Ethical issues in engineering
Anke van Gorp
Engineers have to make decisions concerning ethical issues during technologicaldesign processes. In this thesis the kinds of ethical issues that engineers encounter are described, together with the way engineers deal with them, with a focus on ethical issues related to safety and sustainability. Four design processes were studied, the design process for an ultra light car, for piping and equipment for chemical installations, for a bridge and for a lightweight open truck trailer. A difference can be seen between normal and radical design. During the normal design processes for the bridge and piping and equipment for chemical installations engineers referred to regulative frameworks to account for decisions about safety and sustainability. These regulative frameworks give minimal requirements, (parts of) operationalisations, rules and guidelines for use in normal design. Engineers do not, or only partly use, the regulative frameworks in the radical design processes of an ultra light car and a lightweight open truck trailer instead they relied on internal design team norms for making decisions about ethical issues. Following the descriptive case-study research, the author discusses some preliminary notions for conditions for warranted trust in engineers making normal and making radical designs.
‘Wonder en is
Ethic al is sue s in en gineerin g de sign Ank e van Gorp Simon St ev in Serie s in the Phi lo soph y ofUitnodiging
tot het bijwonen van de
verdediging van mijn
proefschrift:
Ethical issues in
engineering design
op maandag
14 november 2005
om 10.30 uur
in de senaatszaal
van de aula van de
Technische Universiteit Delft
aan de Mekelweg 5
Voorafgaand aan de verdediging
zal ik om 10.00 uur een korte
toelichting geven.
Na afloop bent u van harte
welkom op de receptie.
Anke van Gorp
Westeinde 21c
gheen wonder’
Safety and sustainability
Ethical issues in engineering design; Safety and sustainability
Proefschrift
ter verkrijging van de graad van doctor aan de Technische Universiteit Delft,
op gezag van de Rector Magnificus prof. dr. ir. J.T. Fokkema voorzitter van het College voor Promoties,
in het openbaar te verdedigen op maandag 14 november 2005 om 10:30 uur door Anke Christine van GORP
materiaalkundig ingenieur geboren te Tilburg
Dit proefschrift is goedgekeurd door de promotoren: Prof. dr. ir. P.A. Kroes
Prof. dr. M.J. van den Hoven Samenstelling promotiecommissie Rector Magnificus, voorzitter
Prof. dr. ir. P.A. Kroes, Technische Universiteit Delft, Promotor Prof. dr. M.J. van den Hoven, Technische Universiteit Delft, Promotor Prof. dr. A. Grunwald, Forzungszentrum Karlsruhe GmbH
Prof .dr. B.A.W. Musschenga, Vrije Universiteit Amsterdam Prof. dr. ir. P. Kruit, Technische Universiteit Delft
Prof. ir. A. Beukers, Technische Universiteit Delft Dr. H. Zandvoort, Technische Universiteit Delft
Dr. ir. I.R. van de Poel heeft als begeleider in belangrijke mate aan de totstandkoming van het proefschrift bijgedragen.
ISBN-10: 90-9019907-1 ISBN-13: 9789090199078 ISSN: 1574-941X
Simon Stevin Series in the Philosophy of Technology
Delft University of Technology & Eindhoven University of Technology Editors: Peter Kroes and Anthonie Meijers
Volume 2
© Anke van Gorp, 2005
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without prior permission in writing of the publisher.
e-mail: anke_van_gorp@yahoo.com
ISBN-10: 90-9019907-1 ISBN-13: 9789090199078 ISSN: 1574-941X
Contents
1 Introduction 9
1.1 Research question and objective 13
1.1.1 Ethical issues 14
1.1.2 Ethical issues in engineering design 15
1.2 Research approach 17
2 Engineering ethics and design processes 21
2.1 Engineering ethics 21
2.1.1 Ethics in design processes 23
2.2 Design 25
2.2.1 Design process 25
2.2.2 Design problems 27
2.2.3 The design process as a social process 29 2.2.4 Organisation of the design process 30 2.3 Characteristics of design processes in relation to ethical issues 32
2.3.1 Design type and design hierarchy 32
2.3.2 Normative frameworks 36
2.3.3 Moral responsibility and the trust relationship between
engineers and society 37
2.4 Summary 42
3 Introduction to the case-studies 43
3.1 Working hypotheses 43
3.2 Selection of the case-studies 44
3.3 Acquisition of empirical data 46
4 DutchEVO, safe or sustainable? 49
4.1 A light family car 50
4.2 The design team 53
4.2.1 ‘If you have thought it through then it is ok’ 56 4.3 What does it mean for a car to be safe? 59
4.3.1 Active safety 60
4.3.2 Passive safety 61
4.3.3 Partner protection 63
4.3.5 Regulation 64
4.4 Light throw away after use? 65
4.5 Sustainable and / or safe 70
4.6 Summary of the case and the regulative framework 70
4.6.1 Ethical issues 71
4.6.2 Decision making on ethical issues 73
4.6.3 Regulative framework 74
4.7 Acknowledgements 76
5 Piping and Equipment 77
5.1 The design of a (petro)chemical plant 77 5.2 Regulation regarding pressure vessels and piping 79
5.2.1 Regulations 80
5.2.2 Codes regarding pressure vessels and piping 81 5.2.3 Standards regarding pressure vessels and pipes 82
5.3 Clear responsibilities and tasks 83
5.4 Ethical issues 84
5.5 Summary of the case and the regulative framework 90
5.5.1 Ethical issues 91
5.5.2 Decision making on ethical issues 92
5.5.2 Regulative framework 92
5.6 Acknowledgements 94
6 Designing a Bridge 95
6.1 The design problem 95
6.2 Trying to reconcile all requirements and stakeholders 98
6.3 Legislation and codes 102
6.3.1 Safety during construction 102
6.3.2 Safety in use 104
6.3.3 Sustainability 110
6.4 Responsibility and liability 112
6.5 Summary of the case and the regulative framework 113
6.5.1 Ethical issues 114
6.5.2 Decision making on ethical issues 116
6.5.3 Regulative frameworks 117
6.6 Acknowledgements 121
7 Design of a lightweight trailer 123
7.1 A lightweight truck trailer 124
Contents
7.3 Safe in what sense? 131
7.3.1 Structural reliability 133
7.3.2 Misuse and overloading 141
7.4 Ascribing responsibilities 143
7.5 Summary of the case and the regulative framework 147
7.5.1 Ethical issues 147
7.5.2 Decision making on ethical issues 150
7.5.3 Regulative framework 150
7.6 Acknowledgements 151
8 Conclusions of the empirical study 153
8.1 Summary of the results 154
8.2 Ethical issues and design type and hierarchy 159 8.3 Approaches to resolve ethical issues and design type and hierarchy 160
8.3.1 Decision makers and design type and hierarchy 162
8.4 Regulative frameworks 165
8.5 Design problem formulation 167
8.6 Generalisation of the conclusions 169
9 Towards warranted trust in engineers 175
9.1 Normal design 175
9.1.1 Required competence of engineers in normal design 176 9.1.2 Grunwald’s requirements reconsidered 177
9.2 Radical design 181
9.2.1 Identifying what affected actors value 182
9.3 Further research 186
9.4 Recommendations for engineering education 187 Literature 191 Samenvatting 199 Appendix 1 211 Appendix 2 215 Dankwoord 219 Curriculum Vitae 221
9
1 Introduction
On March the 6th, 1987, the roll-on/roll-off (ro/ro) passenger and freight ferry the Herald of Free Enterprise capsized just outside the Zeebrugge harbour.1 Water
rapidly filled the ship, leading to the death of 150 passengers and 38 crewmembers. The main cause of the disaster was that the inner and outer bow doors were open when the ship left port.
The assistant bosun should have closed the doors, but he had fallen asleep. The absence of warning lights made it impossible to see from the bridge whether the bow doors were closed. On at least two previous occasions, similar negligence with sister ships owned by the same company had led to the ships leaving port with the bow doors open. These incidents however passed without disastrous results [London Crown, 1987].
Pressure to depart quickly and poor communication had contributed to leaving port with the bow doors still open in the case of the Herald. As is often the case, it was human error that preceded the disaster, but it was the design of the ferry that made the occurrence of such a disaster possible in the first place. It was the inherent instability that ro/ro ferries encounter when water enters a deck that played an important role in the disaster. This is an aspect of ship design. It might be expected that while designing the Herald and her sister ships, the engineers were aware that if water were to flood on the decks the ship might quickly become unstable. Following the Herald disaster there was a similar disaster with another ro/ro ferry, the Estonia. Water filled one of the decks and the ship capsized killing nearly 800 people. This happened despite the fact that Estonia’s owners had complied with the proposed new regulation concerning ro/ro ferries formulated after the Herald disaster [Van Poortvliet, 1999].
In the following, more detailed description of the Herald disaster I will focus on decisions made in the design process that made the ship very vulnerable to water flooding the car decks. This example demonstrates the possible ethical impact of decisions made in the design process.
One ethical question that arises in relation to the design of the Herald of Free Enterprise, and other ro/ro ferries, is whether it should have been designed to be more safe given the fact that it was known that water entering the deck might result in rapid capsizing. This is a moral problem because passengers, crew and ——————————————————————————————————
1 This description of the Herald of free enterprise disaster is based on “Ethical considerations in
their families are harmed when a ship capsizes. There were, and are, simple technical solutions if one wants to prevent rapid capsizing when water enters a deck. Bulkheads created on the decks can easily prevent water from flowing freely over a deck and prevent rapid capsizing [www.safetyline.wa.gov, 2005]. Bulkheads on decks, however, give rise to longer loading / unloading times and take up space on the decks, hence this costs money.
When we look at ethical problems in relation to the design of the Herald and comparable ships, ethical issues become relevant at different stages of the design process and during the use of the product. Ethical issues are relevant during the formulation of criteria and requirements for the design and in the acceptance of trade-offs between requirements. I will focus on the formulation of safety requirements for a ro/ro passenger or freight ferry, and on the trade-offs that exist between safety and economic requirements. This description will explain why ro/ro ferries were not designed in a way that would prevent rapid capsizing if water floods a deck.
When it comes to formulating legal safety requirements, the International Maritime Organisation (IMO) has an important part to play. This international organisation is responsible for adopting legislation for ships. IMO’s safety legislation deals with the ship and passengers. The SOLAS (Safety of Life at Sea) convention is especially concerned with passenger safety and with lifesaving equipment on passenger ships. IMO officials knew as early as 1981 that if water entered the car decks of a ro/ro ferry, the ship could be lost in a rapid capsize [Van Poortvliet, 1999, 52]. Water entering the car deck will flow to the lowest point leading to a greater inclination, resulting, if the inclination exceeds a certain angle, in a rapid capsize. This fact has been regarded as common knowledge in the maritime world, at least since 1981. The IMO did not adjust its regulations to solve this problem, even though simple technical solutions, e.g. bulkheads, were available.
Legislation adopted by the IMO needs to be implemented by governments, and only governments accepting the IMO convention will implement it. Thus when making a convention, it is important to make it acceptable for as many governments as possible, otherwise only a small percentage of all fleets will be obliged to abide by the convention. A shipping company can decide to sail under the flag of another country which has not ratified an IMO convention, if, in the opinion of company management, complying with the convention will cost a lot of money. So there is a certain amount of pressure on the IMO not to issue safety requirements that are considered by some governments to be too tight or too costly to implement.
Introduction
11
Most IMO conventions affect new ships but do not apply retrospectively to ships already at sea. This is known as the grandfather clause. The grandfather clause protects the poorer states, because for them it would be too costly to adapt their older fleets to new legislation. IMO legislation may thus be said to be weak and ro/ro vessels complying with IMO legislation are still prone to rapid capsize.
Apart from the IMO, insurance and classification companies also have a part to play in the formulation of safety requirements. To obtain hull insurance from insurance companies such as Lloyd’s of London, a ship needs to be classified. Classification organisations are private organisations that monitor compliance with legislation during construction and certify sea worthiness during a ship’s lifetime. Only the ship’s equipment and construction are taken into account by the classification organisations, they do not deal with passenger safety [Van Poortvliet, 1999].
There is little incentive for shipping companies to ask for, or for shipyards to design, ships that are safer than required by IMO conventions and hull insurance regulations. When disasters occur the investigation that follows usually concludes that it was a human error that led to the disaster. Little attention is given to the design of the ship as long as, on completion, the ship complied with the current regulations of the time.
Six actors are important in the formulation of the safety requirements laid down for ro/ro ferries. These actors are: the IMO, governments, insurance companies, classification companies, shipyards and shipping companies. To understand why these six actors have not formulated tighter safety requirements, it is important to realize that when safety requirements are formulated a trade-off is made with economic requirements.
Economic considerations are important for insurance and classification companies because they depend on shipyards and shipping companies. When the safety requirements they impose are more costly than those of competitors they will lose customers. Insurance companies will want the requirements to be tight enough to prevent them from having to pay out too frequently for hull loss and damages. However, they usually do not want to impose more or tighter requirements than their competitors as they are afraid of loosing their customers.
Shipyards do not have loyal customers. To be competitive the price needs to be kept as low as possible or at least lower than that of the competitors. Safety measures are usually only built in when there is a legal obligation to do so. Shipyards may not be held liable if, at the time they were built, their ships complied with the relevant legislation.
Shipping companies in Northwest Europe are in strong competition with trains and planes, therefore they do not want to face increasing costs or longer
loading times. In the case of ro/ro ferries, shipping companies do not want to have bulkheads on the decks because it takes time to put them in place while loading the ferry. Moreover, fewer cars can be transported because the bulkheads have to be designed in such a way that larger and smaller cars can fit between them; this requires spacing between the bulkheads that is less efficient if large cars and freight-trucks are considered. So shipping companies also trade off safety against economic considerations.
Finally, the IMO and governments of individual countries also trade off safety considerations against economic ones. As we saw earlier for IMO conventions to be effective as many countries as possible have to support them. For many countries, economic considerations will play an important part when it comes to deciding which safety requirements they consider acceptable. This is reinforced by the fact that shipping companies can choose which flag they sail under. Governments could forbid ships that do not meet their stricter national regulations from entering their harbours. There are economic reasons not to do this. A government’s national harbours, where stricter regulations are enforced, will have a competitive disadvantage compared to harbours in countries that do not impose stricter regulation than the IMO regulation. This, in turn, reinforces competition between countries when it comes to devising attractive rules for shipping companies. Such competition may well water down safety requirements.
To summarise, some of the important ethical issues in the case of the Herald of Free Enterprise are the following: the ship’s design was inherently unstable once water entered the car deck. Is it ethically justifiable to design, produce and use ships that are in certain circumstances inherently unstable? What is the responsibility of engineers in this complex situation? There were no warning lights on the bridge, therefore it was not possible to establish from the bridge whether the bow doors were closed or not. Should engineers attempt to anticipate human errors during the design process? Is it the responsibility of the engineers to design in a way that prevents human errors as much as possible or even to design idiot proof ships? It is, for example, desirable to design ferries that cannot leave port unless the bow doors are fully closed and secured. As we have seen in setting the design requirements, trade-offs are made between safety and economics. There is economic pressure to water down safety requirements. Are trade-offs between economics and safety acceptable? Which of the choices regarding this trade-off can be justified? Does following the regulations lead to morally acceptable designs? Ethical issues that come up in design processes like the ones mentioned above will be central in this thesis.
Introduction
13
1.1 Research question and objective
Technology has a profound influence on society. New possibilities and new risks arise as a consequence of the employment of new technologies and products. Decisions made during design processes shape the possibilities and risks of products. These decisions are ethically relevant. Some decisions, for example, can have a large influence on the safety of people using the product. Although there is extensive literature on design processes and on engineering ethics, specific attention for ethical issues in design processes is relatively new. A lot of the literature on engineering ethics has been developed from the study of disasters such as the Herald of Free Enterprise or cases of whistle-blowing but in these studies only little attention is given to the design process. In this research I will focus on daily practice in engineering design. It might seem strange to start this chapter with the description of a disaster given that I will look at daily engineering practice, the description of the Herald disaster is only intended to make it clear that engineers make choices regarding ethical issues during design processes. These decisions can, but need not, have detrimental consequences. With or without the actual occurrence of the Herald disaster, the design of this ro/ro ferry may be seen as an example of daily engineering practice. My research question is as follows:
What kind of ethical issues come up and how do engineers deal with these ethical issues during design processes?
This analysis of design practice will contribute to engineering ethics.2 It will
provide detailed information on which ethical issues play a part in engineering design and how engineers decide about these issues. This information should enrich discussions on the moral responsibility of engineers in design processes. The objective of this research can be summarised as follows:
To provide a contribution to discussions on the moral responsibilities of engineers in engineering design processes.
The contribution will consist of detailed descriptions of engineering design practices and a normative analysis of these design practices. As can be seen in the Herald case there can be regulation pertaining to the design of the product. This research should provide information for answering, amongst others, the following question: do engineers behave in a morally responsible manner if they follow the existing regulations or should responsible engineers do more than just follow the regulations?
——————————————————————————————————
2 People interested in design research and not engineering ethics might be interested in the
1.1.1 Ethical issues
Until now I have assumed that the reader knows intuitively what ethical issues are. I will now explain in more detail the meaning of the term “ethical” as used here.3 I will call a problem an ethical or moral problem if moral values are at
stake. In characterising moral values I will follow Thomas Nagel. According to Nagel, there are different sources of value, special allegiances, general rights, utility, perfectionist ends of self-development and individual projects, that cannot be reduced to each other or to more fundamental values. Values based on special allegiances are, according to Nagel a result of a subject’s relationships to others and consist of special obligations to other people or institutions. General rights are rights that everyone has as a human being. These rights constrain action; actions that violate these rights are morally not permitted. According to Nagel, ‘utility includes all aspects of benefit and harm to all people (or sentient beings)’ [Nagel, 1979, 129]. Perfectionist ends of self-development refer to the intrinsic value of certain achievements. Nagel provides examples of the intrinsic value of scientific discovery or artistic creation. The fifth type of value derives from individual projects. Nagel says that ‘this is value in addition to whatever reasons may have led to them in the first place’ [Nagel, 1979, 130]. An example Nagel gives is that if you have set out and started to climb to the top of Mount Everest then this project gains importance. Ethical theories usually focus on one of the sources of value. Kantianism focuses on universal rights. Utilitarianism only accounts for utility. Virtue ethics concentrates on perfectionist end of self-development. I do not want to limit myself to one source of value by choosing a definition of ethical issues that refers only to utility or virtues or universal rights. In this thesis, issues that are related to one of the sources of moral values identified by Nagel are called ethical issues and decisions concerning ethical issues are called “ethically relevant” decisions. For example, issues concerning safety are related to utility but also to universal rights, therefore safety is an ethical issue. The term “ethical issue” only indicates that the way engineers deal with an issue can be evaluated from an ethical point of view.
This conception of ethical issues is used independently of what engineers themselves think are ethical issues. Engineers may or may not share this conception of ethical issues. Even in cases that engineers do not consider an issue to be ethical, if it is an ethical issue according to the above conception I will treat it as such in this research. There might also be issues that engineers call ethical but that are not ethical issues according to the above conception; these ——————————————————————————————————
3 Some philosophers indicate that morality describes a code of conduct of a society or another
Introduction
15
issues will not be considered ethical issues in this thesis. An example of this is that some industrial designers conflate aesthetic and moral values.
Some of the ethical issues are also legal issues, for example safety issues. There is a lot of legislation, standards and codes pertaining to safety and design. This makes decisions regarding safety no less ethically relevant, it only provides engineers with rules they should follow from a legal point of view when making decisions. In these cases the way engineers deal with these issues can be evaluated both from an ethical and a legal point of view. Decisions about the safety of a product might then be morally right or wrong and legal or illegal. A question that can be raised in such instances is whether a design that is safe enough according to legislation is also ethically acceptable and vice versa. Legislation, codes and standards regarding safety can also be evaluated ethically.
1.1.2 Ethical issues in engineering design
To take into account all ethical issues connected in one way or another to a design process would be impossible. It is not that difficult to point out the ethical relevance of what seems to be a very trivial choice, like which tea to drink during meetings of a design team. Tea can be produced organically or with the use of herbicides and under good or bad working conditions. The choice of what tea to drink is therefore related to utility and universal rights. Lots of ethical issues might play a role in the design context, for example some parts might be produced in countries where child labour is usual and therefore it might be assumed that these parts are made by children. Although all these issues like child labour, exploitation of underdeveloped countries, use of herbicides and pesticides are indeed ethical issues, these issues will not form the main focus of interest in this research. I will concentrate on ethical issues that have a direct influence on the design of a product and the way it is used. In particular, I will focus on ethical issues concerning safety and sustainability. The reason for the focus on safety and sustainability is that these play a dominant role in many design processes. Given the conception of ethical issues it is clear that safety and sustainability may give rise to ethical issues. Decisions made about these issues are related to utility and general rights. Decisions regarding safety and sustainability are made in almost every design process, although the importance of these subjects may differ. In some cases, sustainability or safety will not be regarded or discussed by the engineers, but this does not mean that there are no choices made regarding sustainability or safety.
In the following two examples I will show that the impact of decisions made during design processes concerning safety and sustainability may be far reaching. In everyday life choices about safety and sustainability with regard to the use of technological devices are often made, but the consequences of the choices made
by an individual user about safety and sustainability are usually of a much smaller magnitude than those for decisions made during a design process.
When designing a printer/copier, a choice needs to be made as to whether the printer/copier will be able to print two sided or not. Once a choice is made for two sided printing and copying, an additional choice needs to be made about the default properties. If two sided printing is the default option, users have to make an explicit choice to print one sided. Usually the prints and copies coming of the machine will be printed two sided. Only in exceptional cases, where the two sided copies and prints option is switched off by the user, will papers be printed one sided. This default option will probably save a lot of paper compared with a printer/copier that can only print one side. The environmental effects of saving paper are not that big if a single printer/copier is regarded but when the total number of printers/copiers in use is considered the amount of paper saved by printing two sided copies and prints is enormous. As paper is produced from wood, a reduction in paper use will also reduce the amount of wood used. The production of paper, the transportation of wood and the transportation of paper all require energy. The amount of energy used will also be reduced and the total reduction in the resources used will be significant on a global scale. This example shows that decisions made during a design phase of a product, and that seem trivial during that phase, can have large environmental effects.
Another example of the ethical impact of design decisions is the following. A person may decide not to drive too fast as this is usually dangerous and not environment friendly. The government of a country might decide to regulate the speed of cars by imposing speed limits. If there are speed limits imposed drivers can still drive as fast as they wish, and is possible in their car, but they will run the risk of being fined when exceeding the speed limits. Car engineers might decide to design a car in which it is impossible to exceed the speed limits. Trucks for example in the Netherlands are equipped with a speed regulator that makes it impossible for the driver to drive faster than 90 km/h. This example illustrates the influence engineers may have; they can promote or prevent speeding. Independently of what regulation requires or what speed limits are legally enforced, engineers can design cars with lower top speeds.4 Cars with top speeds
of 300 km/h make speeding possible and might perhaps invite drivers to test the top speed while installing a speed regulator or designing a car with a less powerful engine would make such speeding impossible. Designing cars with ——————————————————————————————————
4 In the Netherlands only trucks are legally required to be equipped with speed regulators.
Engineers might, however, decide to equip cars with speed regulators even if this is not legally required. There is a gentlemen’s agreement between German car producers to limit the speed of a car to 250 km/h, examples are the Mercedes Benz CLK 55 AMG cabriolet, the BMW M5 (with speed limiter 250 km/h without 338 km/h), Audi A3 sportback 3.2 Quatro. Although MG is not a
Introduction
17
lower top speeds would also save a lot of fuel as the fuel consumption is higher at higher speeds. Lower fuel consumption also decreases CO2 production. Smaller
speed differences, for example between trucks and cars may possibly decrease the number of accidents occurring on roads and thereby the number of people injured and killed on roads. So by choosing to design a car with lower top speeds, be it by actively limiting the top speed of the car or designing a less powerful engine, engineers can reduce fuel consumption, CO2 production and the amount
and severity of accidents on highways.
1.2 Research approach
Descriptions of design practices have to be obtained to answer the research question. This will be done through studies [Yin, 1984/1989]. In case-studies, different ways of obtaining data can be used. In my case-case-studies, I have interviewed engineers, observed design teams at work and read official and informal design documents.
Observing of design meetings allowed me to collect information about the way the decisions are made by engineers. Observing design meetings is also a way to get information about what engineers perceived to be the difficulties and challenges of design processes. The meetings used for the case-studies were taped and the tapes transcribed. See appendix 1 for a list of the meetings that were observed.
Design documents, especially the official ones meant for customers, give a kind of reconstruction of the decisions made in design meetings. The design documents can sometimes provide additional information, for example, in some design meetings arguments for or against certain choices were given but the actual decision was not taken during the meeting although the decision was documented in the design documents. The more informal design documents often gave information on specific aspects of the design process. This information can be used for later official design documents for customers. Sometimes the informal documents were more detailed than the official design documents.
Interviews were used to get further information on what role specific engineers had in the design process and whether they encountered ethical issues. In the interviews engineers were asked what they considered to be the ethical issues in their design process. Most interviews were held at or near the end of the observation period, so it was possible to ask engineers about anything that was not yet clear to me after having read the design documents and observed the design meetings. All the interviews were transcribed and the transcriptions were approved by the interviewees, see appendix 1.
After the interview and observation periods I gave a presentation of my results to each of the design teams. This presentation was followed by a discussion. Engineers could indicate whether I had made some incorrect factual statements about the design process and whether or not they recognised the results. The presentation was also a last opportunity for me to ask about some details that were not yet clear to me. The reason for giving a presentation at the end of the observation and interviewing period was that the engineers were curious about, and interested in, my research and results. If engineers asked me about the results, I told them that I would present my results, and give them the opportunity to react to these results, later on. This gave me the opportunity to postpone discussions on safety and sustainability until the presentation in order to influence the design process as little as possible.
I have chosen to change the names of the participating engineers in the descriptions of the design processes The engineers did not ask me to do this but I have chosen to protect their privacy in the main text, see appendix 1 for more concrete information. The exact identity of the persons involved does not matter for the case descriptions, his or her arguments, decisions and formal position in the design team are relevant for this thesis.
In doing the case-studies, I made the choice to present myself as an engineer among engineers. A large advantage of this choice was that the design team members knew that I could understand the “language” of engineers; and although my participation was kept to a minimum during the observation and interview period, the members of the design teams knew that I was a qualified engineer; this made communication easier. The design team members did not feel compelled to explain everything they were doing in a simplified way.
In line with keeping my participation at a minimum, I did not contribute to the solution of the design problem. This choice was also made because participating in a design process would require a lot from me as I had no previous design experience. The design task would probably completely absorb my attention and time, making it difficult also to observe what was going on in the team. Therefore, I was involved but not as a member of the design team. I had some input in the design process at the end of the data gathering period when I presented my results.
With regard to the validity of my results, I tried to minimise the influence of my presence. I did this by indicating that I was interested in ethical issues in design processes without a thorough explanation of what I meant by the term “ethical issues”. Roughly, engineers in the cases interpreted the term “ethical issues” in four different ways. First, some engineers thought that I wanted to look at human rights and wondered what I was doing in their company because they
Introduction
19
could not imagine human rights issues in their company. Second, others thought that ethics was only concerned with what kind of life a person should live and wondered what designing had to do with that. Third, some people, Bachelor students especially, thought that I would study the etiquettes in their design team. Fourth, some engineers shared the interpretation of ethical issues used in this study and expected me to look at decisions concerning the safety of the design or the prevention of disasters. I deliberately did not correct the engineers who thought that I was interested in etiquette, the good life or human rights. My presence might have influenced the design team but as most of the engineers did not know exactly what I was looking at it is not likely that they placed more emphasis on decisions concerning safety and sustainability. I taped whole design meetings and made notes throughout design meetings and interviews, not only when safety or sustainability issues were under discussion. If engineers asked questions about my research results I usually referred to the presentation that I would give later on.
As little is known about the way engineers deal with ethical issues in daily engineering design, this was an exploratory research project. Based on ideas taken from the literature on design processes, and to be presented in chapter 2, working hypotheses were formulated. These working hypotheses and the selected cases are introduced in chapter 3. The cases are described in chapters 4 to 7. Conclusions are drawn from the cases and an effort is made to generalise the results of the case-studies in chapter 8. The results from this research are used to make a start with defining conditions for warranted trust in designing engineers, in chapter 9. These conditions lead to a preliminary delineation of the moral responsibilities that engineers have during a design process.
21
2 Engineering ethics and design processes
The research question and objective formulated in chapter 1 indicate that this thesis contributes to engineering ethics. This research should lead to descriptions of the ethical issues engineers encounter and how they deal with these issues in design processes. The focus on ethical issues in engineering design processes is relatively new. As yet in engineering ethics there has not been a lot of systematic attention for design processes, as will be indicated in section 2.1. An overview of the literature on the nature of design processes is presented in section 2.2. This overview is relevant because the ideas about the nature of design processes are used to guide the gathering of data in the case-studies. Ideas about design processes that are particularly relevant for this thesis, because these ideas explain and inform the formulations of working hypotheses presented in the next chapter are introduced in section 2.3.
2.1 Engineering ethics
Research into ethics and design is part of the research field of engineering ethics. In this section I will not give a complete overview of engineering ethics literature. I will restrict myself to a description of the main issues that are focussed on in engineering ethics and how this research is positioned with regards to these issues.
Engineering ethics is the field of study that focuses on the ethical aspects of the actions and decisions of engineers, both individually and collectively. A rather broad range of (ethical) issues are discussed in engineering ethics: professional codes of conduct, whistle-blowing, dealing with safety and risks, liability issues, conflicts of interests, multinational corporations, privacy etc (see for example [Harris et al., 1995] [Davis, 1998] and [Bird, 1998]). A substantial amount of literature on the teaching of engineering ethics to engineering students has been developed since the beginning of the 1980’s, (cf.[Baum, 1980] [Unger, 1982] [Martin and Schinzinger, 1989] [Harris et al., 1995], [Birsch and Fielder, 1994]).
A salient feature of engineering ethics literature is that a lot of it has been developed based on studies of disasters like the Challenger disaster ([Vaughan, 1996] and [Davis, 1998]). Another feature of engineering ethics is that, especially in the United States, there are a lot of proponents who regard engineering ethics as a kind of professional ethics (cf [Schaub et al., 1983] [Davis, 2001] and [Harris, 2004). The idea is that the engineer as a professional has obligations not only to his or her employer but also to the general public, as for example doctors or
lawyers also have obligations. Engineers should adhere to professional codes of conduct that state, for example that engineers shall hold the safety and welfare of the public paramount. Based on descriptions of the Challenger disaster, Davis emphasizes that there is a difference between engineers and managers. Engineers should adhere to their professional norms and hold safety paramount and managers do not do this [Davis, 1998]. This tendency to regard engineering ethics as a kind of professional ethics has led to a focus on the individual engineer and his or her responsibilities in his or her job and profession in most (American) engineering ethics textbooks. This can also explain the focus on whistle blowing that can be found in some of the engineering ethics literature. The individual engineer should in certain cases take his or her moral and professional responsibilities seriously and blow the whistle.
According to Zandvoort et al. [Zandvoort et al., 2000] and Devon et al. [Devon, et al., 2001] engineering ethics should focus on more than the individual engineer. They argue that the ethical problems that engineers encounter are partly due to the context they work in. Some of the ethical problems cannot be solved by individual engineers or the profession.
In contrast to most of the literature on engineering ethics, I will not focus on disasters and isolated individual engineers. In this research the focus will be on daily practice in engineering design. Fortunately, not every engineer has to deal with disasters or with a decision as to whether to blow-the-whistle or not. There is not much literature on ethical issues in daily practice, yet every engineer will be confronted with these ethical issues. Furthermore, no distinction will be made between engineers and managers in this research as is sometimes done in engineering ethics literature. I will regard every member of a design team as a designing engineer regardless of their job title or education. I have two reasons for this choice. One, in the Netherlands engineers are usually not regarded as professionals in a formal sense. It would be difficult to indicate who is a professional engineer in the Netherlands as engineers are not licensed or certified. Having taken a degree at a University of Technology in the Netherlands gives you the right to use the title “Ingenieur”. There are some professional organisations for engineers but not all engineers are members and some are members of more than one professional organisation. Moreover, some professional organisations are open to everyone doing a certain type of work regardless of whether that person is entitled to use the title “Ingenieur”.1 The
second reason is that in design processes engineers, managers and marketing specialists cooperate to design a product. It would be artificial to exclude some ——————————————————————————————————
Engineering ethics and design processes
23
persons who clearly cooperate in the design process just because they have another job description or another educational background.
2.1.1 Ethics in design processes
Engineering design is an interesting topic for research from the point of view of engineering ethics because design is one of the core activities of engineers. Moreover, technology has social and ethical implications because of the kinds of products produced, as the outcomes of design processes.2 Only recently has more
attention been given to ethics and engineering design [Lloyd and Busby, 2003], [Devon et al., 2001], [Van de Poel, 2001]. Interesting developments in ethics and design can be found in the field of software design and computer ethics. Efforts to incorporate values into the design of software have been labelled “value sensitive design”.
Lloyd and Busby use empirical data to describe how engineers deal with ethical issues in design [Lloyd and Busby, 2003]. They use three main ethical theories to see whether reasoning and argumentation during the design process fit within these theories. They refer to the three ethical theories as “consequentialism”, “deontology” and “virtue ethics”. They looked at all reasoning, not just at reasoning about issues that are clearly ethical like safety [Lloyd and Busby, 2003, 514]. They relate, for example, reasoning about making a better product to consequentialist reasoning. They conclude that, contrary to their expectations, consequentialist reasoning is not prevalent in engineering design. Engineers also use deontological reasoning and engineers identify what Lloyd and Busby call virtues of engineers like collectivity, consistency and emphasising evidence. Lloyd and Busby have considered normal day-to-day situations in which design decisions are made. According to Lloyd and Busby a great number of small design decisions that each seem to be ethically neutral, can add up to ethically relevant consequences:
‘Although it is simply a fact that not much of engineering designing is specifically about what one might normally consider to be ethical issues, the products of engineering design- and particularly the use of those products- undoubtedly are.’ [Lloyd and Busby, 2003, 514]
In many design processes, ethical problems are indeed difficult to recognise and less specific than some of the examples given in the literature on disasters. I agree with Lloyd and Busby that in every design process “smaller” ethically relevant decisions are made. I think however that it is problematic to regard all decisions as being possibly ethically relevant. Some values like, for example, efficiency are not moral values (see section 1.1.1). Efficiency is therefore not ——————————————————————————————————
necessarily an ethical issue. Decisions regarding efficiency can, however, be ethically relevant if they are related to, for example, sustainability. Making a more energy efficient product is ethically relevant because it is a part of designing a more sustainable product. The same holds for trying to design a product as simply as possible. Simplicity is a normative term but it need not be a moral term. Decisions concerning simplicity are only sometimes related to moral values. Simplicity might be an ethical issue if it is related to ease of operation. A simple product can probably prevent accidents related to unintended misuse. If operation of a machine requires a complex procedure, there is a chance that operators will make a mistake when carrying out the procedure. Another issue related to simplicity might be that a simple product can be used by everyone, unlike some video recorders or microwave ovens that people find too difficult to use. Simplicity can therefore sometimes be related to moral values but this need not be the case.
In contrast to Lloyd and Busby, who studied the (ethical) reasoning that engineers use in design processes, Van de Poel distinguished five actions during the design process that may be ethically relevant.
‘1) The formulation of goals, design criteria and requirements and their operationalisation.
2) The choice of alternatives to be investigated during a design process and the selection among those alternatives at a later stage in the process. 3) The assessment of trade-offs between design criteria and decisions about the acceptability of particular trade-offs.
4) The assessment of risks and secondary effects and decisions about the acceptability of these.
5) The assessment of scripts and political and social visions that are (implicitly) inherent in a design and decisions about the desirability of these scripts.‘ [Van de Poel, 2000, 3]
Van de Poel’s approach would imply that, for example, the formulation of requirements is an action that can be expected to be done during design processes. Formulating requirements can be ethically relevant, for example, if safety requirements are formulated. These requirements need to be operationalised and this operationalisation is also ethically relevant. Different alternatives that score differently with respect to different requirements and different operationalisations of requirements may have to be assessed. Trade-offs between different requirements may have to be made. In accordance with Van de Poel’s approach these actions can all be ethically relevant if related to moral values, and therefore these are included in this research.
The concept of value sensitive design has been developed within computer ethics and human computer interface design. According to Friedman and others:
Engineering ethics and design processes
25 ‘Value Sensitive Design is a theoretically grounded approach to the design of technology that accounts for human values in a principled and comprehensive manner throughout the design process.’ [Friedman et al., 2003, 1]
This definition does not imply that value sensitive design is only applicable within software and computer design. Yet, the concept has until now been mainly used within these fields and not with regard to the design of other kinds of technology [www.nyu.edu/projects/valuesindesign/index.html]. In using a philosophical analysis of values and sociological research into the use and development of technology, value sensitive design is an attempt to make software designs that account for moral values like privacy and autonomy.
The research under taken here may be considered research into the way engineers deal with moral values during engineering design processes and therefore research into value sensitive design. There is, however, a difference: whereas researchers into values sensitive design are trying to develop a method for dealing with moral values, I will concentrate on describing how engineers deal with ethical issues like safety and sustainability. Another difference is that some ethical issues that are very important in software design like, for example, privacy and identity are not that important for my case-studies of engineering design.
2.2 Design
The features of design problems and design processes relevant to the topic of this thesis are presented in this section. At the end of this section I will present the conception of design processes used in this work.
2.2.1 Design process
According to Cross, the design process can be seen as a process in which products or tools are created to suit human purposes [Cross, 2000, 3]. The starting point of a design process is usually some stated or perceived customer’s needs. A material structure that meets these functional requirements is designed.3 The design process is usually constrained by economic and time
restrictions. A design should be finished by a certain date and the costs of the whole design process should not exceed a certain amount of money.
In the literature on design methodology a lot of different models of design processes can be found ([Cross, 1989], [Roozenburg and Cross, 1991] and [Baxter, ——————————————————————————————————
3 One could also speak of the design of an organization in which an organizational structure is
1999]). Cross presents a model of the design process that consists of three phases: generation, evaluation and communication. A concept is generated in the first phase of the design process. A designer needs to understand the design problem and to find possible solutions for it; this usually happens simultaneously. Possible solutions help the designer to get a better understanding of the design problem. The concept is evaluated in the second phase. During the evaluation, a decision is made as to whether the possible solution meets the requirements. The concept is adapted in an iterative process. Often, more than one iterative step is necessary because adaptation of a part of the design can lead to problems in other parts of the design. The design is communicated to the people who are responsible for production in the third phase. Drawings, computer drawings and descriptions of the design are used in this communication [Cross, 1989].
Another more detailed model is proposed by French [cited in Cross, 1989, 21 –22]. French divides the design process into four activities:
• analysis of the problem • conceptual design • embodiment of schemes • detailing.
An analysis of the design problem should lead to a clear statement of the problem. The requirements and constraints are formulated in this phase. The designer searches for different possible solutions and makes schemes of them in the conceptual design phase. In the next phase, embodiment of schemes, a choice is made between the schemes. The scheme is further detailed in the detailing phase.
Although there are different models that can be used to divide the design process into different phases and use different terms to name the phases, there are similarities between the models, (see also [Roozenburg and Cross, 1991] and VDI 2221: Systematic Approach to the Design of Technical Systems and Products cited in [Cross, 2000, 39]). The design process can grosso modo be described as follows. The goal, requirements and constraints are defined at the beginning of the design process. This is sometimes done by the customers or in co-operation between customer and engineers. After this a creative part follows in which concepts are generated and evaluated. In the next phase, one concept is chosen and that concept is further detailed. Finally, drawings and descriptions of the design are made for the production of the product. The design process is not a linear process; it is iterative. It may always be necessary to go back one or more steps and then move forward again.
Engineering ethics and design processes
27 2.2.2 Design problems
If design problems are problems in which the requirements alone determine the solution then engineers can say that they are not responsible for ethical issues because the requirements determine everything and the customers define the requirements. Some authors maintain that engineers are not, and should not be, involved in the formulation of design requirements, criteria or goals [Florman, 1983]. According to Florman, the formulation of requirements and goals is ethically relevant, but this should not be done by engineers. Managers, politicians, customers etc should formulate the requirements. In this line of thinking, the task of engineers is to discover what is technologically the best solution given certain requirements. This task is seen as ethically neutral. Ethical questions may arise in the user phase when technologies are used for certain purposes and produce certain (social) effects. According to Florman these ethical questions concerning use are also outside the scope of the engineers and should be solved by the user (see figure 2.1). In this model, the sole responsibility of engineers is to carry out a task formulated by others in a competent way.
Figure 2.1: Division of labour with respect to engineering design if design problems were well-structured problems in which the requirements fully determine the solution, after [Van Gorp and Van de Poel, 2001].
Design problems are, however, usually not problems where a clear set of requirements is available that completely determines the solution. Design problems are more or less ill-structured problems ([Simon, 1973] and [Cross, 1989]). Simon states that in the ill-structured problem of designing a house:
‘There is initially no definite criterion to test a proposed solution, much less a mechanizeable process to apply the criterion. The problem space is not defined in any meaningful way,’ [Simon, 1973, 311]
For Simon the main characteristics of an ill-structured problem are that the solution space is not well-defined and that there is no criterion to test different solutions and decide which is best. Cross gives the following characteristics of ill-structured problems:
‘1. There is no definite formulation of the problem. 2. Any problem formulation may embody inconsistencies. 3. Formulations of the problem are solution-dependent.
4. Proposing solutions is a means to understanding the problem. 5. There is no definitive solution to the problem.’ [Cross, 1989, 11-12]
Some design methods require that engineers formulate the requirements and solutions separately and independently, but this is impossible if design problems are ill-structured. In a redesign of an existing design it might be possible to formulate most of the requirements at the start of the design process but this is not a definition of the requirements independent of the solution. The solution space is, in these cases, limited because a redesign is made; certain features of the product will remain the same. Other design problems aiming at designing a completely new product are very ill-structured and only some vague requirements can be formulated at the start of the design process. So design problems can be more or less ill-structured.
An example of an ill-structured problem is the following. In the mid nineteen nineties substitutes were sought for replacing CFCs as coolants in refrigerators, because CFCs damage the ozone layer [Van de Poel, 1998 and 2001]. Two alternatives were considered: HFC 134a and hydrocarbons, both have their advantages and their disadvantages. Hydrocarbons are for example flammable and existing refrigerator design needed to be changed if hydrocarbons were used. HFC134a has a long atmospheric lifetime and if released would therefore still damage the environment, although to a lesser extent than CFCs. There were different operationalisations available for the environmental, health and safety criteria. Both proposed solutions scored differently under different operationalisations of the criteria. There was no solution that was best under all operationalisations. No definite criterion was available to say which solution was the better one. This example shows that even for the seemingly simple case of looking for a substitute coolant in existing refrigerator design, there are features of the problem that make it ill-structured.
In cases where a design problem is an ill-structured problem, there may be more than one solution; each of these solutions can be valid. Engineers, in this case, have to make a choice: it is not the case that the requirements will lead to just one solution. At the start of a design process, there may not even be a clear and unambiguous set of requirements. During the design process it may be proved that there is no solution to the ill-structured problem. In some cases it
Engineering ethics and design processes
29
might prove to be necessary to adjust or drop some requirements because no solution meeting all the requirements can be found. So a design problem can be under, or over, determined by the requirements. Either way, engineers need to make choices during the design process for example regarding which requirements can be dropped or which of the possible solutions to the design problem is the best.
2.2.3 The design process as a social process
Most designs are made by a team of engineers. Designing is in these cases a social process. Choices are made in, and by groups of people. During the design process, communication, negotiation, argumentation, (mis)trust between engineers and power differences between engineers influence the design. This has consequences for design research as the design process should be conceptualised as a social process. There is some research into actual design processes with design teams [Bucciarelli, 1994], [Lloyd and Busby, 2001], [Lloyd, 2000] and [Baird et al., 2000]. Bucciarelli describes the design process as a social process in which negotiation is necessary:
‘Contemporary design is, in most instances, a complex affair in which participants with different responsibilities and interests…. must bring their stories in coherence’ [Bucciarelli, 1994, 83]
The different engineers, with their different educational backgrounds and experiences, will all conceive the design task differently. Take for example the cage construction and bodywork of a car. A mechanical engineer looks at stresses and strains within the cage construction and bodywork of a car. He or she tries to design them in such a way that stresses and strains remain low during normal use and absorb energy during a crash. An aerodynamics engineer might look at the same bodywork and sees a body that needs to have a low frontal area and a low drag coefficient. Although both the mechanical and the aerodynamics engineer look at the same parts they see something different and think of different requirements the parts should meet. All these different views have to be ‘brought in coherence’ [Bucciarelli, 1994], just like all the parts have to fit and function together. This ‘bringing into coherence’ is done as a process of communication and negotiation.
Other authors also recognise the importance of social processes during the design process and stress the importance of communication. Lloyd stresses the importance of storytelling within the design process [Lloyd, 2000]. Engineers construct stories during the design process. These stories are used to come to a common understanding:
‘Engineering design as a social activity consists in the construction of social agreements. We have observed storytelling to be a mechanism that aids this construction.’ [Lloyd, 2000, 370]
The stories can be used within the entire company or within single departments. The whole company or design team might know the story of a previous design failure. Stories of specific difficult customers might only be known in the sales department. Knowing the stories is part of being part of the design team or department. Stories can therefore be inclusive or exclusive in their use [Lloyd, 2000].
Baird and others conducted an ethnographic study at Rolls-Royce Aerospace [Baird et al., 2000]. They conclude that personal interaction between engineers is crucial for information to be disseminated throughout an organisation. Engineers who know each other from other projects tend to ask each other for advice when working on new projects. At the beginning of the design process more experienced senior engineers are very important. They help younger engineers and point them ‘to the sequences and sources of expert opinion they should seek’ [Baird et al., 2000, 350]. According to Baird et al., this helps with structuring the design problem.
2.2.4 Organisation of the design process
A division of labour exists in most design processes. For example, the design of a car can be divided into the design of the drive shaft, the engine, the seats, the electronic systems, the suspension and the styling of the car etc. The partitioning of a design team into smaller design groups responsible for a part of the design is, from an ethical point of view, noteworthy because it may lead to the problem of many hands [Bovens, 1998] and [Thompson, 1980]. This problem presents itself with regard to active and passive responsibility. Passive responsibility is responsibility after something has happened: being held accountable. Active responsibility refers to being or feeling responsible for something or some task [Bovens, 1998].
With regard to passive responsibility, the problem of many hands is the following. It might seem to be quite clear who is officially responsible for what as this depends on formal job descriptions in organisations, but in practice it is very difficult to point out the people responsible for acts of organisations that have caused damage. Organisations are often opaque to people outside of the organisation. It is not clear who is responsible for what and who was able to influence a certain decision. Above this, when an organisation is organised hierarchically, people lower in the hierarchy can indicate that those higher in the hierarchy are responsible while those higher in the hierarchy claim to have no knowledge of the situation.
Engineering ethics and design processes
31
With regard to active responsibility, the problem of many hands can be seen when no one feels or thinks that he or she is responsible for certain issues. If issues are not specifically part of someone’s task description, everyone can avoid taking responsibility for them. These issues may then be neglected in the design process.
In a paper on the relationship between how companies are organised and harm they cause other people, Darley studied a case of the design and testing of landing gear for a military aircraft that failed during landing after test flights [Darley, 1996]. People in the company knew that there were calculation errors and because of these errors there was a large chance the landing gear would fail. Certain social mechanisms made the people in the organisation actively conceal the calculation errors for their customers and tinker with data in certifying documents. In this case people felt either forced by their superiors or felt they were already too involved and caught up with the tinkering of the data to stop the concealing of calculation errors.
Darley also points to the way a decision is framed. Stopping a production process or changing a design requires action, while going on is often seen as not requiring action. Action has to be defended to other people while in the case of doing or changing nothing, such defence is not required. Action is only taken when there is proof of harm. This frames the decision in an “innocent until proven guilty” way. This is different from using a precautionary frame for the decision, where no harm has to be proven. A suspicion of harm can be enough to warrant acting against it. Framing a similar decision differently can lead to different actions [Darley, 1996]. This might be relevant for design processes, especially when decisions have to be made to stop or go on with a design process. The aspects of design processes mentioned above are all relevant from an ethical point of view and therefore they were included in the conception of design processes used in this thesis. In view of the foregoing, I regard design processes
as organised social processes which aim at solving more or less ill-structured design problems in this work. All of these aspects mentioned above were used to support
data-collection in the case-studies. For example, in the case-studies attention was paid to the organisation of the design team and the social processes within the design team because this helped me to determine who was involved in what decisions when dealing with ethical issues.
2.3 Characteristics of design processes in relation to ethical issues4
So far, some very general characteristics of design processes have been discussed. When studying ethical issues in engineering design it may be useful to make further distinctions between different kinds of design processes. It might be expected for instance that during the design process for a bolt for a car wheel, different ethical decisions have to be made than those that have to be made during the design process for a completely new personal transportation device. Ideas drawn from the literature are used to characterise the different kinds of design processes. Working hypotheses are formulated in the next chapter based on these ideas. I conclude this chapter with a discussion of some preliminary ideas regarding what this research, and its results, might contribute to discussions about the moral responsibility of engineers during design processes. These ideas will be further elaborated in the last chapter of this thesis.
2.3.1 Design type and design hierarchy
Let us go back to the example of a design for a wheel bolt on a car as opposed to the design of a new personal transportation device. The bolt design has to comply with dimensional constraints, safety norms, standards, financial constraints etc. It is a small part of a known product. Moreover, most designs of bolts are redesigns of existing bolts. Norms, standards or dimensional constraints are absent for a new personal transportation device, or it is questionable whether existing norms, standards or dimensional constraints can or should be used. The design problem for a new personal transportation device is more ill-structured than that of designing a bolt. The reason why the design problem for the bolt is better structured than that for the new transportation device is that there are more external constraints pertaining to the design of the bolt.5 I use the term
“external constraints” here for all constraints that are taken for granted during a design process. Some of these constraints may be set by the engineers at the start of the design process, for example by the engineers deciding to redesign an existing bolt instead of designing a new one. Other constraints are set by other stakeholders and not the engineers, such as the customer’s requirements, governmental regulations or codes and standards. These external constraints are usually already operationalised into specific and clear requirements. In a redesign it is also usually obvious how these requirements can be implemented.
——————————————————————————————————
4 The ideas in section 2.3.1 and section 2.3.2 are based on the paper ‘The need for ethical
reflection in engineering design; the relevance of type of design and design hierarchy’ [Van de Poel and Van Gorp, 2006].
