The precautionary principle is frequently referred to in various
momentous decisions affecting human health and the environment. It has
been invoked in contexts as diverse as chemicals regulation,1
Commission of the European Communities (CEC).
Communication from the Commission on the precautionary principle.
Brussels, Feb 2, 2000. COM 2000:1.
regulation of genetically
modified organisms,
2 Löfstedt RE,
Fischhoff B, Fischhoff IR. Precautionary principles: General definitions
and specific applications to genetically modified organisms. Journal
of Policy Analysis and Management 2002;21(3):381–407.
and research into life-extending therapies.
3 Glannon W. Extending the human life span. Journal of
Medicine and Philosophy 2002;27(3):339–54.
Precaution
is not an unknown concept in medical contexts. One author even cites the
Hippocratic Oath as a parallel to the precautionary principle.
4 Ozonoff D. The precautionary principle as a
screening device. In: Raffensperger C, Tickner J, eds. Protecting
Public Health and the Environment: Implementing the Precautionary
Principle. Washington, D.C.: Island Press;
1999:100–5.
In a recent paper commenting on Walter Glannon's view, John
Harris and Søren Holm argued that the precautionary principle is
incoherent and leads to paradox.5
Harris J,
Holm S. Extending human lifespan and the precautionary paradox.
Journal of Medicine and Philosophy
2002;27(3):355–68.
Their arguments are not aimed merely at
Glannon's standpoint, but are quite general. In this paper, I argue
that their arguments are directed against interpretations of the
precautionary principle that are taken out of context, and that their
arguments fail when the precautionary principle (or one or another version
of it) is understood in context. The next section introduces the
precautionary principle and presents two arguments that have been raised
against it. Against the background of this, Harris and Holm's
critique is discussed in the third section. In the fourth section, the
importance of their critique for proponents of the precautionary principle
is acknowledged and commented on. Conclusions are offered in the last
section.
The Precautionary Principle
Harris and Holm are well aware that there is no universally accepted
canonical formulation of the precautionary principle. As has been noted,
it may be misleading to talk about “the precautionary
principle” with the definite article,6
Graham JD. Perspectives on the precautionary principle.
Human and Ecological Risk Assessment
2000;6(3):383–5.
as the phrase has several distinct uses. I
will try to clarify the matter somewhat.
First, the phrase the precautionary principle may refer to
one or another principle of national or international law. Various
precautionary principles, to use the plural, have been included in several
international legal documents. Second, the phrase the precautionary
principle is used more broadly, referring to some principle that can
or should be applied by decisionmakers and policymakers in general.7
Sandin P. Dimensions of the precautionary
principle. Human and Ecological Risk Assessment
1999;5(5):889–907.
This is relevant also in medicine and
healthcare. The distinction is important, for the following reason.
Consider a formulation of a rule or a principle, such as the
prohibition of killing: “Killing is prohibited.” Taken in
isolation, it can be easily shown that such a principle leads to
counterintuitive or at least problematic consequences. Is killing in
self-defense always prohibited? Is it in every case prohibited to kill a
dictator in order to save thousands of innocent lives? Or, to take an
example from medicine, is it always wrong to kill someone who is suffering
from terminal and painful illness, and who requests to be killed, a brand
of killing usually termed euthanasia? (This is not to say that there are
no proponents of an absolute prohibition on killing. There are. Neither
would I argue that such a position is necessarily wrong. What I would
argue, however, is that such an absolute prohibition to me, and many
others, carries with it prima facie counterintuitive and problematic
consequences, and hence cannot be accepted without convincing arguments.
This is sufficient for the point I make here.)8
Glover J. Causing Death and Saving Lives.
Harmondsworth: Penguin; 1990.
Of course such rules or principles very rarely occur in isolation.
They occur in context, surrounded by a number of auxiliary principles,
presuppositions, and limitations, implicit or explicit. This context might
be religious, as in the case of “Thou shalt not kill,” it
might also be legal, as in the case of a paragraph prohibiting murder, or
secularly philosophical.
The above holds also for the precautionary principle. If we by
“the precautionary principle” mean a specific principle of
national or international law, we have to consider it in that context.
This is highly relevant for some of the arguments against the
precautionary principle that have been discussed.9
Sandin P, Peterson M, Hansson SO, Rudén C, Juthe A.
Five charges against the precautionary principle. Journal of Risk
Research 2002;5(4):287–99.
One such argument is that
the precautionary principle is ill defined. That might be true, but
interpretations of the precautionary principle are in fact emerging,
albeit slowly. This is a task for policymakers, legislators, and courts.
For instance, the recent EC communication,
10 the broad lines of which were
endorsed by the European Council's meeting in Nice, December
2000,
11 European Council. Presidency
Conclusions. Nice European Council Meeting 2000 Dec 7–9. SN
400/00.
is a modest step in this direction. The Commission
notes that “it would be wrong to conclude that the absence of a
definition has to lead to legal uncertainty” (p. 10), and that
“[t]he Community authorities' practical experience
with the precautionary principle and its judicial review make it possible
to get an ever-better handle on the precautionary principle” (p.
10).
Arguments from Absolutism and Risk Trade-Off
Similar remarks apply to another objection to the precautionary
principle that has been termed the argument from absolutism. The
argument from absolutism says that the precautionary principle, at least
some versions of it taken literally, will prohibit every action,
and thus offer no action guidance whatsoever. Consider the following
hypothetical and obviously absolutist version of the precautionary
principle:
If an action might lead to severe damage, then the action
must be avoided.
Of course, every activity is associated with some risk of severe
damage. My writing this paper might, for instance, through an extremely
complex causal chain, result in the end of the world. Fortunately, it is
not at all probable, but it is not impossible. In this way, an absolutist
interpretation of the precautionary principle would prohibit, in
principle, every action. Because any action might have unforeseen
catastrophic consequences, the action of carrying it out will be
prohibited, and so will the action of not carrying it out. Thus,
the precautionary principle is not coherent. The argument from absolutism
is not at all uncommon in the debate around the precautionary
principle.12
McKinney WJ. Prediction and
Rolston's environmental ethics: Lessons from the philosophy of
science. Science and Engineering Ethics 1996;2(4):429–40;
Manson NA. Formulating the precautionary principle. Environmental
Ethics 2002;24:263–74.
A related but less radical objection to the precautionary principle is
the argument from risk trade-off. The phenomenon of risk
trade-off is the following: If I do something to reduce a risk, as when I
take prophylactic medication to reduce the risk of catching malaria, my
risk-reducing action might have the consequence of increasing another
risk, such as the risk of nausea or worse side effects of the
medication.13
Graham JD, Wiener JB. Risk
versus Risk: Tradeoffs in Protecting Health and the Environment.
Cambridge, Mass.: Harvard University Press; 1995.
The relation between the argument from absolutism and the argument
from risk trade-off is the following: The argument from absolutism says
that the precautionary principle is incoherent, for logical reasons, as it
would prescribe that an action should be carried out and that the very
same action should not be carried out. The argument from risk trade-off
says that concrete applications of risk reduction measures—not only
the precautionary principle—will (or at least may) lead to worse
consequences or larger risks than would abstaining from these measures.
This may come about directly when risk reduction measures themselves
impose new risks or indirectly when risk reduction measures are so costly
that the resultant loss of wealth imposes risks.14
Wildavsky A. Richer is Safer. Public Interest
1980;60:23–39; Cross FB. When environmental regulations kill: The
role of health/health analysis. Ecology Law Quarterly
1995;22(4):729–84; Cross FB. Paradoxical perils of the precautionary
principle. Washington and Lee Law Review
1996;53:851–925.
According to the argument from risk
trade-off, principles prescribing such measures are not necessarily
incoherent, but merely irrational or counterproductive.
Types of Precautionary Principles
Despite the plethora of formulations of the precautionary principle,
there are nevertheless at least two distinct types.15
See note 9, Sandin et al. 2002; Morris J. Defining the
precautionary principle. In: Morris J, ed. Rethinking Risk and the
Precautionary Principle. Oxford: Butterworth-Heinemann;
2000:1–21; Soule E. Assessing the precautionary principle.
Public Affairs Quarterly 2000;14(4):309–28.
The
first type contains the
argumentative versions of the
precautionary principle. These versions are not action prescribing, but
principles for what reasons or arguments are valid. The most prominent
example is the version of the precautionary principle found in Principle
15 of the Rio Declaration.
16 UNCED. The
Earth Summit: The United Nations Conference on Environment and
Development Rio De Janeiro, 1992. Introduction and commentary by
Stanley P. Johnson. London: Graham & Trotman; 1993.
This
version, which is included almost verbatim also in several other
documents, requires that “lack of full scientific certainty shall
not be used as a
reason for postponing cost-effective measures to
prevent environmental degradation” (my italics). Argumentative
versions of the precautionary principle do not seem very demanding. They
say little more than that arguments from ignorance should not be used.
Argumentative versions, are thus, if not unproblematic, at least less
problematic than the second type of version of the precautionary
principle:
prescriptive versions. Prescriptive, as opposed to
argumentative, versions prescribe actions. They say that, under certain
conditions, precautionary measures are prescribed (in a sense to be
discussed later). A typical prescriptive version of the precautionary
principle is the formulation of the Wingspread Statement of 1998:
When an activity raises threats to the environment or human
health, precautionary measures should be taken, even if some
cause-and-effect relationships are not fully established
scientifically.17
Wingspread Statement on
The Precautionary Principle. In: Raffensperger C, Tickner J, eds.
Protecting Public Health and the Environment: Implementing the
Precautionary Principle. Washington, D.C.: Island Press;
1999:353–5.
Prescriptive versions of the precautionary principle share some common
elements and a common structure.18
They can be recast into the following
if-clause:
If there is (1) a threat, which is (2) uncertain,
then (3) some kind of action (4) is mandatory.
Here, (1) and (2) together indicate when precautionary measures are
applicable, (3) states what kind of action is prescribed, and (4) states
the way in which the action is prescribed. We may call (1) the threat
dimension, (2) the uncertainty dimension, (3) the action dimension, and
(4) the prescription dimension. Of course, the phrases in the dimensions
vary greatly between different formulations. For example, in the threat
dimension, we find phrases like “activities raising threats to the
environment or human health,” and in the action dimension we find
phrases like “regulatory measures” or “action to protect
the environment.”19
Obviously, prescriptive versions of the precautionary
principle vary greatly in stringency depending on the content of these
dimensions. However, the elements (1)–(4) and the overall structure
is common to most existing prescriptive versions of the precautionary
principle.
Harris and Holm's Objections
Let us now turn to Harris and Holm's critique of the
precautionary principle. Their “initial approximation to the core
content of [the precautionary principle]”20
See note 5, Harris, Holm 2002:358.
fits the structure just described, as does their modified version
(PP1).
21 See note 5, Harris, Holm
2002:359.
They did not discuss the distinction between
argumentative and prescriptive versions of the precautionary principle.
Instead, they distinguished, first, between the precautionary principle as
an epistemic rule (E-PP) and as a rule of choice (C-PP). Second, they
distinguished between C-PP as a principle of rational choice and as a
moral principle.
Epistemic Precaution
By the E-PP, they meant a principle that requires that evidence
suggesting a causal link between an activity and possible harm should be
given greater weight than it would in other circumstances. They rejected
such principles, for the reason that “systematic discounting of
evidence would systematically distort our beliefs about the world, and
would necessarily, over time, lead us to include a large number of false
beliefs in our belief system.”22
See
note 5, Harris, Holm 2002:362.
That may well be true, even if it
is not obvious that this makes the principle invalid, as Harris and Holm
claimed. It is not only the number of correct beliefs in a belief system
that counts. Suppose that we have exactly three possible epistemic
principles, P1, P2, and P3. P1 leads us to include a large number of false
beliefs about mostly irrelevant matters (the exact number of grains of
sand in Iraq, what Caesar had for dinner three days before his death,
etc.) in our belief system, but it also leads us to the correct answer to
the question of what substance would cure AIDS. P2 does not lead us to
include any false beliefs, but true beliefs about irrelevant matters
(Caesar's dinner, etc.). P3 does not lead us to include any beliefs
at all. In this case, the reasons for rejecting P1 are indeed weak. Harris
and Holm's argument apply at best
ceteris paribus. However,
there is a more down-to-earth problem with this argument. It is that the
E-PP, as Harris and Holm presented it, is something of a straw man. The
argument is not relevant to existing versions of the precautionary
principle. The following extremely simplified example may illustrate the
point: Suppose that we have scientific evidence concerning the toxicity of
a chemical substance
S. Moderately credible evidence is pointing
in the direction that
S is toxic, and equally credible evidence
is pointing in the opposite direction, that it is nontoxic. The E-PP, as
Harris and Holm interpreted it, would prescribe “Believe that
S is toxic” or perhaps “Believe that the proposition
‘
S is toxic’ is
x% more probable than the
proposition ‘
S is nontoxic’” (add
technicalities to taste). This is the only reasonable way of making sense
of the argument that E-PP will lead us to include a large number of false
beliefs in our belief system. However, this is not what existing versions
of the precautionary principle require us to do. It is a different sense
of “giving weight to evidence” that is at play here. It is not
about which proposition is to be believed, but which is to be acted upon.
Harris and Holm admitted that it may be rational to have epistemic rules
that “instruct us always to resolve ties in justification in a
specific way.”
23 See note 5, Harris,
Holm 2002:361f.
As an example, they mentioned court cases. But in
court cases, epistemic rules are to be acted upon. A typical example is
the rule that a suspect is presumed innocent until it can be shown beyond
reasonable doubt that he is guilty. Those who have argued in favor of
evidence-weighing or burden-of-proof versions of the precautionary
principle rather have this in mind. In fact, I have not encountered any
formulation of the precautionary principle that is like Harris and
Holm's E-PP, either in the original or the weakened version. I will
therefore not dwell further on the refutation of the E-PP and instead turn
to the C-PP, that is, the precautionary principle as a rule of choice.
The Precautionary Principle as a Rule of Choice
As we saw, the C-PP may be interpreted both as a principle of
rationality and as a principle of morals. Both as principles of rational
choice and as moral principles, the versions of the precautionary
principle they discussed are prescriptive and fit well into the structure
described above. Principles of rationality and moral principles can both
be prescriptive. (The distinction between rationality and morals is
traditional and not unquestionable, but I will not pursue the matter
here.)
The difference in the case of the precautionary principle is to be
found in the prescription dimension. If we take the precautionary
principle to be a moral principle, we would insert a phrase like “is
morally right” or “is morally permissible” in the
prescription dimension. If we take it to be a principle of rational
choice, we would use a phrase like “is rational.” Of course, a
neutral phrase may also be used, such as “should be taken,”
where it is implicit whether the “should” is a moral term or a
rationality term.
Harris and Holm rejected the precautionary principle as a rule of
choice. Their argument is in essence identical to the argument from
absolutism. Using the example of GM plants, Harris and Holm talked of an
“infinite regress of precaution,”24
See note 5, Harris, Holm 2002:362.
and, commenting
on Glannon's paper on research into life-extending therapies,
25 See note 3, Glannon 2002.
they termed
this “the paradox of precaution.” Another way in which they
put it is that the precautionary principle “requires us to use an
infinite safety factor.” They are right in their claim that this
would be irrational, but wrong in that the precautionary principle
requires us to do so. The problem with their argument is that they took
the precautionary principle out of context.
Prescriptive versions of the precautionary principle prescribe that
precautionary measures should be taken. What, then, is a precautionary
measure? In the everyday sense of the term, a precautionary measure is
roughly an action that is intended to prevent an undesirable outcome that
is possible but not certain to occur. There is more to it than that, but
this definition is sufficient for the purposes of the present paper.26
Sandin P. The precautionary principle and the
concept of precaution. Environmental Values
2004;13:461–75.
What must be noted is that precautionary
measures are precautionary
with respect to something undesirable.
Suppose that you are a doctor performing an operation to fixate a femur
fracture. There is a slight possibility that there will be an infection in
the wound, and as a precautionary measure, the patient is given
prophylactic antibiotics treatment. In this case, the antibiotics
treatment is precautionary with respect to a certain undesirable outcome
(the infection). It is not precautionary with respect to other undesirable
outcomes. For instance, it is not precautionary with respect to evolution
of bacteria that are resistant to antibiotics.
The argument from absolutism obviously bites if the precautionary
principle is thought of as prescribing measures that are precautionary
with respect to everything. But this is not the case. If the
precautionary principle is interpreted as a principle of national or
international law, it is implicitly or explicitly stated to what threats
it is supposed to apply. In the case of the Rio Declaration cited
above,27
we are clearly in the context of environmental degradation. In that
context, precaution is not necessarily thought to apply to every other
possible threat as well.
Furthermore, many prescriptive versions of the precautionary principle
require that actions be taken when there is lack of full scientific
certainty. (In fact, uncertainty is a prerequisite for precaution. We do
not take precautions against things we expect to happen.) But this does
not mean that precautionary measures are required by the precautionary
principle when there is no particular evidence, scientific or other, of
the presence of a threat. This is recognized in increasing degree in
several documents. For instance, in some documents it is explicitly
required that the possibility of harm should at least be identified. One
example is found in the conclusions of the European Council Meeting in
Nice in December 2000.28
See note 11,
European Council, Annex III, §7.
I do not believe that any
proponent of the precautionary principle would actually demand precaution
with respect to everything, and nothing in texts I have encountered so far
indicates that they would.
Well, it might be objected: It is perhaps the case that proponents of
the precautionary principle did not intend to be precautionary with
respect to everything. But this is what follows from their principle,
whether they intended it or not, and it leads to unreasonable
consequences, for instance in the form of “the precautionary
paradox.” They do not seem to have realized this.
To this I reply: Well yes, it follows—but only if the
precautionary principle is taken out of context. To do so is unreasonable.
Indeed, the argument from absolutism points at such a glaring problem that
it is highly improbable that anyone would propose a principle that is
vulnerable to it. Instead of concluding that the proponents of the
precautionary principle have not realized the consequences of their
position, one may suspect that this objection, like Harris and Holm's
earlier rejection of the E-PP discussed above, is directed against
something of a straw man.
Importance of the Objections
Even if Harris and Holm's objections might be directed at
something of a straw man, they are not without importance for proponents
of the precautionary principle, whatever version of it is intended, and in
whatever area it is to be applied. The objections are important because
they point at areas in which clarification is needed. Also in this case,
context is of vital importance. In particular, certain implicit
assumptions should be made explicit. For instance, as we saw that
precautionary measures cannot be precautionary with respect to
everything—this is the lesson learned from the argument from
absolutism—it should be explicitly and precisely stated with respect
to what a proposed precautionary measure is meant to be precautionary.29
See note 26, Sandin 2004.
This is not
least important in risk communication. There is a reason why Harris and
Holm, and quite a few other critics, have interpreted the precautionary
principle in the context-free way described above. (Harris and Holm are
not the only authors who have argued that the precautionary principle will
lead to paralysis in decisionmaking, and their claim that “the
fundamental threat that the increasing popularity of the
[precautionary principle] poses for scientific advance and
technological progress” has been “almost unnoticed”
30 See note 5, Harris, Holm 2002:356f.
is exaggerated.) The reason is that appeal to the precautionary principle
has not always been made in the precise and structured way it should be.
This is also indicated by the fact that the principle has been criticized
for being vague and ill defined.
31 Bodansky
D. Scientific uncertainty and the precautionary principle.
Environment 1991;33(7):4–5, 43–4; Starr C. The
precautionary principle versus risk analysis. Risk Analysis
2003;23(1):1–3.
Harris and Holm's arguments thus
pinpoint areas where increased specification is needed.
Thus far, I have argued that the precautionary principle should be
specified in the threat dimension. It should also be specified in at least
two other respects.
The first is already hinted at and concerns the uncertainty dimension.
In the uncertainty dimension, it is stated at what level of uncertainty
precautionary measures are applicable. Here is room for further
specification as to what level of evidence is required to trigger
precaution. This is, as we have seen, a way around the argument from
absolutism. For instance, future formulations of the precautionary
principle could be made to include more precise statements as to the
nature of the suspicions required to trigger precaution, something which
has hitherto been lacking. It might be objected that a phrase requiring,
for example, “reasonable scientific foundation” of a suspicion
to trigger precautionary measures is not very precise. But it is not
obviously less precise than many other phrases that legislators and
policymakers leave to courts and other authorities to interpret. (It is
not even clear that it is more difficult to interpret than phrases such as
“scientific proof” or “scientific certainty.”)
The second respect does not concern any of the dimensions
(1)–(4) mentioned above. Instead, it concerns time. Precaution is
warranted only when information about the possible threat is lacking. When
something undesirable is highly probable to occur, measures to prevent it
are not usually said to be precautionary. We do not sterilize surgical
instruments as a precautionary measure, as we know that an infection is
rather probable if we use unclean instruments. But information changes
over time. Let me take an example that has been cited as an instance of an
early use of the precautionary principle.32
Harremoes P, Gee D, MacGarvin M, Stirling A, Keys I, Wynne B,
Guedes Vaz S. Introduction. In: Harremoes et al., eds., The
Precautionary Principle in the 20th Century: Late Lessons from Early
Warnings. London: Earthscan; 2002:5–8.
I do not think
it is correct to call it an application of the precautionary principle,
for the reason that no principle seems to have been referred to. However
it is an obvious instance of precaution. In an 1854 London cholera
epidemic, Dr. John Snow recommended that the handle of the Broad Street
water pump be removed in order to stop the spreading of cholera. At the
time, the mode of communication of cholera through contaminated water was
not evident, though Snow himself seems to have had such suspicions.
33 Snow J. The cholera near Golden-Square, and at
Deptford. Medical Times and Gazette 1854;9:321–22.
Snow's recommendation was thus an example of precaution. Today,
however, the closing (or perhaps chlorinating) of a cholera-infested
source of drinking water would not be merely precautionary. We know that
things are likely to turn out badly otherwise. The opposite situation is,
of course, also possible, where a precautionary measure is discovered to
be unnecessary as new information becomes available. One example might be
the downgrading of diethylhexylphthalate, DEHP, from “possibly
carcinogenic to humans” to “not classifiable as to its
carcinogenicity to humans” by one of the working groups of the
International Agency for Research on Cancer.
34 These examples both illustrate the relevance of the time factor for
the precautionary principle. Precautionary measures cannot be taken once
and for all, but must be subject to review as new information accrues.
Consequently, the precautionary principle needs to be specified as to how
often the measures should be reviewed. This need is acknowledged in, for
instance, the communication on the precautionary principle from the
Commission of the European Communities.35
Conclusion
In this paper, I have argued that Harris and Holm's rejection of
the precautionary principle is unwarranted. Their rejection of what they
termed the “epistemic” version of the precautionary principle
(E-PP) is based on a formulation of the principle that is at odds with
most existing formulations. Their rejection of the precautionary principle
as a rule of choice (C-PP) consisted primarily of delivering an argument
that has not been uncommon in the debate on the precautionary principle:
the argument from absolutism. In both cases, their rejection was based on
a narrow conception of the precautionary principle that ignores context.
The objections delivered by Harris and Holm, however, pinpoint important
areas—not all such areas, of course—in which clarification is
needed. Thus, their objections might in the end be quite useful for
proponents of the precautionary principle. We should thank them for
that.36
Cf. Häyry M. Precaution and
solidarity. Cambridge Quarterly of Healthcare Ethics
2005;14(2):199–206.
