another problem is that of morality and ethics in the sense of the German tradition. Do we need a material morality for responsibility or rather an abstract ethics or both? And whichever way this question is decided, neither a morality nor an ethics that is universally acceptable is discernible right now. So, coming back to responsibility for technology or for business information technology, we face the problem that there is no generally accepted instance and there is not even a generally accepted set of rules according to which responsibility might be ascribed. This may be one of the reasons for the apparent shift of relevance away from ethics toward the law. Legal responsibility disposes of clear rules, at least within nation-states, and is therefore attributable and sanctionable. For moral responsibility this seems to be much less the case. If this is so, then the attempt to find a framework for responsibility in relation to IS seems a rather hopeless endeavour.

Thus far we have seen that in the case of each of the three basic dimensions of responsibility, one quickly runs into problems when attempting to simply describe them. Questions like who is responsible for IS; in IS, what is he responsible for; and before whom is he responsible, are difficult to answer. This is not only caused by the difficulties that the idea of business information technology itself contains, but even more because of principal problems of the concept of responsibility. However, the problems are not confined to the basic dimension of subject, object, and instance. Apart from the three basic dimensions, there are several other facts and determinants that also have to be taken into account for a comprehensive picture of responsibility.

Types of Responsibility

So far we have made several allusions to different types of responsibility. A complete theory of responsibility has taken into account that there are different sorts of responsibility that in some cases are similar and in other cases seem to be disparate or even contradictory. A lack of clarity concerning the type of responsibility can therefore result in theoretical confusion and practical inapplicability. While most authors in the field of responsibility would agree to this statement, it has not led to a generally accepted classification of different types of responsibility. Depending on the author, one can find between four (Hart, 1968, p. 211f; Johnson, 2001, p. 174) or five (Baier, 1972, p. 56) and up to ten (Lenk, 1998, p. 265f) different types. Typical candidates for types of responsibility that can be found in most of the texts are causal responsibility, task responsibility, role responsibility, legal responsibility, and moral responsibility.

Since causal responsibility was discussed earlier on, we will concentrate on the types of responsibility that are typically ascribed to agents. Most closely linked to agency is task responsibility, where the subject is held responsible for the fulfilling of a specific task (Ricoeur, 1994, p. 30). The subject of action needs to fulfil several conditions that are similar to the ones that the subject of responsibility has to fulfil, which is the reason why task responsibility seems to be a natural first step in the development of responsibility.

Closely related to task responsibility is the responsibility for recurrent tasks that are linked to the roles a subject plays. Whatever these roles may be and however the subject may have acquired them, it is a unifying feature of most different roles that they entail responsibilities. Whether the subject be a policeman, a mother, a CEO, or a computer programmer, all of these roles lead to certain expectations from the environment and to the ascription of responsibility. Role responsibility is interchangeable with duty and refers to what individuals are expected to do in virtue of one of their social roles (Johnson, 2001, p. 174). Those responsibilities that result from the subject s fulfilling a certain role are clearly recognisable and in daily life are probably the ones that can most often be felt. The problem with role responsibility is that it refers to other types of responsibility, mostly to legal and moral responsibility (Hart, 1968, p. 215) without being either one of them.

role responsibility is a limited concept useful for describing certain social and institutional obligations, but not useful as an absolute criterion for making ethical judgments. (Werhane, 1985, p. 106)

Role responsibility is the basis of professional ethics, an approach frequently suggested for dealing with ethical problems of IS.

The two most widely recognised and most frequently discussed types of responsibilityarelegalandmoralresponsibility.Legalresponsibilityistheorigin of the concept of responsibility, and it has the advantage of being most easily discernible and having the clearest structure. The entire notion of responsibility often seems to be fixed in its legal usage, where the ascription takes place between accused, plaintiff/prosecutor, and judge, and results in compensation or punishment (cf. Ricoeur, 1995a, p. 41). Responsibility is often thought of as primarily a legal concept (Long, 1999, p. 118).

However, responsibility is mostly used ”even in legal circumstances ” with at least an implicit reference to morality. Legal and moral responsibility are clearly related, but the exact nature of this relationship is not clear. Some theories state that legal responsibility, in order to be efficient, must be based on moral responsibility. A legal punishment that does not reflect the moral consensus of a society is not acceptable. On the other hand severe moral offences often seem to require a legal punishment (Neuberg, 1997, p. 3; Downy, 1972, p. 67). In the case of the most serious of moral misdemeanours such as murder, assault, rape, etc., this seems to be plausible, but it should not be understood to mean that moral and legal responsibility are fundamentally the same.

One difference between the two is that moral responsibility aims at reasons and causes, whereas legal responsibility is concerned with the results of actions (Vossenkuhl, 1991, p. 52). Another difference can be found in the sanctions that in the legal case are formally institutionalised, but in the moral case tend to be less clear (Collste, 2000c, 1 p. 26; Lenk & Maring, 1990, p. 98). This is also the reason for the ridicule that moral responsibility is sometimes subjected to due to its apparent powerlessness and its function as an excuse in those cases where the legal institutions fail (L ¼bbe, 1998, p. 155).

What then is moral responsibility? It is a type of ascription that is characterised by certain facts. First of all its moral nature implies something like a universal validity. It is equally effective for everybody regardless of particular circumstances. It not only applies to the roles that humans fulfil, but it applies to humans per se or to persons (Lenk, 1991a, p. 67; 1994, p. 128). This universality, in the German tradition usually defended by reference to the universal properties of reason, is also a weakness of moral responsibility because it often does not allow the development of sufficiently clear rules of ascription (Krawietz, 1995, p. 197). Moral responsibility is based on moral norms and rules, and it therefore can only accept potentially moral entities as subjects. One of the reasons for this is the instance which in the case of moral responsibility is often understood to be the individual conscience (Kreikebaum, 1996, p. 180; S nger, 1991, p. 36). The person or group that ascribes responsibility to the subject must extend a certain sort of emotion toward the subject. This is usually described using the word blame. Moral responsibility entails an element of blameworthiness (Johnson, 2001, p. 175; French, 1992, p. 80; Wallace, 1996, p. 56). Most authors see moral responsibility as one of a multitude of different sorts. Depending on the definition of ethics and morality, however, the scope and scale of moral responsibility differs greatly. If one follows the definitions of morality given earlier on as anything that affects human coexistence, anything that has to do with the way people interact, and especially with conflicts and their peaceful resolution, then every responsibility ascription is moral. Responsibility can then be classified as moral, independent of whether it is at the same time of a legal, political, institutional, or any other nature.

Apart from these distinctions, one can also find others in the literature. Corresponding to causality, responsibility can be seen as direct or indirect (Ladd, 1992, p. 293). Responsibility can be negative or positive, which corresponds to the distinction between action and omission discussed earlier on (Velasquez, 1991, p. 111). Some authors see a difference between natural and contractual responsibility, where natural responsibility is usually taken to originate from a non-negotiable role or position. As a prototype for natural responsibility, the example of parental responsibility is often given (Etchegoyen, 1993). If the definition of responsibility as a construct of ascription is accepted, however, then there is no such thing as natural responsibility. A last type of responsibility that one can find in the literature is meta-responsibility, which stands for the responsibility for the conditions of the possibility of responsibility. This idea is the basis of Jonas (1984) theory of responsibility, and it can claim a certain plausibility. The problem, however, is that it is not accessible from within any theory of responsibility. The responsibility for the possibility of responsibilityissimilartothefinalfoundationofmoralityortothequestion:Why be moral? This is a question that does not seem to be answerable from within moral theory. We will therefore not discuss it here, but start at the point where the subjects assume or are ascribed responsibility without considering the final reason for that.

But there are still a few distinctions concerning responsibility left that need to be mentioned in order to supply something like a comprehensive overview of the existing theories.

The Temporal Aspect of Responsibility

Ethics has something to do with time, and the same is true for responsibility. What the relationship between time and responsibility is depends on the definition of time and the theory of responsibility. For Jonas, responsibility is the moral complement of our human ontological state as temporal beings (Jonas, 1984, p. 198). This existentialist idea refers to Jonas attempt to give a final foundation of responsibility that we just discussed. The problem with discussing questions concerning time is that it is one of the most complex issues philosophy has ever addressed. Agustinus (Confessiones X, 27, quoted in Bourdil, 1996, p. 28) said that whenever nobody asks him about time, he clearly knows what it is, but as soon as he wants to answer a question about it, he no longer knows. The same problem still applies today. Is time something objective or subjective? Does it exist or is it just a category of perception? These are open questions we will not discuss here. There seems to be enough of an interpersonal understanding of time for us to coexist and to coordinate our actions. In this sense one can say that time is objective and this objective time is of relevance for responsibility ascriptions.

Such an objective time is usually divided into past, present, and future. With the present being only the infinitesimal border between past and future, we are left with responsibility aiming either in the direction of the past ( responsibility ex post ) or in the direction of the future (responsibility ex ante ). Depending on the temporal direction of responsibility, the concept is filled with differing contents. Classically, responsibility means that the subject has to answer for damages, injustices, and errors that it committed (Ricoeur, 1991b, p. 282). Accordingly, responsibility ex post aims at negative events and looks for their human initiator. These negative events, being in the past, can be clearly described and used for the purpose of imputing punishment. Most of the responsibility theories aiming at sanctions imply that the deeds for which punishment is meted out are in the past (Wallace, 1996, p. 56; Hart, 1968, p. 160). Since legal responsibility has as one of its objectives to impose punishments, there is a close link between legal responsibility and responsibility ex post . Retrospective responsibility is thought to be crucial to legal judgment, primarily because the defendant s responsibility, or lack thereof, for committing an offence is thought to determine whether, and to what extent, it is permissible to inflict legal sanction or punishment (Long, 1999, p. 125).

While backward-looking responsibility is usually clearly defined and has a relatively long history, many authors today share the perception that forwardlooking responsibility is gaining in importance (Birnbacher, 1995, p. 146; Ricoeur, 1990a, p. 341; Jonas, 1984). In this case, the object of responsibility is either in the future or at least it will extend into the future. Responsibility ex ante is characterised by several facts or common interpretations that distinguish it from ex post . Firstly, it is positive rather than negative with regard to the object. While the subject is usually punished ex post for a misdeed, the object ex ante is often a positive state. That can either be the avoidance of damages or the production of something desirable (Bayertz, 1995b, p. 45). The desirable future state and the corresponding positive sanctions are meant to serve as a (positive) motivation to the subject rather than the deterrent ex post (Schlick, 1930, p. 33; Neuberg, 1997, p. 9). Accordingly, the duties resulting from responsibility ex ante are generally duties relating to action rather than duties of omission (Birnbacher, 1995, p. 151). The inherent uncertainty of future development demands less reliance on external supervision and instead a higher degree of internal identification with the object. internalise responsibility when we take responsibility for the future (Schmidtz, 1998, p. 10). Finally, where responsibility ex post seems to point towards legal proceedings, responsibility ex ante is clearly closer to ethics and morality. If the subject has to give an answer concerning the object, that means it must think about it beforehand. This at least implies that the subject accepts its obligation to consider the other as a person and show its respect by giving her a reason for the action. Also, the horizon of the future points most clearly toward the general moral good (Ricoeur, 1990a, p. 341). Morality contains the element of duty, and duty as something one has to do directs the attention towards the future (Etchegoyen, 1993, p. 22).

The two temporal aspects have a close connection. While being held responsible often hints at the past and at legal sanctions, the purpose of this is usually to be found in results it will produce in the future. Most of the purposes of responsibility ex post can best be justified by their capacity to influence future developments. The economy of threats, for example, punishes past deeds with the explicit aim of deterring future perpetrators. Thus responsibility ex post forces us to assume responsibility ex ante (cf. French, 1992, p. 15; Vanberg, 1990, p. 94; Schmidtz, 1998, p. 11; Ricoeur, 1991a, p. 282; Goodin, 1998, p. 150). Beu & Buckley (2001, p. 62) express the same idea by saying: Knowing you will be held to account in this way for what you have done, in the past, may well make you do better, in [the] future. Some authors therefore believe that the main function of any sort of liability is prospective (Seebass, 2001, p. 89).

On the other hand there is also a reverse relationship. If the knowledge of sanctions forces the subject to assume responsibility ex ante, then it is in many cases hard to see how responsibility can be ascribed ex post if the subject had no knowledge of it ex ante (Birnbacher, 1995, p. 145). This refers to some of the conditions of responsibility, namely knowledge and power. These imply that the subject must at least be aware of the fact that its actions may lead to a responsibility ascription.

The prospective aspect of responsibility also refers to another problem which has led to the prominence of responsibility, to the problem of risk. Risk is an ex ante concept which expresses the degree of uncertainty of possible events (Holzheu, 1993, p. 265). The relationship between risk and responsibility has several aspects. Firstly, responsibility for future events is always uncertain and therefore risky (Richardson, 1999, p. 221). Secondly, the future is the realm of the possible where the idea of freedom makes sense (Picht, 1991, p. 28). Thirdly, the interplay between the two temporal directions and the corresponding relationship between the two types of responsibility finds its analogy in risk as well. From a personal perspective, risk is what lies in the future, what is to come. Systemically, risk is something that has to be dealt with after it happened, thus ex post . Society has to decide not only how to deal with risks that will come, but also how to distribute the damages of past risks. Again, the two aspects refer to one another (Bon , 1995, p. 31).

Depending on the type and temporal direction of responsibility, there are different mechanisms of ascribing responsibility that also need to be explained for the concept to make sense.

Reflective and Transitive Ascription

There are two ways in which responsibility can be ascribed ”reflexively and transitively. Transitive ascription stands for the process of someone or some group ascribing responsibility to the subject, whereas reflexively, the subject ascribes responsibility to itself. Traditionally and in the legal sense, responsibility is usually called for and ascribed transitively by a person or group that was affected by the subject s action. The subject may have a chance to defend itself, but the eventual sanctions are imposed despite and often against itswill.

Modernity, complexity, uncertainty, risk, and all the other developments that led to the development of responsibility keep such transitive ascription from being comprehensive, and one can see a development towards the selfascription by the subject. Responsibility then becomes the functional equivalent of traditional norms in situations where the traditional instruments of normative control no longer work (Kaufmann, 1990, p. 71). In this sense responsibility can be seen as a correlative to human freedom, and it turns into a self-obligation of the subject (Kaufmann, 1992, p. 41; 1995, p. 80).

This self-ascribed responsibility constitutes the centre of the approaches of moral philosophy that are sometimes called ethics of responsibility (Weber, 1992, p. 71). Since in this case subject and object are identical, the instance is usually thought to be an internal one, usually the personal conscience (S nger, 1991, p. 36; Lenk, 1998, p. 177). The fact that the conscience is the instance of self-ascribed responsibility also means that the sort of rules applying to it must be moral instead of legal. The absence of external instances and sanctions also requires some of the classical characteristics of moral personhood such as freedom and autonomy (von Foerster, 2001, p. 58). The moral and voluntary nature of self-ascribed responsibility has led some authors to believe that it is of a higher order than transitively ascribed legal responsibility (van Luijk, 1990, p. 42).

Excuses and Exemptions

One last aspect of responsibility that any theory must incorporate consists of the limits of the concept. There must be clear rules outlining under what circumstances responsibility is applicable and when it is no longer valid. These limits are usually called excuses or exemptions. Generally, responsibility can or should not be ascribed if the subject does not fulfil the conditions discussed earlier. This lack of fulfilment of conditions is called excuse if it is temporary; it is called exemption if the subject is not able to meet the conditions for a longer time or ever. Put differently, there are excuses, which function locally (examples include physical constraint or coercion); and exemptions, which operate more globally (such as insanity, childhood, or perhaps addiction ) (Wallace, 1996, p. 118). However, the question whether a subject can be responsible is usually not black or white. There are many different shades of grey that often have something to do with possible excuses. De George distinguishes three different categories of excusing conditions: those conditions that preclude the possibility of the action, those conditions that preclude or diminish the required knowledge, and those conditions that preclude or diminish the required freedom (De George, 1999, p. 112). Accordingly, a subject is excused from moral responsibility if the action in question is impossible to perform, if the subject does not have the required ability, if the subject does not have the opportunity to perform the action, or if the circumstances are beyond the subject s control.

Excuses or exemptions can be analysed in a more detailed way by looking at each of the conditions of responsibility. Each of the conditions discussed above can pose a problem if it is not fulfilled or if it is impossible to discern whether it is fulfilled. If there is no causal relationship between subject and object, responsibility is impossible to ascribe. The same is true in the case of a clear absence of freedom, power, or the necessary personal properties of the subject. In many cases the problem is to find out whether or not the subject lacked the conditions completely or partly.

This problem of clearly discerning whether a subject can be held responsible or whether it has a valid excuse is probably the most serious threat to the functioning of a valid theory and practice of responsibility. The problem is fundamental for any theory of responsibility because it is caused by the very reasons of the importance of the concept. Uncertainty, contingency, risk, and the modern perception of life as being unpredictable have been major factors in bringing about the success of the concept of responsibility. At the same time they now threaten its applicability.

Two Sorts of Responsibility in Information Systems and Their Resulting Problems

Looking at the different dimensions and determinants of responsibility, it becomes clear that there is an almost infinite number of combinations, each of which will lead to its own questions and problems. It is not possible to discuss all of them here. However, there are two typical configurations of responsibility whose impact on IS we will take a look at. The first one is based on clearly defined formal ”mostly legal ”norms, is addressed at individuals, and refers to an object that is also clear and in the past. The ascription happens transitively and from an external point of view. The second one is based on moral norms, ascribed most likely by the subject itself for an object that lies in the future and is unclear and uncertain. The subject does not have to be an individual, it can be a group or an organisation. We will later refer to these two poles as responsibility because of IS and responsibility for IS (Chapter 6). Between these two poles most responsibility ascriptions can be found.

Let us go back to the beginning of the section titled "Problems of Responsibility and Information Systtems", where several examples of responsibility in IS were enumerated. The first two of them fit the distinction between the two types quite well. The programmer who is ascribed responsibility for the programming bug has to answer for something that is in the past that may lead to serious external sanctions such as dismissal. The CIO on the other hand is responsible for the future working of the information infrastructure of the company, which is an object that is much harder to grasp with formal norms. The object is at the same time shrouded in uncertainty. Basically, we are looking at two examples of what responsibility can mean in the business use of information technology. Both examples are too brief to allow any meaningful analysis of the situation, but they can serve as typical illustrations of our topic. At the same time they also serve as a reminder of how difficult it is to realise responsibility. In both cases it is very easy to find excuses even without knowing the details of the example.

The programmer may be causally responsible for the bug in the program, but that is usually hard to determine. Modern programs are simply too complex to be understood by any one human being, including our programmer. She can therefore claim a lack of knowledge of the problem. Also, in all likelihood she will not have been the only programmer working on the program, and thus the complexity rises. The programming bug may have happened due to unclear program specification or a lack of communication that the programmer herself was not responsible for. But even if the causal link were clearly established, it is still possible that the programmer can excuse herself with a lack of freedom. She may have seen the problem coming but had to follow the customer s specification. On the other hand she may not have had the freedom to want to program better because she may not have been aware of programming standards because of a lack of education. Maybe she is just not the right person for the job and lacks the mental capacity to live up to it. Generally, she may not have been aware of the potential problem beforehand and much less of the fact that she might be held responsible for it later on. It is easy to see that there may be others who potentially share the responsibility for the programming fault: the colleague who did not tell her the specifications, the supervisor who neglected to check on her, the company that had no adequate standardised procedures, the computer profession that did not develop a sufficient curriculum, the industry that failed to regulate procedures, the society that did not take action, or the state that did not pass laws. One can see that it is easy to find quite a few excuses even in a case that is potentially as clear-cut as this one.

The second example of the CIO who is responsible for the information systems of her company leads to a number of different problems that are by no means easier to handle. The CIO s responsibilities are uncertain because they are mainly in the future. The definition of her responsibilities ”the development and functioning of the company s information systems ”is already so wide that it is hard to fill it with life. Additionally, the CIO can usually not realise all of this by herself. She depends on her staff for the fulfilment of the tasks that she cannot handle herself. This means that she has to deal with all of the problems of future responsibility on the one hand and of collective responsibility of the other hand. Future causalities can at best be estimated; there can be no knowledge about the future, and the individual s powers to influence it are clearly limited.

Furthermore, in the case of organisational responsibility that the CIO has to rely on, there are problems concerning the organisation and distribution of responsibility. One would have to ask whether there is the possibility of one individual being responsible for the actions of another and what that means. In addition, bothexamplesshowthemixofdifferenttypesofresponsibility.Theprogrammer s as well as the CIO s responsibilities are clearly role responsibilities. At the same time both have a moral aspect and a legal aspect. Even though one may be directed into the past and the other into the future, it is clear that they contain both temporal dimensions. Both also mix transitive and reflexive ascription, and due to the mix of components , it is not clear who or what might be the instance deciding about the final outcome of the ascription.

Having spent this much time on the discussion of responsibility and IS, the result of this analysis of the concept seems to be that there is no coherent theory that would allow an unequivocal ascription. Even though the different theories of responsibility discussed so far agree in a great number of points, this agreement is not far-reaching enough to allow for an undisputed ascription. There are several conclusions one can draw from this. First, one can simply refrain from discussing moral or normative questions in terms of responsibility. This is not a good idea because one would at the same time lose the practical advantages that the term has, namely its general familiarity and the positive reception it is usually greeted with. Also, it might be hard to find another concept that allows the discussion of the different ethical and moral theories that were introduced as the German and French tradition.

Another solution would be to narrow the term down more specifically and return to one of the theories introduced earlier on. One could for example concentrate solely on the moral or the legal aspect of responsibility, and in both cases it might be possible to come up with clearer rules of ascription and the resulting consequences such as sanctions or rewards. While this is certainly a manageable way of dealing with responsibility, the disadvantage would be that it would have to shield out the relationships between the different sorts of responsibility.

For these reasons we will try a different approach to responsibility in IS. Instead of allowing the multitude of facts and theories about responsibility to ruin the applicability of the term, we will attempt to clarify it by applying it reflexively in the hope that this will allow the determination of the theoretical heart and the practical consequences. We will therefore in the next chapter introduce and discuss the idea of reflective responsibility, and after that we will apply it to the normative problems of business information technology.

^[9] See, e.g., Neuberg, 1997, Part I; French, 1992, Chapter 4; Fischer, 1999; Wallace, 1996.

^[10] May (1992) argues that it is the solution to the problem of collective responsibility to hold people responsible for their attitudes. If someone has racist attitudes and lives in a society where racially motivated crimes occur, then that person would be considered responsible for the crimes even if he did not participate in the racist activities himself. While this is an interesting approach that might allow dealing with some of the most difficult problems of responsibility, it is not visible how this would lead to clear ascriptions and sanctions.

^[11] Wieland (2001a, p. 23) points out that this argument is in fact tautological. The classical argument against collective responsibility does nothing more than define responsibility in individual terms and then show that a collective cannot fulfil it.

^[12] Some staunch defenders of the field of artificial intelligence (AI) such as Brooks (2002) are convinced that we will be able to build computers that become like humans, that eventually the limits between information systems and humans will vanish . But even outspoken critics of this approach ”who think that computers as we know them are fundamentally unable to ever develop understanding or meaning, such as Dreyfus (1993) ”are careful not to say that building self-conscious computers is impossible. This is a good indication of how unclear the notion of consciousness still is.

^[13] For a thorough discussion of responsibility, see Birnbacher (1988).