Ethical Issues Facing the Food Industry

by

Paul B. Thompson
Center for Biotechnology
Policy and Ethics
Texas A&M University
College Station, TX

The array of ethical issues facing the food where ethical concerns bear upon food biotechnol-
industry is extensive. It includes fair and just ogy, the paper makes no attempt to survey the full
treatment to food industry employees, especially range of ethical concern. What is more, the paper
as disproportionate numbers of minorities take does not present a normative argument favoring
jobs in food processing and food service. Issues one policy option rather than another. The idea
of food distribution and hunger continue to be that ethics requires a particular set of policies for
important. In the calendar year 1992, however, food biotechnology is not argued in this paper.
these issues pale in comparison to those raised by Instead, the purpose is to examine how ethical
food safety and labeling. The impetus for this arguments establish a burden of proof for policy
issue is multiple, The Nutritional Food Labeling evaluation. The thesis is that effective policy
Policy Act has mandated new labels intended to making requires an ability to understand how
provide consumers with consistent information on different types of ethical criteria bear on policy,
ingredients that will be useful in dietary planning. Insensitivity to contrasting ethical approaches will
Questions over labeling of foods derived from the only prolong policy conflict.
transfer of genetic materials are being asked by
regulators, the food industry and by consumer The framework is then brought to bear on
advocates. The so-called “Delaney paradox,” has the question, “What ethical considerations should
raised questions about the ethics of limiting risk be brought to bear on labeling policies for food
from additives, while risk associated with whole products?” This question does not equate labels
foods is unregulated. At the same time, lingering with wjirnings. There are many ways to configure
questions about risk from pesticide residues and a labeling policy that do not imply health claims.
microbial contamination frame a continuing debate The short answer to the question is that there are
over food safety, one which frequently returns to two kinds of ethical considerations. The first has
labeling as a strategy for addressing consumer to do with the use of labels as tools to produce
concerns. ethically desirable ends such as good health and
consumer satisfaction, The second has to do with
The balance of this paper outlines a frame- the role of labels in protecting the principles of
work for policy analysis, and demonstrates how consent. While these two ways of evaluating
ethics bears upon each element of the framework. labeling policies may converge, they may also
Contested issues in food biotechnology policy are indicate contradictory directions. The long
used to illustrate the applicability of the frame- answer to the question uses a general approach to
work for interpreting policy conflict. Although ethics and policy to show why this is the case.
this approach addresses several of the key points

February 93/page 12 Journal of Food Distribution Research

A Framework for Policy Analysis Conduct includes the production, processing
distribution, and consumption of food.
Schmid (1987) presents a theoretical frame-
work for policy analysis where the laws, proce- 4. Pe@ormance: a given pattern of conduct will
dures and administrative decisions that seem to be produce an end ‘state ‘which consists of the
instruments of policy are analyzed in terms of the policy’s consequences for affected parties.
incentives they create for key actors. Schmid’s Health, disease, injury, profit and loss all
framework develops a public choice/transaction qualify as components of this end state.
cost approach to public policy that permits an (Schmid, Shaffer and Van Ravenswaay, 1983)
analysis of how informal norms and standard
operating procedures interact with the formal The framework provides general categories that
apparatus of law and administrative decision mak- SI1OWa competent analyst to bring implicit fea-
ing. Conventional economic policy analysis tures of policy out more clearly, and to examine
assumes that a policy’s costs and benefits can be how policies produce end states. It is admittedly
computed simply by examining the impact of laws quite general and is undoubtedly commensurate
and administrative decisions upon production costs with many different methods of policy analysis.
and consumer demand. Here the key insight of
Schmid’s approach is that the formal policy appa- It is worth noting a few additional points
ratus is one component in an ensemble of laws, before examining how ethics bears upon the
norms and standard operating procedures. The framework. First, the framework is interpretive
totality of this ensemble imposes a structure upon in that it will require judgment to assign specific
an existing reality, and the combination of struc- variables to any of the four elements. For exam-
ture and the situation as determined by physical ple, the technology that is used to detect the pres-
and biological facts determines individual incen- ence of a substance in foods uses physical and
tives and opportunities. According to Schmid, chemical principles. Technology is, in that sense,
economists have naively assumed that individuals’ a part of the reality or situation on which a struc-
behavior is shaped merely by preference rankings ture is imposed. However, this technology has
of exchangeable goods and have failed to examine changed so dramatically in the past four decades
how shared norms and public policy shape oppor- that it is probably more useful to think of it as a
tunities for choice. component of structure. The interpretation of
administrative guidelines for food safety decisions
For purposes here, Schmid’s framework includes standard operating procedures for the use
will be telescoped into four key elements. They of specific technological tests. As such, when
are defined here with explicit attention to the technology changes, there is a sense in which
analysis of food safety and nutrition policy. policy changes, too.

1. Situation: the things that cannot be changed. Second, the general category of perfor-
This should be understood to include the physi- mance can be taken to include the full range of
cal, chemical or biological processors that criteria that would be applied in evaluating a
determine food production and consumption. policy. As will become clear shortly, some such
criteria have little to do with the end state pro-
2. Structure: the ensemble of laws, shared duced by the policy. The dominant practice in
norms, procedure+ and rules that are either public policy analysis is to predict policy out-
proposed or in place in the status quo. In comes, and to report them as an end state, often
addition to the obvious elements of policy, as costs and benefits. This practice leaves the
structure includes norms that govern what decision to the responsible party or parties, be
people regard as food. they an administrator, a court or the Congress.
Decision makers can and do apply criteria that
3. Chduct: the behavior that will be produced as make little if any use of predicted end states, but
a result of the opportunities created when a the typical practice among analysts is to equate the
given structure is imposed upon the situation. predicted end state with the policy’s performance.

Journal of Food Distribution Research February 93/page 13

Analysts writing on the banning of Alar, for that rationality consisted in actions chosen as
example, typically evaluate the policy in terms of means to an end. While one might disagree about
a trade-off between the economic value of the ends, Bentham thought that a rational person must
apple crop for producers and some minimal, even accept that acts which fail to achieve the desired
tentative, reduction in risk for consumers. This end are to be rejected. The point here is that
approach leaves open the possibility of comparing structure becomes a means toward an end. While
these outcomes using a variety of criteria, but citizens in a democracy can be expected to have
presumes that the decision is based upon projected different preferences, they must evaluate the rules
policy consequences to the extent that it is defen- and regulations adopted by policy makers only as
sible at all (Roberts and Van Ravenswaay, 1989), means to some end. Hence end states are the
dominant performance criteria for public policy.
How Ethics Bears on Policy
Nozick’s Anarchy State and Utopi&(1974)
The assumption that consequences (or end is an extended philosophical attack upon this
states) provide the basis for evaluating public notion of political rationality. He offered a now
policy has its philosophical basis in the ethical famous analysis of why the basketball player Wilt
writings of Jeremy Bentham and John Stuart Mill. Chamberlain is entitled to great wealth, despite
These utilitarian philosophers argued that action what Nozick thought to be the lack of any propor-
can be justified only in light of the consequences, tionate social value produced by his play. The
and they proposed the twin norm of counting argument stressed that, given any initial distribu-
consequences for all affected parties and of maxi- tion of wealth, policies that confiscate money
mizing aggregate utility. Traditionally utilitarian voluntarily exchanged between Chamberlain and
philosophy has been criticized for its insensitivity paying fans must necessarily violate individual
to the distribution of costs and benefits. In more liberties. While a performance focused analyst
recent times, John Rawls (1971) has argued that might argue that the market structure permitting
policies should benefit the worst-off groups rather such exchanges is efficient (in that it maximizes
than maximize aggregate utility. As Nozick utility or produces a Pareto better outcome), such
(1974) noted, both utilitarian and Rawlsian egali- considerations are irrelevant for Nozick’s argu-
tarian theories evaluate policy by applying a norm ment. The point was that policies are justified
or decision rule to the end state that the structure only when they conform to an antecedently deter-
is expected to produce. Many economists who do mined set of moral or political rights. The conse-
not think of themselves as either utilitarian or quences produced by structures conforming (or
egalitarian would also assume that end states failing to conform) to this set are irrelevant.
provide the side basis for evaluating policy. The
search for Pareto better solutions or et%cient Just as it is possible to differ over the per-
levels of pollution begins by predicting policy formance criteria used to evaluate policy, it is
outcomes. The debate is over whether the ac- possible to disagree about which rights belong in
counting is accurate and complete or which princi- the template used to evaluate policy, For Nozick
ples to apply in evaluating end states. and other libertarians, the template will be nar-
rowly confined to those that protect individuals
To use the language developed here, these from interference by others. For liberals such as
are all pe~ormance or end state focused Ronald Dworkin (1977) or Henry Shue (1980),
approaches to ethics. Their philosophical pedigree the list of rights may be much more expansive.
extends back to Bentham, who hoped to reform an The point here is that these approaches to policy
English legal system based upon status and privi- are structurefocused. They establish a burden of
lege. By turning the debate toward consequences, proof met only by 1) demonstrating that policy
Bentham established a burden of proof for which conforms to the antecedently chosen structural
social rank and divine right were irrelevant. template; or 2) refuting the claim that a given
While common law based policy on past practice, right belongs in the template. Arguments that
Bentham’s theory held that it should be based stress trade-offs, efllciency or other features of
entirely on expected outcomes. Bentham assumed end states do not meet either test.

Febrwy 931page 14 Journal of Food Distribution Research

 FRAMEWORK ETHICS

Situation

Structure 4 ➤ Rights Theory

Conduct ● ➤ Virtue Theory

Performance 4 ➤ Utilitarian Theory

Figurel: How Ethica Beara on Policy

The tension between end state and structure basis for rejecting the policy. For example, a
focused policy evaluation has a ready analogy in pesticide policy which encourages producers to
the deba~e ov~r food labels. whethe~ requir~ by misrepresent their use of chemicals would be
law or custom, labels are clearly a component of judged unethical, even if no rights are violated, or
structure for food policy. The debateis: how if no harm is done.
should labels be evaluated? Given a performance
focus, labels will be evaluated in terms of the end Many authors have taken up conduct-
state they produce for the producers and consum- focused ethics in the past two decades. Bernard
ers of food products. Labels will be seen as Williams (Smart and Williams, 1972) criticized
educational tools. One will want to know whether utilitarian arguments because they fail to address
the label allows the consumer to make food pur- the character of the moral agent. Alisdair
chases that more fully satisfy preferences. Policy MacIntyre (1981) has criticized both rights theory
will be seen as a trade-off between producer costs, and utilitarian arguments for the emphasis that
consumer preferences, and health; and labels can they place on an individual’s self-regarding wants.
affect each of these outcomes in a variety of ways. He argues that a better approach would take up
A structure focused evaluation will see the matter virtues and vices that are the reference points for
almost entirely in terms of informed consent. moral character. These authors do not discuss
Labels will be seen as transforming the conditions policy, however. The public choice approach to
of consumer choice from those of mild coercion to policy analysis makes it possible to see how these
implied consent. Policies that protect individual philosophical ideas bear on policy by making it
consent are acceptable; those which foreclose clear how situation and structure produce conduct.
individual consent must bear a very heavy burden There is little doubt that most citizens would
of proof before being judged acceptable. Labels regard the conduct-focus as the most obviously
will be preferred even if people choose to ignore “ethical,” despite the relative lack of attention it
them, or even if consumers have false beliefs that has received by policy analysts. Public outrage
lead them to make less than optimal choices. over Congressional check bouncing, for example,
is almost certainly focused upon conduct, rather
Utilitarian or consequentialist ethical argu- than structure or performance.
menta, then, express norms or policy criteria that
focus on performance, while human rights argu- With respect to food safety policy, the most
ments identify characteristics of structure that likely relevance of conduct-focused evaluation is
must be in place without regard to consequences. not to producers and consumers, but to the ccm-
The matter does not end there, however, It is duct of policy makers themselves. Arguments
often a person’s conduct that is judged ethical or against the practice of pricing life, for example,
unethical. If a policy structure induces individuals are best analyzed as an objection to the practice of
to behave in ways that are unethical, there is a quanti~ing the value of life. Amette Baier

Journal of Food Distribution Research February 93/psge 15

(1986), Allan Gibbard (1986), and Douglas on the technology, and, short of that, for labeling
MacLean (1990) have expressed the concern with of milk produced using recombinant BST
pricing life in this way. The problem cannot be (Hanson, 1991). The ethical evaluation of label-
resolved by adjusting the amount of value ing policies for food biotechnology has therefore
assigned to lives; nor is the objection based on the already assumed practical importance (see also
suggestion that human lives should be assigned Hopkins, Goldberg and Hirsch, 1991).
infinite value. The point is that persons of good
character do not make decisions by attempting to The case of BST should serve as motivation
decide how much others are worth. A policy for thinking about ethical issues, but it is a poor
procedure which requires public servants to example of the scientific issues that will arise in
engage in such conduct is, in this view, a corrupt connection with the use of biotechnology for food
and indefensible procedure, and the policies that products. At least some of these products will
result from it are tainted. pose diftlcult questions for risk analysis, and some
may pose quantifiable risks. By contrast, there is
To sum up, ethics bears on policy in three virtually no scientific support for questioning the
ways. Traditional human rights arguments focus safety of BST milk (Kroger, 1992). Future policy
upon structure, applying a template of anteced- decisions will almost certainly be characterized by
ently determined constraints in their assessment of the kinds of uncertainty that have hindered the
policy. The language of virtue and integrity application of science to public policy in the regu-
focuses on conduct, evaluating a policy in terms lation of artificial sweeteners (Merrill and Taylor,
of the patterns of behavior it promotes. Finally, 1986) or of chemicals (Graham, Green and
the end state or performance evaluation of policy Roberts, 1988). In such cases, questions about
that has become a staple of the social sciences the extrapolation of data from animal studies, or
draws upon the utilitarian tradition in ethics. about the applicability of epidemiological data
Each approach establishes a burden of proof that caused regulatory policy to become embroiled in
cannot be met by arguments grounded in either of technical and methodological disputes. Criteria
the other two approaches. At the same time, each for scientific judgment and cross disciplinary
approach is deeply grounded in the culture and conflict over patterns of scientific inference are
habits of contemporary Americans. crucial to the policy debate. Because the proce-
dures and norms for scientific enquiry are them-
Food Biotechnology, Policy and Ethics selves matters of philosophical dispute, it is accu-
rate to describe risk policy debates as philosophi-
An analysis of food labels that stresses cal controversies (Hollander, 1991), Debates
rights and consent will establish very different over acceptable evidence extend philosophical
burdens of proof than does an analysis that evalu- controversy into the interpretation of situation, of
ates labels according to their consequences. the basic facts that must be accepted as constraints
Matters of character and conduct enter only indi- upon available policy options, These debates are
rectly into the disputed areas of policies for con- not, h~ever, ethical debates that conform to the
sumer information, but could be decisive to the pattern described above.
extent that they break a deadlock between those
focused on structure and those focused on perfor- Although the debate over acceptable evi-
mance or outcome criteria. Recombinant bovine dence will almost certainly recur in future food
somatotropin (BST) is the case that has spurred policy decisions, the lack of scientific or technical
debate. The substantive ethical issum raised by controversy over BST makes it a good example
the development and proposed release of BST for considering ethical issues precisely because
concern unintended consequences for the dairy disagreement about the probability of harm does
industry, dairy animals, and for environmental not confuse tie ethical issues. The biological
impact of dairying. However, it is public accept- facts that make up the situation for BST milk are
ability of milk produced using the new technology not themselves a source of controversy, at least
that has produced the greatest anxiety (Thompson, not among scientifically informed participants.
1992). Opponents of BST have called for a ban Nevertheless, labeling requirements for BST milk

February 931page 16 Journal of Food Distribution Research

have been proposed. A policy that certified or direct costs of labeling, costs which maybe signif-
required labels should be understood as an altern- icant when their impact upon processing is
ative to policies which regulate by removing or assessed, Even the approximate value of these
approving products tout court. Labels thus become costs and benefits is largely speculative, but the
a component of policy structure, to use the termi- point here is to see how consequence assessment
nology introduced here. It is worth noting, how- provides an ethical basis for the evaluation of
ever, that labels might become a component of policy, The policy is justified in terms of the
structure in any of several ways. One might acceptability and desirability of its consequences.
require labels that proclaim the presence of BST The historical standard has been the utilitarian
milk, or one might permit the use of labels that maxim proposed by Jeremy Bentham in 1789: act
certify its absence. In either case, the precise so as to produce the greatest good for the greater
wording of labels will be extremely important, as number of people. Although there are many cases
will the procedures for assuring the integrity of in which pure optimization rules such as the utili-
labels. The diversity of approaches to labeling tarian maxim may need to be modified (Thompson
implies that it is not one policy proposal that is and Stout, 1991), policies which do not provide
being discussed here, but a general class of poten- benefits that compensate for their direct cost of
tial policies that would be evaluated in similar implementation to government and to affected
ways. industries will always be difllcult to justify.

Whatever labeling strategy is employed, a Structure Focussed Evaluation of Labels:

policy using product labels can be expected to Structure focused evaluation centers upon protec-
stimulate certain patterns of conduct by consumers tion of rights as a precondition to ethically legiti-
and by processors and manufacturers. Some mate application of state power. Two key criteria
consumers will read labels and will use informa- are consent and fairness. The principle of govern-
tion as a basis for food purchases; others will not. ment by consent of the governed is, in many
One would presume that consumers expressing respects, the foundational norm of democratic
concern over BST in milk would use the label, government, while fairness, understood as equality
while others might not. These patterns of conduct before the law and protection of minority rights,
will lead to consequences that determine the per- constrains the excesses of democratic decision
formance of a labeling policy. Relevant conse- rnahg. Labels are an attractive component of
quences certainly include health benefits or risks structure because they make it possible to argue
to food consumers that are incurred as a result of that individual food consumers have been placed
their conduct. They also include costs to con- in a position to grant or withhold consent to food
sumers in the form of higher food prices, and in borne risks (real or alleged), A commmer who
the trouble and inconvenience required for reading chooses to purchase a labeled food item can legiti-
labels. Costs to processors and manufacturers are mately be understood to have consented to the
also a component of the policy’s performance. transaction, so long as meaningful alternatives are
Given this account of situation, structure, conduct available. A policy structure which does not allow
and performance, it is possible to examine the consumers to discriminate on criteria they have
ethics of a labeling policy for BST. judged to be important violates consent criteria.
However, labels can also raise questions about
Pe~ormanceFocussedEvaluationofhbels: fairness to the food industry. If the institutional
As noted, product labels can be expected to pro- practice is to use labels only in cases where seri-
duce certain costs and benefits for consumers and ous risks to health have been scientifically demon-
producers. In the case of BST, the scientific strated, as has been the case for tobacco, then the
consensus is that the health benefit to a person application of a label to BST milk may violate the
who would use such a label is zero, Consumers rights of industry by unfairly prejudicing consum-
who express concern over the use of BST would ers against the product.
derive some benefit from reduced anxiety if they
are able to satisfy their preference for non-BST btduct Focussed Evaluation of Labels:
milk. These benefits must be weighed against the Character and virtue are less clearly related to

Journal of Food Distribution Research February 93/page 17

labeling policy than are rights and consequences. ture focussed evaluation, consumer choice is
It is not obvious how the presence or absence of important to the extent that it satisfies criteria of
labels for BST milk would induce consumers to consent; questions of whether consumers are made
engage in unethical conduct, Some Americans better off by the choices they make are irrelevant.
may be inclined to make moral judgments of
character based upon a person’s dietary choices. Rights based approaches to social theory
Some religions require a dietary regimen for the have never assumed that governments have any
deviant, for example, and it is already common responsibility to make socially optimal policy
for vegetarians and non-vegetarians to make moral decisions. Rather, the first ethical responsibility
judgments about one another. Even so, there is of government is a negative one: not to interfere
little public consensus for such judgment. The in the personal liberties or freedoms of its citi-
more relevant conduct is that of industry. To the zens. Accordingly, structure focussed evaluation
extent that labels represent a form of disclosure of public policy stipulates a list or template of
that would be required by norms of honesty or rights, liberties, and possibly opportunities, much
truth-telling, a practice of labeling might be like the U.S. Bill of Rights. A policy is evaluated
thought to promote ethical conduct on the part of in terms of its ability to satisfy or fit this anteced-
industry. Ironically, disclosure will win far more ently determined template of rights. The list of
praise if it is voluntary. Hence, a labeling policy rights is adopted prior to entertaining any particu-
will win more praise from those who focus on lar policy option, so performance evaluation of
conduct if it facilitates, but does not require, specific policies is not a component in justifying
disclosure of relevant information. Conduct eval- the inclusion of a given right. The key right with
uation does not provide a clear mandate for or respect to food consumption is a general right to
against labels, however, and it will not be empha- non-interference in personal choices, provided that
sized in the comparison which follows. personal choices do not violate complementary
non-interference rights of others.
Comparing Ethical Approaches
for Policy Evaluation The example of labels for BST provides an
illustration of why performance focussed and
Structure focussed and performance structure focussed approaches to understanding
focuss~ approaches to policy evaluation establish how ethics bears on policy introduce distinct
different and sometimes contradictory burdens of burdens of proof. A more detailed analysis of
proof. Evidence and argument which is highly labels would need to take up additional issues.
relevant to a performance evaluation is often For example, the traditional norm of caveat
irrelevant to an evaluation in terms of rights and emptor has historically served as an informal
consent. component of structure for consumer food deci-
sions. The previous discussion of labels has
There are a number of difficult philosophi- assumed that consumers have a right to any infor-
cal and economic measurement problems that must mation4hey deem relevant about food choices, but
be addressed within a performance focussed evalu- caveat emptor might be taken to qualify this right.
ation (Giere, 1991). Two general points of dis- Further discussion of issues specific to labeling
pute concern the quality of models used to predict policies cannot be undertaken within the con-
health consequences, and the choice of a decision straints of this paper. However, readers should be
rule to compare consequences, once they have advised that what has been said to illustrate the
been assessed. The debate over linear vs. thresh- contrast between structure and performance is far
old extrapolation of data (Schaffner, 1991) is an from being a complete ethical analysis.
example of the first problem. The debate over
Delaney vs. de minimus is an example of the The contrasting burdens of proof for perfor-
second (Jasanoff, 1991). The point here is to see mance philosophies and structure philosophies
that the crucial burdens of proof established within present a general problem for biotechnology poli-
a consequence evaluation approach differ from cy and food safety. Interested parties, including
those in a rights-based procedure. Within a struc- scientists and regulators, can be so closely wedded

February 931page 18 Journal of Food Distribution Research

to one of these philosophies that they fail to predictions for regulation of foods involve risk
understand the force of their opponents’ argu- assessment and economic impact. If citizens are
ments. Someone who insists upon interpreting invited to participate in decision making after such
structure-focussed rights arguments in terms of the information has been wllected, analyzed and
end state produced by policy will simply miss the presented, the opportunity for a structure or con-
point of that argument. Rights arguments will duct focussed evaluation has been reduced, if not
appear as irrational or naive, failing to grasp the foreclosed, by the preponderance of evidence
importance of trade-offs that are thought to be the relevant only to performance criteria. Since
main focus of policy evaluation. It is not difllcult structure focussed criteria often stress the impor-
to find authors who appear to exhibit virtually tance of participation and consent, the lack of
total insensitivity to the burdens of proof entailed citizen participation in the scientific assessment of
by a focus on structure. In a 1990 article on food risk and of economic impact is doubly troubling.
safety policy for recombinant DNA derived ani- Citizens have been denied participation in the
mal growth hormones, Fred Kuchler, John early stagea of the process, and are faced with a
McClelland and Susan Offutt characterize the final decision procedure in which evidence that is
issue entirely in terms of performance or end-state potentially irrelevant to consent criteria dominatea.
criteria. The idea that rights or consent are rele-
vant is absent from their analysis, Doyle and Structure and conduct criteria should not be
Marth (1991) also offer an analysis focussed arbitrarily excluded from a decision, as they are
exclusively on end states. when scientists make risk assessments without
substantial citizen participation. But this is not to
Milton Russell (1990) has written that it is say that citizen assessment of risks should simply
irresponsible for a public of!lcial to make policy be substituted for scientific ones. Frank Cross
decisions without attempting to assess the conse- (1992) accuses me of doing just this in a recent
quences to the fullest extent possible. He qualifies paper @hompson, 1990). The point is that end
this commitment to performance evaluation by state assessments cannot meet burdens of proof
also stating that “. . . legitimacy flows from an established by criteria that stress participation and
acceptance of the decision, or at least the decision consent. Deborah G. Mayo (1991) has an
process, by those affected” (p. 22). This qualifi- extended discussion of how scientific and citizen
cation stresses the importance of consent. Con- assessments of risks might be compared. People
sent can involve the prediction of consequences, who interpret the criticism of scientific risk assess-
so long as predictions are part of an information ment as a call for citizen assessment are display-
sharing process designed to build consensus. But ing a myopic focus upon performance criteria.
traditional consent criteria stress the protection of Structure and conduct criteria establish burdens of
rights, and may leave the burden of predicting proof in which anyone’s assessment of likely end
consequences to the affected parties themselves. states is largely irrelevant. Put another way, if
Despite his qualifying comment, Russell may the goal is to implement a policy that is likely to
favor performance criteria in making the predic- minim@e food related illness, then it seems obvi-
tion of consequences a strict requirement of ethi- ous that the best scientific techniques should be
cal policy making, even while he endorses the used to predict the incidence of illness associated
principle of consent, One aspect of the tension with a given product or practice. However, if the
between end state evaluation and principles of goal is to ensure that consumers have confidence
consent is the question of who assesses conse- in the food supply, an altogether different policy
quences. may be indicated. Consumer conildence may be
very imperfectly correlated with the probability of
Biologists or economists who predict the illness. Confidence may be more closely corre-
consequences of biotechnology do not, generally, lated with participation and consent, with the
need the advice of citizens in the process of col- structure under which dietary decisions are made,
lecting data and making projections. They can rather than the end state that is produced.
produce a prediction of end states without seri-
ously involving citizens in their activity. The key

Journal of Food Distribution Research February 93/page 19

 The implication is that, while performance Bibliography
evaluation is “objective, “ in the sense that it does
not systematically favor any specific interests or Baier, Annette. 1986. “Poisoning the Wells,” in
political ideology, the practice of performing Values at Risk, Douglas E. MacLean, ed.,
extensive scientific studies can introduce a bias Totowa, NJ: Rowman and Allanheld, pp.
against ethical criteria that emphasize structure 49-74.
and conduct. Yet it is easy to find examples of
structure or conduct focused criteria in the history Cross, Frank, B. 1992. “The Risk of Reliance
of American government. The framers of the on Perceived Risk,” in Risk: Issues in
U.S. Constitution were themselves structure- Health and Safety, 3(l): 59-70.
focussed in adopting the Bill of Rlghta, and con-
duct-focused in proposing the division of powers. Doyle, Michael P. and Elmer H. Marth. 1991.
It is therefore reasonable to presume that demo- “Food Safety Issues in Biotechnology,” in
cratic decision making ought not be systematically Agricultural Biotechnology: Issues and
biased against structure and conduct criteria, C%oices,Bill R. Baumgardt and Marshall
Barring specific arguments to the contrary, policy A. Martin, cd., West Lafayette, IN: Purdue
making procedures should avoid domination by University Agricultural Experiment Station,
any one of these three philosophical approaches to pp. 55-80,
ethics, and should weigh evidence and arguments
in terms of the burden of proof to which they are Dworkin, Ronald. 1977. T&ng Rights Seri-
most clearly relevant. ously, Cambridge, MA: Harvard University
Press,
None of this is to suggest that structure-
focussed or conduct-focussed ethical criteria ought Gibbard, AlIan. 1986. “Risk and Value,” in
to determine policy choices unilaterally. The risks Values at Risk, Douglas E. MacLean, ed.,
and economic consequences are obviously relevant Totowa, NJ: Rowman and Allanheld, pp.
to policy choice. The obvious ideal is when all 94-112.
three approaches to policy evaluation converge on
the same choice, but there may be good reasons Giere, Ronald. 1991, “Knowledge, Values, and
why impacts or end-states should be the overrid- Technological Decisions: A Decision
ing considerations in making some policy determi- Theoretic Approach,” in Acceptable Evi-
nations where they conflict. It will be very diffi- dence: Science and Values in Risk iUan-
cult to understand or state those reasons, however, agement, Deborah G, Mayd and Rachelle
if one is so deeply committed to consequentialist D. Hollander, eds., Oxford: Oxford Uni-
or utilitarian thinlchg that it is impossible to versity Press, pp. 183-203.
understand how someone could see it another
way. The main contribution of ethics to better Graham, John D., Laura C. Green and Marc J.
policy making is to show how the opportunities Roberts. 1988. In Search of Safety:
for conflict and consensus rest upon alternative Ckemicals and Clmcer Risk, Cambridge,
visions of the right and the good. A policy analy- MA: Harvard University Press.
sis framework in which situation, structure, con-
duct and performance have been clearly distin- Hanson, Michael. 1991. “Consumer Concerns:
guished can contribute to that end. Give Us The Data,” in Agricultural Bio-
technology at the Crossroads: Biological,
Social and Institutional Clmcerns, Ithaca,
NY: National Agricultural Biotechnology
Council, pp. 169-178.

February 931page 20 Journal of Food Distribution Research

Hollander, Rachelle D. 1991. “ExpertClaims Mayo, Deborah G. 1991. “Sociological Versus
and Social Decisions: Science, Politics, and Metascientific Views of Risk Assessment,”
Responsibility,” in Acceptable Evidence: in AcceptableEvidence: Science and Values
Science and Values in Risk Management, in Risk Management, Deborah G. Mayo
Deborah G. Mayo and Rachelle D. and Rachelle D. Hollander, eds., Oxford:
Hollander, eds., Oxford: Oxford University Oxford University Press, pp. 249-279.
Press, pp. 160-173.
Merrill, Richard A. and Michael R. Taylor.
Hopkins, D. Douglas, Rebecca J. Goldberg and 1986. “Saccharin: A Case Study of
Steven A. Hirsch. 1991. A Mutable Feast.’ Government Regulation of Environmental
Assuring Food Safety in the Era of Genetic Carcinogens,” in Agriculture and Human
Engineering, New York: Environmental Values, 3(1 and 2): 33-73.
Defence Food).
Nozick, Robert. 1974. Anarchy State and
Jasanoff, Sheila. 1991. “Acceptable Evidence in Utopia, New York: Basic Books.
a Pluralistic Society,” in Acceptable Evi-
dence: Science and Values in Risk Man- Rawls, John. 1971. A Theory of Justice,
agement, Deborah G. Mayo and Rachelle Cambridge, MA: Belknap Press of Harvard
D. Hollander, eds., Oxford: Oxford Uni- University Press.
versity Press, pp. 29-30.
Roberts, Tanya and Eileen van Ravenswaay.
Kroger, Manfred. 1992. “Food Safety and 1989. ‘The Economics of Safeguarding
Product Quality, “ in Bovine Somatotropin the U.S. Food Supply,” Economic
and Emerging Issues: An Assessment, Research Service, Agriculture Information
Milton, C. Hallberg ed., Boulder, CO: Bulletin Number 566, U.S. Department of
Westview Press, pp. 265-270. Agriculture.

Kuchler, Fred, John McClelland, and Susan E. Russell, Milton. 1990. “The Making of Cruel
Offutt. 1990. “Regulatory Experience Choices, “ in Valuing Health Risks, Cbsts,
with Food Safety: Social Choice Implica- and Benejits for Environmental Decision
tions for Recombinant DNA-Derived Making: Report of a Clmference,
Animal Growth Hormones,” in Biotechnol- Washington, DC: National Academy Press,
ogy: Assessing Social Impacts and Policy pp. 15-22.
Implications, David J. Webber ed.,
Westport, CT, Greenwood Press, pp. 131- Schaffner, Kenneth F. 1991. “Causing Harm:
144. Epidemiological and Physiological Concepts
of Causation, ” in Acceptable Evidence:
MacIntyre, Alisdair. 1981. Afier Virtue:A Study Science and Values in Risk Management,
in Moral Theory, Notre Dame: University Deborah G. Mayo and Rachelle D.
of Notre Dame Press. Hollander, eds., Oxford: Oxford University
Press, pp. 204-217.
MacLean, Douglas E. 1990. “Comparing Values
in Environmental Policies: Moral Issues and Schmid, Allen. Property, Power, and
1987.
Moral Arguments. ” ValuingHealth Risks, Public Choice: An Inquiry into J2ZWand
Cbsts, and Benejits for Environmental Economics, 2nd Edition, New York:
Decision Making: Report of a Conference, Praeger Publishers.
Washington, DC: National Academy press,
pp. 83-106.

Journal of Food Distribution Research February 93/page 21

Schmid, Allen, James D. Shaffer and Eileen O. Thompson, Paul B. 1990. Tlisk Subjectivism
van Ravenswaay. 1983. “Community and Risk Objectivism: When Are Risks
Economics: Predicting Policy Red’?” Risk: Issues in Health and Safety,
Consequences,” Department of Agricultural Concord, NH: Franklin Pierce Law Center,
Economics, Michigan State University, East 1(1):3-19.
Lansing, MI.
Thompson, Paul B. 1992. “Ethical Issues in
Shue, Henry. 1980. Basic Rights: Subsistence, BST,” in Bovine Somatotropin and
Afluence, and U.S. Foreign Policy, Emerging Issues: An Assessment, Michael
Princeton: Princeton University Press. C. Hallberg ed., Boulder, CO: Westview
Press, pp. 33-50.
Smart, J. J. C. and Bernard Williams. 1973.
UtilitarianismFor and Against, New York: Thompson, Paul B. and Bill A. Stout. 1991.
Cambridge University Press. Beyond the Large Farm: Ethics and
Research Goalsfor Agriculture, Boulder,
CO: Westview Press.

February 931page 22 Journal of Food Distribution Research