An Ethically Neutral Decision Aid
Department of Agricultural Economics,
University of Gšttingen
Platz der Gšttinger Sieben 5
D- 37073 Gšttingen, Germany
Tel: +49-551-394805. Fax: +49-551-394812. Email: firstname.lastname@example.org
Decision Analysis can not be done without some reflection on the normative grounds of the
analysis. The aim of decision analysis is to recommend an action. This recommendation con-
stitutes a normative act. There are mainly two economic decision analyses: Cost-benefit
analysis and multi-criteria-analysis (MCA). The latter uses evaluations on several criteria to
recommend a decision. The discussion about the superiority of one of these instruments is
usually based on the question whether the interests of the decision maker are more adequately
integrated into a mono- or multi-criteria analysis (see for example MUNDA 1996). In my paper
I go one step further: I take it as evident that in certain contexts − which are here environ-
mental decisions in the framework of sustainable development − interests are more adequately
reflected by several criteria. I ask how these criteria would have to be constructed to integrate
all possible arguments of the decision maker. The answer to this question, an explicitly norma-
tive discourse, leads to the idea of the implementation of an ethically neutral MCA. The cho-
sen ethical foundation, pluralism, enforces the positive aspects of participation. The final
paragraph illustrates the argument through a (still hypothetical) example.
Decision analysis, ethics, multi-criteria analysis, decision making or decision aid, participation,
1 I am indebted to Rainer Marggraf for his structuring aid and to Thomas Sikor for his language skills. Faults
remain my own, though.
We analyse decisions in order to find better solutions in complex situations. Before starting
the analysis we have to state, though, why some decision would be better than another. A
statement of this kind is not possible without having decided on which normative grounds the
analysis rests. If the grounds are not clear then it is not possible to judge whether or not a so-
lution is better than another. Any decision might be ÔbetterÕ than another depending on the
point of view. In decision analyses, it is the point of view of the decision-maker or of society
which should count and not the analystÕs point of view. The context of the decision decides
whether we may take the point of view of the decision-maker or a (supposed) point of view
of society. Questions appearing in business and mainly concerning business (for example port-
folio management) might not need normative discussions. The criteria do not seem to be ethi-
cally relevant, as it is only profit which counts.2 On the other side, decisions in the public
domain which involve several individuals, living or not-living humans or non-humans, busi-
ness, social, religious or quasi-religious spheres etc. are ethically relevant. Here it is not evident
why any decision might be better than another (except for trivial situations). Not only the
results but also the choice of the method of decision analysis depends on the decision about
the normative foundation of the analysis. Consequently, deeper insights are necessary into the
different arguments in ethics and, in our case, in environmental ethics.
The main question requiring clarification is whether it is only the consequences of a decision
which should count (consequentialistic ethics), whether it is the decision (process) which is
decisive for the goodness of a decision (deontological ethics), or whether it is necessary to
combine both approaches.
On this issues I take the (normative) point of view that it should be possible to integrate all
arguments of the people affected by a decision a priori. This seems to be a necessary condition
for a democratic decision-making process (hereby excluding unaffected humans and non-
humans). This point of view is in accordance with a WEBERian understanding of science: valua-
2 This is debated, though. See the discussions about ethical and ecological investment.
tions, assertions of norms or objectives are not accessible to a scientific rationality. The en-
counter with fundamental values or norms limits the possibility to rationalize valuation state-
ments via hierarchisation etc. (NIDA-R†MELIN 1996: 47). This means in our context that deci-
sion analysis may not take side for or against certain fundamental norms or values. Instead, the
chosen instrument should be as open as possible for different normative arguments.
In the following two chapters I will examine whether decision-making or decision-aiding
methods can fulfil this democratic-normative condition in the context of environmental policy.
Normative grounds of decision-making instruments
It would be ideal if we could identify a method which would be able to supply the best deci-
sion in all areas without any doubts on its empirical or ethical legitimacy. Such a decision-
making method would have to be based on an indisputable ethics which helps judge an action
following only one master-principle (monistic ethics). If an ethics, however, contains different
juxtaposed rules decisions in complex fields will usually be governed by different rules with no
master-principle ordering the application areas of the rules. Main candidates for monistic eth-
ics are found in two ethical schools: consequentialistic and deontological ethics.3 Most MC
analysts implicitly or explicitly base their assumptions on consequentialist foundations, i.e.
they compare the world after the action with the world before it and value the differences with
the help of preferences elicited before the action. The preferences might cover human-related
areas such as social or economic criteria and nature-related areas. Human preferences about
states of the world are in the centre of valuation, which means in the centre of the decision
Another great school of ethics is deontological ethics. According to this ethics, an action is not
judged according to its consequences. The action itself is judged by a rule. A well-known deon-
tological ethics is Kantian ethics.
3 I do not consider contractarian ethics which can not apply in environmental fields nor RAWLSian ethics which
may be applied (see SINGER 1988, with some difficulties, though − see PFORDTEN 1996: 189ff.). It will be clear
that the inclusion of RAWLSian ethics would not change my argument.
Empirical economic surveys (SPASH AND HANLEY 1995, STEVENS ET AL. 1991) and analyses of
political decision-finding processes (see BOOTH 1994: 173ff.) show that people use both con-
sequences- and action-based judgements as arguments on behalf of nature conservation.
Consequently, as some of the affected people do not only Ônot preferÕ either the decision-
finding process, the decision, or its consequences, but find some component of them ÔevilÕ,
utilitarianism is not valid as a foundation for a decision-making method. Scientists from differ-
ent fields agree that it is due to its utilitarian (which is often used synonymous to consequen-
tialistic) foundation that cost-benefit-analysis cannot do the decision-making job (see, e.g. SEN
1979, OÕNEILL 1993, SAGOFF 1994, STIRLING 1997). For mainly the same reasons, an only
preference-based MCA is not able to do it either (even if the MAUT-based approaches of
BEINAT 1997, BEINAT AND RIETVELD 1997 or others can be considered as advances to standard
decision-making in environmental policy).
On the other side, any ethics which excludes consequentialistic thinking rules out important
arguments for or against certain actions and do not fulfil our main condition for decision-
The concept of discursive ethics (a member of the family of deontological ethics) is quite seri-
ously discussed as a normative foundation for public decisions − but mainly in contributions
in fields outside philosophy. At first glance, it is an approach which might be able to include
all arguments. From an empirical perspective, all arguments are permitted in the rational dis-
course, which constitutes the morally right by consensus in a symmetrical and argumentative
process (KREBS 1997: 271). Its goal positions discourse ethics firmly within deontological
ethics: Normative environmental discursive ethics is no process which aims at finding values
for ecosystems, its objective is to ãestablish moral norms or duties to or vis-ˆ-vis natureÒ (OTT
1995: 328 - own translation), i.e. to judge actions on their goodness. Yet things are more com-
plicated if looked upon from a normative perspective. First, arguments are valid not for the
sake of their own ethical foundation, but because all participants of the discursive process can
accept them. Secondly, the validity of non-anthropocentric arguments is not secured by the
4 The difference beween consequentialistic and deontological arguments can be refound in the discussion about
substantive or procedural rationality (e.g. FAUCHEUX ET AL. 1997).
process. A participant of the discourse has to raise his voice on behalf of non-human entities.
If no one raises his voice it would not be ethically incorrect to mistreat dogs, dolphins etc.
Furthermore, it would be quite impossible to include non-anthropocentric arguments in discur-
sive ethics. A communication-like relation with trees, stones or ecosystems would be neces-
sary for an integration in discursive ethics (HABERMAS 1991: 219ff.), but this integration is not
plausible for most people. It is not possible to integrate the most extreme range of eco-ethical
arguments, i.e. the holistic arguments of, for example, deep-ecology, in discursive ethics.
To conclude, I did not identify any monistic ethics, i.e. an ethics reducing all sorts of argu-
ments to one or few master-principles, which can allow for the whole range of arguments used
in environmental ethics (see for example the journal Environmental Ethics) or in environmental
decision-making. But for decision-making methods we need an ethics permitting to judge all
possible actions coherently to be able to determine the best, second-best, etc. solution without
any doubts on the judgement. Consequently, it is not possible in environmental contexts to
use MC decision-making methods to find an ethically sound and democratically acceptable
Pluralism as a basis for MC decision aid
We have to accept, at least in interpersonal relations, that we have different coexisting argu-
ments without a master-principle being able to order them and decide about their respective
validity and weight. There are, nevertheless, approaches which claim to provide frameworks
for integrating different arguments: Some of them are − without any claim on completeness −
pluralism (STONE 1987), moderate pluralism (WENZ 1993), expressive rationalism (ANDERSON
1993) and practical coherentism (NIDA-R†MELIN 1997). They do not look for a single axiom in
the sense of immediately evident assumptions from which single theorems and criteria would
be deduced. The theories do not want to base their normative content on a speculative founda-
tion of a theory of objective and intrinsic values (NIDA-R†MELIN 1997: 181). Instead, they
part from central and apparently convincing elements of our belief system. These approaches
try to arrange the different elements of individual ethical life systematically and coherently.
An example: I want to write a good paper on ethics and MCA, but I also want to take a walk
on this sunny autumn afternoon, and I feel obliged to stay with my partner to help her with
our children. My decision what to do can best be described in terms of my capability to use
reason for mediating between different duties, wishes and passions.
Extrapolated to interpersonal situations, the pluralistic approaches can also be used to enhance
the degree of coherence in interpersonal decision-making. Using these approaches one is not
sure to find a specific solution. But the set of possible solutions might contain some which are
ÔbetterÕ than a ÔbestÕ solution found through an only superficially accepted MC decision-
making process. When a normative discussion is successful intrapersonal coherence will aug-
ment and there will be less intrapersonal competition between different ethical elements. Indi-
vidual argumentation will be less contradictory and it will be easier to identify consensual and
dissensious fields. In this process new solutions might emerge which would have been hidden
in a more mechanistic process (see BOHNET AND FREY 1994). This optimistic view contrasts
with ROYÕs (1990A: 178): The role that each criterion could and should play ãis, in the main, a
reflection of a system of values, but also of more fragile opinions, which too detailed a discus-
sion will disturb.Ò But BOHNETÕs AND FREYÕs view accords with ROYÕs definition of MC
decision aid (MCDA): It should help the decision actor shape his preferences (1990: 28).
There is a second argument for the use of pluralistic ethics as a basis of MCA, which is used
more often and not based on an ethical rejection of monistic ethics as a foundation of interper-
sonal decision analyses. It is argued (see ROY 1990B) that, in environmental contexts, MC
decision-making is not feasible for several reasons: the fuzziness of the set of feasible alterna-
tives, the absence of well-stated preferences of the (group of) decision actor(s), imprecise data,
institutional or cultural impacts on the quality of the decision etc. For these reasons ROY, in
some contexts, prefers MC decision aid which ãshould help an actor taking part in the deci-
• either to shape, and/or to argue, and/or to transform his preferences,
• or to make a decision in conformity with his goals.Ò (ibid.: 28).
MC decision-making methods need strong computational assumptions on the revealed
preferences which only utilitarian ethics can fulfil. If, for other than pure ethical reasons,
MCA should be used as a decision aid these strong assumptions are not necessary anymore.
There is consequently no need to use an utilitarian rather than any other ethical approach. The
selection of the ethical foundation of MCDA requests an investigation of the question which
approach is more reliable in achieving the aims of MC decision aid. Only an approach which
can accept all possible judgements of the decision actor can either shape, and/or argue, and/or
transform the whole range of his judgements. As his goals consist of elements coming from
different ethical approaches, a monistic decision method offers more limited help than a plu-
ralistic method when making a decision in conformity with his goals.
• If one judges pluralism to be the right normative foundation then decision-making methods
are not possible anymore.
• If one prefers decision-aiding methods then pluralism is the most restrictive normative
foundation which is necessary.
So far the theoretical reflections on ethical foundations of MCA.
Ethically neutral MCA
How could these theoretical reflections be integrated into a decision-preparing method? All
parts of MCA change if one adopts a pluralistic5 point of view: (1) the understanding and aim
of MCA: as discussed above, MCA can only be understood as a decision aid in complex fields
such as environmental decision-making; (2) the criteria: they should respect the differences of
ethical arguments more than those of accustomed sciences; (3) the calculating method: it has to
allow for different scales and for incommensurability between different ethical criteria.
The last point can be treated by the choice of an appropriate method: FUSTIER (1994) dis-
cusses an interesting qualitative approach on the question of scale. Another approach is taken
by MUNDA (1995), who integrates the concept of incommensurability, as well as other out-
ranking methods (see ROY 1990A). It is the second point which remains new and unsolved.
How can we establish criteria?
The classical point of view is formulated by BOUYSSOU (1990: 59): ãIn a multiple-criteria ap-
proach, the analyst seeks to build several criteria using several points of view. These points of
view represent the different axes along which the various actors of the decision process jus-
tify, transform and argue their preferences.Ò They ãshould be understood and accepted by all
the actors of the decision process, even if they disagree on the relative importance that they
would like each of them to haveÒ (ibid.: 61). BOUYSSOU underlines the ethical neutrality of
criteria by demanding that ãthe method allowing to arrive at the evaluation on the criterion for
each alternative É should be as free as possible from elements deeply linked to a particular
value systemÒ (ibid.). But he himelf uses terms deeply linked to utilitarianism in concluding
that a ãcriterion thus appears as a tool allowing to Ôsum upÕ a set of evaluation on conse-
quences related to the same point of view so as to be able to establish partial preferencesÒ
(ibid.: 62). For him, as for other MC analysts, the points of view are not ethical issues but
criteria such as Ôimpact on environmentÕ, Ôsocial impactsÕ, Ôeconomic impactsÕ etc. Does a
criterion like Ôimpact on environmentÕ fulfil his basic guidelines, i.e., is it ãa point of view
along which it seems adequate to establish comparisonsÒ (ibid.: 62)? On the surface it seems
so. Yet scientific comparisons depend on value judgments, which are impossible without a
normative foundation. Is it feasible to collect data on Ôeconomic impactsÕ without a deep link
to the efficiency paradigm and to utilitarianism? I do not think so. Only if we explicitly ac-
knowledge the normative foundations of criteria can we make the use of decision analysis more
fruitful. The use of empirically derived criteria, which carry normative implications, will, how-
ever, not use all the capabilities of MCA.
STONEÕs pluralistic ethics offers a possibility for the building of normative criteria. He pro-
poses to follow each ethical approach separately, for example using land ethics, utilitarianism
etc. as a single approach (STONE 1987: 115ff.). His approach ãis characterized by alternations
among several ethical theories. Although each such theory is accepted in its entirety, the range
of application of each is limitedÒ (WENZ 1993: 65). Apart from the fundamental difficulties of
his pluralistic approach to decisions (see WENZ 1993 and CALLICOTT 1990), it does not seem
practical or feasible that each person is able to play the game: How would I decide if I was a
utilitarian, Kantian, deep ecologist etc.? Instead, we should recognize the central elements in
the ethical foundations of the affected people and discuss them in an interactive way so that
5 In my paper, I refer with the word ãpluralisticÒ to the ethical foundation of a MCA, not to the understanding of
every participant in the discursive process has the possibility to structure his own arguments
coherently. In the ensuing normative discussions in the decision-preparing or decision-taking
group, ecological sub-criteria, for example, would be integrated into criteria such as intragenera-
tional fairness, respect for non-human living beings, intergenerational fairness,6 self apprecia-
What would change? First, scientific advisors would have to consider the ethical values af-
fected by proposed measures in their respective fields. Secondly, the discussion in the deci-
sion-finding process would be more about the values of importance and not about the concrete
facts, which are only made meaningful by a whole range of ethical values. These two points
would imply an open, truthful, and credible style of exchange, which would be most unusual.7
The discourse could be organised along the idea of OÕHARAÕs descriptive discourse ethics as a
way of explicating normative judgements. OÕHARA (1996: 97f.) departs from normative dis-
course ethics and defines her ã(d)iscourse as descriptive ethics (which) does not aim at estab-
lishing ethical norms or at seeking to identify conditions for ethical judgement, but simply
acknowledges that people are ethically motivated, just as they are socially, culturally, eco-
nomically, or ecologically motivatedÒ. Descriptive discursive ethics has so far not claimed to
integrate all ethical judgements into a normative judgement which could substantiate a decision.
It can thus be understood as a kind of moral psychology aiming at integrating normative values
into the decision-making process, but it can not be a normative basis for MCA.
In a broader sense the proposed process could apply OÕRIORDANÕs (1997: 170) definition of
valuation: ãvaluation is achieved by the process of reconciling to the interests of other legiti-
mate player, and in so doing, revealing inner feelings about the moral qualities of a collectively
reached outcome.Ò STIRLING (1997: 190) accords with him in stating that ãno purely analytical
procedure can fulfil the role of a democratic political process. In other words, there is no
uniquely ÔrationalÕ way to resolve contradictory perspectives, divergent values or conflicts of
administrations as pluralist or rationalist (cf. FIORINO 1995: 223ff.).
6 It might be more equitable to use a multigenerational majority rule voting, as proposed by KUNSCH (1994).
But if his norms of intergenerational fairness are not accepted then we would need an Ôequitable dictatorÕ to
How could the implementation of an "ethically neutral MCA" look like? In our actual political
systems, one or few democratically or bureaucratically legitimised persons generally make
public decisions. Implementing MCA in the existing decision structures would imply that
these persons would decide after receiving adequate information about the consequences of
possible actions and after a discussion with a supporting decision analyst about their personal
ethical values. There would be no need for an interpersonal ethical discourse with people who
are unknown to the decision-making body. However, this process does not meet the normative
requirements of ethically neutral MCA.
Beside the well-known arguments for public participation in general and in the case of nature
conservation in special (see for example WELLS AND BRANDON 1993), the use of ethically neu-
tral MCA requires participation for two further reasons:
(1) It is dubious whether the elected or bureaucratic representative could represent the ethical
criteria of the affected people;
(2) A non-participatory process does not use the ability of ethical MCA to generate better
ad 1. While a decision-maker could take into account material consequences for affected peo-
ple, it would be very difficult for her to decide on the foundation of ethical values which
are strange to her. The point is still more valid in new contexts, as pluralistic values
emerge in the process of deliberation. This process has to take place and results would
be of limited benefit if the deliberation had only been hypothetical. The aim of the par-
ticipatory process is ãto form a collective understanding and to construct a coherent ba-
sis for choiceÒ (VATN AND BROMLEY 1994: 143).
ad 2. It is this process of deliberation which creates the usefulness of pluralistic MCA.8 The
information transmitted and created in a public discussion helps individuals to clarify
their respective ethical convictions and to state commensurabilities and incommensura-
7 A friend of mine suggested that this is exactly the reason why it is not done.
8 I do not want to hide the difficult aspects of group decision-making: see for example JELASSI ET AL. (1990).
But I argue that they are more than compensated by the positive aspects.
bilities (BOHNET AND FREY 1994: 345). Only after having made clear the individual ethi-
cal values, new or unconventional emerged solutions can be accepted or rejected on ra-
tional grounds. If the analyst can conduct the discussion well then individuals might
combine moral norms (as equity) and pragmatic norms (as economic efficiency) in such a
way that well structured arguments result from the communicative process. The new
structures enable the development of new options. By this way the probability of col-
lectively irrational solutions is diminished and public decisions allocate public goods bet-
ter. Suboptimal solutions appear less often, and the public will trust its administration
to a greater extent. The view of pluralistic ethics as a process accords with the view of
sustainable development as a process rather than an end state (ACHTERBERG 1996: 158).
Which forms of participation could be useful? This depends most heavily on the problem it-
self. RENN ET AL. (1995b) classify environmental conflicts into three groups: Some can be un-
derstood as conflicts due to incommensurable values. While these tend to be the conflicts with
the highest degree of complexity and/or conflict intensity, two other types of debates are rele-
vant: debates about factual arguments and debates about public confidence in institutions to
deal with environmental threats. The classification is not always apparent as there is a strong
tendency to re-frame the conflicts: business and government tend to present value-based con-
flicts as factual or confidence conflicts. ãThis is an attempt to focus the discussion on techni-
cal evidence, in which the agency is fluent (DIETZ ET AL. 1989). Citizens who participate are
thus forced to use first level (factual) arguments to rationalise their value concerns. Unfortu-
nately, this is often understood by experts as ÔirrationalityÕ on the part of the publicÒ (RENN
ET AL. 1995B: 357). Consequently, the attempt to find a value-based decision aid which might
be implemented in participatory structures can be understood as an aid to the fluency of agen-
cies in value matters.
Not all models of participation seem to be equally appropriate. Existing proposals for citizen
participation have to be examined upon their ability to be used as a frame for a pluralistically
based MCA. Here I ignore some of the approaches described in the integrative volume of
RENN ET AL. (1995a) which are either not easily repeatable (Dutch study groups), in sharp
contrast with the pluralistic approach (compensation/benefit sharing), too specific (citizen
initiatives) or too close to other models (citizen juries). The following list only considers
methods which meet a sufficient number of criteria guaranteeing a fair, competent and value-
open participation process:
• Citizen Advisory Committees: An often publicly sponsored consensus-oriented forum for
participation with representatives of the most important interests, who have little repre-
sentative power (LYNN AND KARTEZ 1995). Frequent use of MCA methods is reported
(VARI 1995: 108).
• Planning Cells (Citizen Panels): small groups of representatively chosen citizens who work
on a problem with the help of experts (RENN ET AL. 1995B: 343). There are often some
doubts about their legitimacy (RENN ET AL. 1993: 204f.).
• Regulatory Negotiation: representatives of interests affected by a proposed rule negotiate
about language and content of the rule (FIORINO 1995: 225). Public criticism and limited ac-
ceptance are due to the resented intransparency and a suspected reciprocal Ôhand-washingÕ
(RENN UND WEBLER 1996).
• Mediation: disputing partners try to find a consensus with the help of a neutral agent. Ne-
gotiations take place confidentially and are difficult to evaluate (NOTHDURFT 1995).
In practice, the different models do not appear as pure forms, but as mixtures of different ele-
ments. Combinations of at least two models are recommended often (e.g. HADDEN 1995: 250,
VARI 1995: 113).
It is possible to include explicitly normative decision analysis in all models, but it does not
seem necessary in conflicts that mainly address questions of knowledge and expertise. Ethi-
cally neutral MCA is most urgent in cases of conflicts due to different worldviews and values.
Having seen the attempts to downsize the conflicts, an explicit openness to normative ques-
tions can be very useful in conflicts about experience and trust in institutions. This said, I will
pick one example of participatory decision-making which is based on the Planning Cell (RENN
ET AL. 1993). After a critical description of the used decision process, I will include ethical
MCA instead of the multi-attribute utility-based decision-making which has been used. The
Planning Cell with its openness to value and trust conflicts seems to be a useful model for
showing the changes in participatory decision-making which occur due to the use of ethically
neutral MCA. Here, I do not want to discuss the major advantages or the disadvantages of
Planning Cells, but use it merely as a fair example.
RENN ET AL. (1993) start from the premise that a ãmodel is needed that combines technical
expertise and rational decision-making with public values and preferencesÒ (p. 189). In the
first step, their model includes stakeholders in order to identify and select concerns and
evaluative criteria. The stakeholders build, with the help of value-trees, a list of hierarchically
structured values. In the second step, experts identify and measure the impacts of the different
decision options. ãThe desired outcome is a performance profile for each optionÒ (ibid.: 191).
In the third step, randomly selected citizens aggregate and weigh the expected impacts to ob-
tain a priority list of the options and to give recommendations to the political actor. The steps
are not hermetically separated from each other. All actors involved ãplay a role in each step,
but their influence is channelled to the type of knowledge and rationality they can offer bestÒ
(ibid.: 193). Experts use the concerns and criteria of the stakeholders who might comment later
on the work done by the former. Yet stakeholders are prevented from bargaining. They cannot
do so because the tasks of aggregation and weighing are assigned to citizens, who use the in-
formation condensed in the value tree (and in its variations). The numerical results in the value
tree ãare not used as a expression of the final judgement of the participant but as a structuring
aid to improve the participant's holistic, intuitive judgement. By pointing out potential dis-
crepancies between the numerical model and the holistic judgements, the research team encour-
ages the participants to reflect upon their opinions and search for potential hidden motives or
values that might explain the discrepancy. The final recommendations are always based on a
holistic judgement by individuals or groupsÒ (ibid.: 198).
This approach of the planning cell including stakeholders responds to the critics in respect to
the democratic legitimisation of planning cells.
As an example for their procedure RENN ET AL. (1993: 201ff.) cite a participatory process
which could help define the ÔbestÕ energy policy for Germany on the basis of four predefined
options. This process is also described by KEENEY ET AL. (1990), who concentrate on the
15 Download full-text
weighing process to elaborate a multi-attribute utility function which I will describe shortly.
Stakeholder representatives identified eight independent criteria: financial, technical and mate-
rial requirements; security of energy supplies; national economic impacts; impacts to the natu-
ral environment; health and safety; social impacts; political impacts; international impacts.
The staff of the study aggregated the criteria in only one value tree. Experts added the relevant
sub-criteria which, in some cases, were reformulated by the citizens. The citizens evaluated
intuitively the four different energy futures on a scale from 0 (worst) to 100 (best). After elici-
tating the trade-offs between the different criteria, and after normalising the weights, the staff
were able to calculate a multi-attribute utility evaluation of the four energy futures for each of
the participants. In the following discussion, participants were asked to resolve the inconsis-
tencies between the model and their first evaluation. Most subjects achieved ordinal consis-
tency between the two evaluations. ãApparently, the value forum and the multi-attribute util-
ity modelling had a significant influence on their thinking É they felt that these two sets of
evaluations should agree, refused overwhelmingly the Ôeasy solutionÕ to reject the objectives or
the model, and worked hard, up to three hours, to achieve ordinal consistency. In this recon-
ciliation phase the primary changes were in weights and in intuitive evaluationsÒ (1028f.).
The example questions the above cited statement of RENN ET AL. (1993: 198) that the numeri-
cal results in the value tree ãare not used as a expression of the final judgement of the partici-
pant, but as a structuring aid to improve the participants holistic, intuitive judgement.Ò Their
interpretation of ÔholisticÕ and ÔintuitiveÕ is rather unusual. After having analysed and com-
partmentalised an object, an intuitive and holistic judgement has become impossible. The new
judgement might be more adequate in some respects, but it can not be holistic and intuitive. It
is surely not wrong to question the first, ãintuitiveÒ evaluation. But it is questionable whether
it makes sense or not to start the process with a model which is based on commensurabilities.
Asking for an intuitive evaluation creates another context than questions about trade-offs.
Following SUNSTEIN (1993: 247), it would be most unusual, if both evaluations would accord.
Based upon the three-step model of RENN ET AL. (1993), I propose a four-step model to im-
plement ethical MCA. Figure 2 shows the basic concept and the elements of the four-step