In “A Paradox out of Context: Harris and Holm on the
Precautionary Principle,” Sandin criticizes the earlier paper
“Extending Human Lifespan and the Precautionary Paradox”
wherein John Harris and I argued that the precautionary principle (PP) is
incoherent.1
Harris J, Holm S. Extending
human lifespan and the precautionary paradox. Journal of Medicine and
Philosophy 2002;27(3):355–68.
These criticisms offer
me the possibility to briefly expand and clarify some of our previous
arguments, and to show that the paradox of precaution is not dispelled by
attention to context as Sandin maintains. Even when context is fully
acknowledged, application of the PP will still lead to paradox and
paralysis.
The Criticisms Summarized
Sandin raises a number of criticisms of our position, but I take the
following three to be the main criticisms. First, that the PP, although
existing in different and possibly ill-defined versions, actually works as
a decisionmaking principle when properly interpreted. Second, that our
refutation of the epistemic version of the PP is misdirected because no
one holds the principle in this form, and furthermore it is fallacious.
Third, that our refutation of the PP as a principle of choice, or in
Sandin's terminology a prescriptive principle, is an example of what
Sandin calls the argument from absolutism, and can be dispelled by proper
attention to context. I will briefly discuss each of the three objections
to our position and show why they fail.
The Properly Defined Precautionary Principle
Works?
Is the problem with the PP only a problem of specification or
definition? Sandin admits that we should perhaps give up talking about the
PP with the definite article because there are many different versions of
the PP, but he quickly reverts to the use of the definite article, as if
there was only one canonical version. Be that as it may, he goes on to
point out that policymakers, legislators, and courts are slowly developing
interpretations of the PP and cites the European Commission for the views
that “it would be wrong to conclude that the absence of definition
has lead to legal uncertainty,” 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.”2
Commission of the European Communities (CEC). 2000.
Communication from the Commission on the precautionary principle.
COM/2000/0001/final p. 10.
What Sandin forgets to
tell us is that the European Commission is not a disinterested party. The
European Union relies on the PP in a number of trade disputes with the
United States and other parties, currently being decided in panels of the
World Trade Organization. The Commission communication that is cited is at
least partly an attempt to bolster the credentials of the PP in those
disputes. The Commission itself makes this point rather explicitly on page
9 of the Communication just before the sections that Sandin quotes:
“The Commission also wishes to stress in this Communication that,
far from being a way of evading obligations arising from the WTO
Agreements, the envisaged use of the precautionary principle complies with
these obligations.”
3 We can therefore not take the word of the Commission as the gospel
truth concerning the ease of interpreting and specifying the PP.
We still lack any reliable evidence that the PP is easy to apply when
one keeps context in mind, and if such evidence is forthcoming we would
need to know whether the PP applied by the decisionmakers has any likeness
to the versions of the PP discussed now. The so-called “PP”
that is actually used may be as far from the PP as a simple statable
principle, as current UK jurisprudence on the killing of human beings is
from “Thou shalt not kill.”
The Criticism of the Epistemic Version of
the Principle Is Fallacious
Sandin makes three points against our analysis of a possible epistemic
version of the PP: (1) that no one actually holds the PP as an epistemic
principle concerning belief; (2) that the epistemic versions people do
hold are concerned with which propositions to act upon, not which
propositions to believe; and (3) that a simple counterexample can show our
analysis to be faulty anyway.
With regard to the first and second of Sandin's criticisms, it is
initially important to note that even if he is right, it will not get the
PP off the incoherence hook. If the PP is an epistemic rule, our original
arguments showed that an epistemic PP concerned with beliefs will lead us
to include many false beliefs in our belief system. These arguments will
work just as well to show that Sandin's preferred epistemic PP
concerned with which propositions to act upon will lead us to act
wrongly/counterproductively/against the common good, and so
forth.
Sandin's epistemic PP is just as hopeless a principle as our
epistemic PP.
With regard to his third claim, it allows me to make an important
clarification. Sandin points out that it is not only the number of correct
beliefs in a belief system that counts, and that our elliptic statement
that applying the epistemic PP will lead us to include a large number of
false beliefs in our belief system is not enough to disqualify the PP as
an epistemic principle. A belief system might be better than another if
it, along with some trivial false beliefs, contains some important true
beliefs. This is obviously true, but the conclusion Sandin draws is highly
misleading. In the case of the epistemic PP, the situation is exactly the
reverse, as in Sandin's counterexample. What the epistemic PP will do
is make us adopt false beliefs about important issues, not about trivial
issues, because the proponent of the PP would presumably not wish to
invoke it in trivial matters. The epistemic PP will therefore produce a
belief system with trivial truths and important falsehoods. It is thus
still clear that the PP is an incoherent epistemic principle, as is any
principle that negatively affects the ratio between important truths and
important falsehoods in our belief system.
The Absolutist Objection
Sandin identifies our argument against the PP as a principle of choice
and merely a version of what he calls “the argument from
absolutism,” which “says that the precautionary principle, at
least some versions of it taken literally, will prohibit every
action, and thus offer no action guidance whatsoever.” And he
further states that any force of the argument from absolutism can easily
be dispelled because (1) it relies on a literal (mis)reading of the PP,
and (2) the real PP involves a specification of the precautionary action,
a specification of the threat context, and a specification of the level of
prior evidence of possible harm necessary to trigger the application of
the PP.
An initial response to this claim is to point out that there are
statements of the PP in official binding treaties that seem to support a
literal and absolutist reading. In the 1992 Convention on the Protection
of the Marine Environment of the Baltic Sea Area we can, for instance,
read the following in Article 3, subsection 2:
The Contracting Parties shall apply the precautionary
principle, i.e., to take preventive measures when there is reason to
assume that substances or energy introduced, directly or indirectly, in
the marine environment may create hazards to human health, harm living
resources and marine ecosystems, damage amenities or interfere with other
legitimate uses of the sea even when there is no conclusive evidence of a
causal relationship between inputs and their alleged effects.4
Convention on the Protection of the Marine
Environment of the Baltic Sea Area, 1992 (Helsinki Convention) Jul
2004 ed. Helsinki: HELCOM p. 4.
The “specification” of the threat context here is so
wide-ranging, from human health and the environment, to amenities or
interference with other legitimate uses, that it is difficult to give it a
nonabsolutist reading. If we had chosen to argue against this expansive
version of the PP, endorsed by the 10 ratifying parties to the convention,
which happens to include the European Union, our original paper could have
been very much shorter.
One might also wonder why a literal reading of the PP is wrong. Are
principles, at least important prescriptive principles, not supposed to be
read literally? If they are not supposed to be read literally should we
not instead call them rules of thumb or guiding metaphors or something
other than principles? Even in the case of prima facie
principles, the literal reading is the one we have to do for each
principle, before we can start prioritizing between them.
Sandin's assertion that our approach is absolutist must partly
rely on a misreading of our article, because the version of the PP that we
argue against and show to be incoherent contains exactly the
specifications that Sandin wants:
PP1: When an activity raises threats of serious and both
irreversible and irremediable harm to human health or the environment,
precautionary measures which effectively prevent the possibility of harm
(e.g. moratorium, prohibition etc.) shall be taken even if the causal link
between the activity and the possible harm has not been proven or the
causal link is weak and the harm is unlikely to occur.5
See note 1, Harris, Holm 2002:359.
Here we have a specification of the precautionary action, of the
threat context (human health or the environment) and of the threat
threshold, and implicitly of the prior level of knowledge necessary to
activate the PP.
Sandin also classifies our approach as absolutist because we point out
that there may be an infinite regress of precaution in certain cases where
scientific uncertainty exists at the beginning of a new development and at
every subsequent step in the development process. He does, however,
strangely fail to realize that his favorite remedies would not have broken
the regress in the concrete example we discuss, the case of GM plants.
Even if a PP specified in the threat dimension is applied in this case,
for instance, restricting application to the context of environmental
degradation, the regress would still run its course, and a specification
of level of prior evidence would not help either. At all points in the
development, important risks of the next step have been identified, risks
that could trigger the PP and stop further development. This was also true
at the very beginning. At the point just before the first GM plant was
produced, important risks had been identified with “a reasonable
scientific foundation.” That this foundation was extrapolation and
speculation does not necessarily show it not to be reasonable, because we
are talking about a point in time when very little empirical knowledge
existed.
The Argumentative Version of the
Principle
Sandin mentions one version of the PP that John Harris and I did not
consider in our article, the argumentative version, which is a principle
for what reasons or arguments are valid, and he points to principle 15 of
the Rio Declaration as an example of this version. Principle 15 states
that “lack of full scientific certainty shall not be used as a
reason for postponing cost-effective measures to prevent environmental
degradation.” For the sake of completeness, it is worth pointing out
that this version of the PP is also incoherent. How can we know that
measures are cost-effective in a situation where there is sufficient
scientific uncertainty to make the PP applicable?
Conclusion
As I have shown briefly above, all of Sandin's main arguments
against the analysis of the PP that John Harris and I presented in our
earlier paper fail—either because they are themselves fallacious or
because they fail to take into account some of the aspects of our
analysis. If anyone is guilty of knocking down straw men, it is Sandin and
not Harris and Holm.
The PP is incoherent and no amount of context can ever make the
incoherent coherent.
What we need to do is give up the idea that there is a coherent and
worthwhile PP to be found, defined, or specified, and instead start
discussing how we can make rational and reasonable policy decisions in
situations of great scientific uncertainty. It is unlikely that any single
principle will give us the answer to that question.
