Case studies are used frequently in the teaching of ethics, to engineers as well as non-engineers. The method has strengths and weaknesses. One strength of cases is that they provide a simple and straightforward way of discussing situations which pose ethical dilemmas. A weakness is that cases are almost always written quite briefly, with many details necessarily left out. As a result, some educators argue that cases are unrealistic, and do not prepare students for the richly-nuanced situations which they will encounter in the real world. Such cases are sometimes called "thin examples" because of their sparseness of detail.
There is an intriguing parallel between the problem confronting the teacher of ethics, and that experienced by the teacher of engineering. In the latter case we also necessarily work with sparse models. Such models must be sufficiently simple so that they allow us to analyze the system in question, and sufficiently complex so that they capture as much of the detail as is required to accomplish a successful design. The parallels between ethical analysis and engineering design provide the teacher of engineering ethics with a way to help students arrive at a deeper understanding of ethical analysis and engineering design through a comparison of the two. This paper enlarges on the parallels between these two teaching disciplines, and explains how we can make effective use of these parallels.
You have probably heard this ethical dilemma, or something much like it. You are the driver of a bus which has lost its brakes, and is careening down a mountain road. Suddenly you round a bend and before you on the road is a young child on one side, and a small group of seniors on the other. You can swerve to miss either the child or the group of seniors, but you can't miss both. Who do you save - or, whose life do you take - and why? Such cases or examples are a favorite means of studying ethical dilemmas in the ethics classroom. But, some philosophers question the use of such simplistic cases, arguing that the lack of detail may mislead students into taking simplistic approaches to complex real-life problems. Julie McDonald 1 calls such cases "thin examples" - thin because so much is left out, so much assumed. In the above case, does the driver have any other alternatives, are there passengers on the bus, how are they endangered by the alternatives available to you the driver, how many seniors are involved? And some would hope that you would ask two much deeper questions, which have the effect of extending the case in time. Just what are you doing driving down a mountain road in a bus with failed brakes, what did you do to get yourself into this situation? And, as for the future, what are the long term effects of your decision for all of the individuals involved, including yourself. We shall explore in this paper the problems associated with the use of "thin examples" in the classroom.
But, we are also going to talk about teaching engineering, because the process of engineering necessarily makes use of what we might call, in analogy, "thin models". It is not my objective here to argue a close connection or analogy between these two fields, but rather to ask whether a consideration of one area helps at all to understand the other.
The structure of the paper is shown on the right. With this introduction behind us we are going to start with the path on the left, looking at the problem of ethical dilemmas, with no reference to the engineering design problem. Then we will switch over, and consider the matter of teaching engineering design. The dashed lines suggest that there are connections between the paths which the reader may wish to make, but which will not be explicitly discussed until the final section. After we have finished following these parallel paths we will discuss the implications which arise for the teaching of ethics and the teaching of engineering.
2. Ethical Dilemmas
Before we discuss thin examples in more detail we consider briefly what might be called a "conventional view" of how moral reasoning happens. We can identify four elements of ethical reasoning.
Perception of a moral problem
Assumption of universality and impartiality
Application of action-guiding rules and principles
Obligation to right action
Unfortunately, in real life each of these steps is fraught with difficulty. Perceiving moral dilemmas can be quite difficult, and often our perceptions are quite different from those of another. (see, for example, Lawrence Blum 2). The application of moral reasoning universally to all, with impartiality, may be a good goal (and may not), but it is hardly realistic. Human beings are partial by nature, and one could argue that partiality is at the heart of the human relationships which are the essence of ethical life. (see, for example, Marilyn Friedma 3, 4 and Lawrence Becker 5). Perhaps it would be good if there were clear and unambiguous rules and principles to guide our actions. But, again, life is usually not that simple. Many of our most important ethical problems are true dilemmas, with no clear solution, or with more than one good solution. And finally, the obligation to right action is not always clear. This is particularly true if an action is considered within a narrow frame of reference. Broadening the dimensions of view over time, or over a richer set of situations, may well cause us to change our choice of apparent right action.
With these general concerns as background let us now turn to the specific task of using examples or cases to teach ethics. It is highly tempting to use "thin examples" in the teaching of ethics. They are usually short, easily posed, and attention-getting, if well designed. They also have a number of drawbacks. Here are just three.
The situations are often quite uncommon
Partiality is often not discussed
Examples are usually adversarial - there are winners and losers
On the first point, Marilyn Friedman has made this observation.
"Hypothetical disasters abound as thought experiments in these discussions. The moral world of mainstream ethics is a nightmare of plane crashes, train wrecks, and sinking ships! Wives and children drown in this literature at an alarming rate."3 or 4
Two points are important here. The first is that dramatic cases are hardly typical of our day-to-day problems. The second is that there is seldom any discussion of what led up to these disasters. Again, we can ask the question - perhaps in a metaphorical sense - of the driver of our bus in the introduction. "How did you get into this situation? Did you check the brakes? Did you accept responsibility for the lives you hold in your hands? What were the steps which led up to this dramatic situation? What was your role in the evolution of these steps?"
Let's apply the questioning to a more likely scenario for the practicing engineer. Your product needs to be shipped, but final testing is not complete. Your boss tells you that she is certain that the tests would be satisfactory, and to fill in the test form without finishing the actual tests. How did you get into this situation? Have you kept up with your timetable for testing? Have you helped create this problem? Have you acted in other ways to make your boss think that you might be willing to fake the tests? What is your company's role? Are there standards of conduct, a strict sense of honest practices, an ombudsman, an ethics committee? Few ethical situations come out of the blue. Students should learn to put today's ethical problem in the context of the past and the future.
Partiality is often not brought into the discussion of a thin example, to some extent because there is not enough detail to establish the nature of the relationships which would permit such discussion. And yet, partiality is a fact of practical ethical life, even though we often try to say that it is not in theory. We do not treat everyone alike. We are partial to members of our family, to friends, and others. This does not mean that we can, in justice, deprive another, toward whom we are not partial, of a right. It does mean that in the midst of a true ethical dilemma, where there is more than one good (or bad) solution, partiality may guide our action. If we are in a situation where we can save only one of two lives, for example, our relation to the two persons is relevant. The complex and intertwined ways in which partiality can impact an ethical decision are difficult to teach through examples which have little richness.
Thin examples are often adversarial. That is, someone must win, and someone lose. It's easier to pose the problem that way. Whose life shall I take, that of the child, or those of the seniors? Posing problems this way helps reinforce the common idea that life is inherently adversarial. But, in fact, it is often possible to look for a win-win solution. When your boss tells you to fake the tests, it is easy to think that either I agree and lose my self-respect or refuse and threaten my job. A far better approach is to look for a solution with no losers. What pressures are on my boss? How can I help her meet them? Can I help her understand the ethical problem that I feel in a sympathetic way? Can I stay late to do some extra tests? How can I get this situation away from win-lose? Thin examples can mislead one into accepting win-lose worlds, when win-win may be a possibility.
3. Engineering Design
Engineering design necessarily makes use of models which are "thin" in the sense of leaving out most of the details of "real life". We want our models to be simple enough so that we can work with them effectively, and yet sufficiently complex that they reflect the essential elements of what we are modeling. They need to have enough detail to predict, with acceptable accuracy, what will happen in real life applications. The discussion in this section has applicability to a wide range of engineering design problems, though sometimes in different ways and to different degrees. Let's pose three classes of problems that are of interest.
The design of a self-contained unit, involving no internal human interactions, such as an electronic amplifier, or a jet engine.
The design of a relatively complex system with some human interaction, such as an automobile or an airplane.
The design of a highly complex system, in which human interaction is a very important, and usually unpredictable, variable, such as a major urban transportation system or an electric power system.
In each of these cases the engineer is faced with the problem of deciding how complex the model must be in order to be sufficiently predictive, within, of course, the constraints of practicality. In teaching engineering we often use fairly simple models because they are easy to present, and because students find them manageable. The danger of such simplistic models is that students will come to see them as realistic for engineering practice. Our task as educators is to try to ensure that our students leave the classroom with a respect for the complexity of the real world and for the difference between "thin models" and "real life".
Dietrich Dorner 6 has recently analyzed engineering systems in a way that can help us see why they can be so difficult to understand . Dorner has identified four features of systems, which make a full understanding of any real system impossible. These are:
ignorance and mistaken hypotheses
Complexity reflects the many different components which a real system has and the interconnections or interrelations among these components. Engineering models necessarily neglect many of these components or features, and even more so their interrelations, but there is always a danger in doing so, and our students need to be aware of such dangers. Complexity is clearly a feature of the latter two of the engineering problems listed above, but it is characteristic also of the first case. Simple electric circuits, or even single components, are in reality infinitely complex. A resistor in an electric circuit has a multitude of complex parasitic effects, that become increasingly important as the frequency or the time rate of change in the circuit increases, which is exactly what is happening today as electronic technology evolves. Combine a dozen such components, or hundreds or thousands, on a single circuit board, and the complexity multiplies very rapidly. When we move to more complex systems, the modeling challenge can become staggering. Our students should have a sense of complexity and an understanding that engineering models are finite reflections of an infinite world, and need to be used with appropriate caution.
Many devices and systems exhibit dynamics, that is, the property of changing their state spontaneously, independent of control by a central agent in charge of the system. One of the most fascinating examples of our time is the Internet, an extraordinarily dynamic system, with no one in charge. There is no way to model the Internet system in a way which will predict its future and the future of the people and things which will be impacted by the Internet. Many of our complex technological systems have this property. Examples might include: a new freeway system, nuclear power, high definition television, genetic engineering. Our students need to understand the role that dynamics can play in the life of such systems.
Intransparence means that some of the elements of a system cannot be seen, but can nevertheless affect the operation of the system. Simple devices and systems have parasitics which are often not visible, either directly or indirectly. More complex systems can have many contributors to intransparence. In the Internet, for example, the list would include almost all of the users at a particular time, equipment failures at user sites, local phenomena, such as weather, which affect use of the Internet at other locations. Students need to understand that what you can't see might hurt you.
Finally, ignorance and mistaken hypotheses are always a possibility. Perhaps our model is simply wrong, faulty, misleading. This last problem is particularly interesting and important, because it is the one we can do something about.
The four problems with systems which Dorner has pointed out should be understood by the student of engineering as limitations on the ability of finite models to represent an infinite reality. It should not discourage the student to realize that these factors exist. After all, engineering does work. Bridges stand, airplanes fly, computers work, but only when chosen models are sufficiently good to account for relevant details necessary to predict outcomes with sufficient accuracy. If this is not the case, then devices and systems can fail, and the student must understand this.
4. The Implications for Education
It is probably necessary that we teach much of our ethics and engineering with thin examples and limited models. But, the temptation is to keep too many of our examples and models very simple. They are easier to write, to present, to discuss. Also, students have a difficult time dealing with too much complexity at first. The natural response is "one step at a time." And yet the real world, with which our students must in time deal, is complex, is dynamic, is not transparent, and this is true in the world of ethics as well as engineering. So, in the end we are left with the question of how we as educators should respond. What can we do in the limited classroom to prepare our students for an unlimited world?
Here are two ways to address this problem: find some new tools for analysis, and find some new ways to think about problems. Let's start with the first. In the closing pages of his book, Dietrich Dorner makes a plea for studying complexity and dynamics through computer simulation. Today it is very feasible to simulate high levels of complexity, and to demonstrate the effects of dynamics in the classroom. The purpose of such simulations is not so much to simulate the real world more accurately, as it is to reinforce in students the reality of complexity and dynamics and to give some idea of their general effects on systems. Here are Dorner's closing words on the subject.
"We have the opportunity today to undertake this kind of learning and teaching. Make-believe has always been an important way to prepare ourselves for the real thing. We should use this method in a focused manner. We now have far better tools for this purpose than we have ever had before. We should take advantage of them.
"Is this idea frivolous? Playing games in dead earnest? Anyone who thinks play is nothing but play and dead earnest nothing but dead earnest hasn't understood either one."
Computer simulation and study of complex engineering environments, whether devices or systems, is commonplace today in many branches of engineering. Students have the opportunity to ask an unlimited number of "what if" questions. Parasitics are easily simulated and studied for the effects which they tend to have on device or system operation. Such activities should be encouraged and expanded.
In the area of ethics, cases have been put on computers, with some degree of adaptation in the presentation of the facts with time as the students works through the problem. But, I am not aware of any attempts to model such important features of an ethical problem as:
the implications of partiality
the impact of the historical background of the problem
the future implications of actions
the role of emotion and intuition in decision making
It is certainly possible to conceive of a computer program that would allow such features to be modeled and studied. Once again, as in the case of the engineering simulation, the primary goal would be to stress the existence and general implications of such features, and not necessarily to simulate a particular real situation. Such a computer simulation is a challenge open to the world of ethical studies.
Robert Baum 7 has argued that it is particularly important to provide engineers with a rich humanistic approach to the teaching of ethics because engineers often tend to look at life in very simplistic and structured ways. Such richness could come in some measure through appropriate computer simulation.
Now let's talk about two ways to think about problems and the teaching of problems in the classroom. The following are not really new ideas, but rather are ideas which perhaps deserve a new emphasis, and a new commitment from teachers of ethics and teachers of engineering.
Discuss the nature and limitations of models and examples explicitly to give the student a more realistic understanding of their purpose and value, and their relationship to the real world.
Explore in class the idea that the real purpose of ethical analysis and engineering design is not to solve today's problems, but to prepare tomorrow's engineers to solve tomorrow's complex ethical and engineering problems, and in the process help the world become a better place.
The limitations of models should be a part of our teaching in engineering and in ethics alike. Engineers should always have a strong sense of the strengths and weaknesses of their engineering models. This can be done through a general discussion of the nature of the modeling process. It can also be done through the medium of computer simulation. For example, in my field a simple electric circuit can be simulated on a computer. Parasitics can then be added one at a time. The student can be asked to predict the effect of the parasitic on the performance of the circuit. This tends to focus the student's attention on the role which parasitics play, and on the need to assess whether they must be considered in a given design. The student should recognize that as the model becomes more complex the simulation becomes more accurate, but also more difficult to perform, and sometimes harder to interpret.
In the area of ethics, a comparable study of the "thinness" of examples is possible. Suppose, for example, that we propose the following problem to the class. You have gone into a store with a friend. Your friend shoplifts an object, and then leaves without you. The store manager has noticed this, and asks you for the name of your friend. What do you do? A class might give the problem some thought, and then give an answer. But now lets add a detail. Suppose that the store owner is your father. Is the problem changed? In what ways? And, of course, this now gives you the opportunity to talk about the issue of partiality, and its role in ethics. Suppose that the object stolen is a heater which is badly needed by your friend's family to stave off the cold of a very bad winter. What new ethical dimensions now arise. And, of course, the problem can be enriched as much as one wishes. The object of the exercise should be to help the student understand better the complexity of real life situations, and to avoid being mislead by thin examples.
The second approach to thinking about problems is to put them into a broader context. While the resolution of a complex problem may be on everyone's mind right now, the real goals and objectives and the healthy operation of a company or an organization transcend today's immediate problem. Recall the story of the manager who has asked you to falsify some tests. Suppose we propose this problem in the classroom, and then ask what could have been done to avoid the problem, what impact will its resolution now have on the future of the company, and how does all of this relate to the long term goals of the company. And how does this relate to the life of the community. Viewed this way problems are seen not as isolated events, but as part of a continuing process, the life of the company and the community, and as opportunities to enhance that life.
In a similar way engineering design can be put in a broad context. A particular design is a part of a company's life, and part of the community's as well. It is not at all unreasonable to ask how a design can contribute to the lives of those whom it will impact. These are complex questions for a young student of engineering, but they are questions which need to be asked if the engineer is to place his or her work in a societal context.
The real world is infinitely complex. We can deal with it in exchanging ideas and in doing our engineering only by creating finite languages and models. It is necessary that our students understand the differences between models and real life, and the importance of selecting the model which is appropriate for the situation.
Presented at the Annual Meeting of the Pacific Southwest Section of the American Society for Engineering Education, San Luis Obispo, CA, March 14-15, 1997.
Tim Healy has BS, MS and Ph.D. degrees in Electrical Engineering from Seattle university, Stanford University, and the University of Colorado. He has taught at Santa Clara University, where he holds the Thomas Bannan Chair, for 31 years, in communications and electromagnetics. He is a Scholar of the Markkula Center for Applied Ethics and is active in the teaching of ethics to engineers.
1Julie McDonald, "Thin Examples of Moral Dilemmas", in Social Theory and Practice, Vol. 19, 1993, p 225.
2Lawrence Blum, "Moral Perception and Particularity", in Ethics, Vol. 101, No. 4, July, 1991, pp. 701-725
3Marilyn Friedman, "The Social Self and the Partiality Debates", in Feminist Ethics, University Press of Kansas, Lawrence, Kansas, 1991
4Marilyn Friedman, What are Friends For: Feminist Perspectives on Personal Relationships and Moral Theory, Cornell University Press, Ithaca, NY, 1993
5Lawrence Becker, editor, Symposium on Impartiality and Ethical Theory, in Ethics, Vol. 101, No. 4, July, 1991, pp. 698-864
6Dietrich Dorner, "The Logic of Failure: Why Things Go Wrong and What We Can Do To Make Them Right", Metropolitan Books, New York, 1989, (EnglishTranslation, 1996)
7Robert Baum, Ethics and Engineering Curricula, The Hastings Center, Hastings-on-Hudson, NY, 1980