USE AND ABUSE OF THE PRECAUTIONARY PRINCIPLE
by Peter T. Saunders
Department of Mathematics,
King’s College,
Strand, London WC2R 2LS, UK.
There has been a lot written and said about the precautionary principle
recently, much of it misleading. Some have stated that if the principle
were applied it would put an end to technological advance. Others argue
that it fails to take science properly into account, though in fact it
relies more heavily on scientific evidence than other approaches to the
problem. Still others claim to be applying the principle when clearly they
are not. From all the confusion, you might think that it is a deep
philosophical idea that is very difficult for a lay person to grasp. (1)
In fact, the precautionary principle is very simple. All it actually
amounts to is a piece of common sense: if we are embarking on something
new, we should think very carefully about whether it is safe or not, and we
should not go ahead until we are convinced it is. It’s also not a new idea;
it already appears in national legislation in many countries (including the
United States), and in international agreements such as the 1992 Rio
Declaration and the Cartagena Biosafety Protocol agreed in Montreal in
2000.
Those who reject the precautionary principle are pushing forward with
untested, inadequately researched technologies and insisting that it is up
to the rest of us to prove that they are dangerous before they can be
stopped. At the same time, they also refuse to accept liability, so if the
technologies do turn out to be hazardous, as in many cases they already
have, someone else will have to pay the costs of putting things right.
The precautionary principle is about the burden of proof, a concept that
ordinary people have been expected to understand and accept in the law for
many years. It is also the same reasoning that is used in most statistical
testing. In fact, as a lot of work in biology depends on statistics,
neglect or misuse of the precautionary principle often arises out of a
misunderstanding and abuse of statistics.
The precautionary principle does not provide us with an algorithm for
decision making. We still have to seek apply the best scientific evidence
we can obtain and we still have to make judgements about what is in the
best interest of ourselves and our environment. Indeed, one of the
advantages of the principle is that it forces us to face these issues; we
cannot ignore them in the hope that everything will turn out for the best
whatever we do. The basic point, however, is that it places the burden of
proof firmly on the advocates of new technology. It is for them to show
that what they are proposing is safe. It is not for the rest of us to show
that it is not.
The Burden of Proof
The precautionary principle states that if there are reasonable scientific
grounds for believing that a new process or product may not be safe, it
should not be introduced until we have convincing evidence that the risks
are small and are outweighed by the benefits. It can also be applied to
existing technologies when new evidence appears suggesting that they are
more dangerous than we had thought, as in the cases of cigarettes, CFCs,
lead in petrol, greenhouse gasses and now genetically modified organisms
(GMOs). (2) In such cases it requires that we carry out research to gain a
better assessment of the risk and, in the meantime, that we should not
expand our use of the technology but should put in train measures to reduce
our dependence on it. If the dangers are considered serious enough, the
principle may require us to withdraw the products or impose a ban or
moratorium on further use.
The principle does not, as some critics claim, require industry to provide
absolute proof that something new is safe. That would be an impossible
demand and would indeed stop technology dead in its tracks, but it is not
what is being demanded. The precautionary principle does not deal with
absolute certainty. On the contrary, it is specifically intended for
circumstances in which there is no absolute certainty. It simply puts the
burden of proof where it belongs, with the innovator. The requirement is to
demonstrate, not absolutely but beyond reasonable doubt, that what is being
proposed is safe.
A similar principle applies in the criminal law, and for much the same
reason. In the courtroom, the prosecution and the defence are not on equal
terms. The defendant is not required to prove his innocence and the jury is
not asked to decide merely whether they think it is more likely than not
that he committed the crime. The prosecution must establish, not absolutely
but beyond reasonable doubt, that the defendant is guilty.
There is a good reason for this inequality, and it has to do with the
uncertainty of the situation and the consequences of taking a wrong
decision. The defendant may be guilty or not and he may be found guilty or
not. If he is guilty and convicted, then justice has been done, as it has
if he is innocent and found not guilty. But suppose the jury reaches the
wrong verdict, what then?
That depends on which of the two possible errors was made. If the defendant
actually committed the crime but is found not guilty, then a crime goes
unpunished. The other possibility is that the defendant is wrongly
convicted of a crime, in which case his whole life may be ruined. Neither
of these outcomes is satisfactory, but society has decided that the second
is so much worse than the first that we should do as much as we reasonably
can to avoid it. It is better, so the saying goes, that a hundred guilty
men should go free than that one innocent man should be convicted.
In any situation in which there is uncertainty, mistakes will occur. Our
aim must be to minimise the damage that results when they do.
Just as society does not require a defendant to prove his innocence, so it
should not require objectors to prove that a technology is harmful. It is
up to those who want to introduce something new to prove, not with
certainty but beyond reasonable doubt, that it is safe. Society balances
the trial in favour of the defendant because we believe that convicting an
innocent person is far worse than failing to convict someone who is
actually guilty. In the same way, we should balance the decision on risks
and hazards in favour of safety, especially in those cases where the
damage, should it occur, is serious and irredeemable.
The objectors must bring forward evidence that stands up to scrutiny, but
they do not have to prove there are serious dangers. The burden of proof is
on the innovators.
The Misuse of Statistics
You have an antique coin that you want to use for deciding who will go
first in a game, but you are worried that it might be biased in favour of
heads. You toss it three times, and it comes down heads every time.
Naturally, this does nothing to reassure you. Then along comes someone who
claims to know about statistics. He carries out a short calculation and
informs you that as the “p-value” is 0.125, you have nothing to worry
about. The coin is not biased.
Now this must strike you as nonsense, even if you don’t understand
statistics. Surely if a coin comes down heads three times in a row, that
can’t prove it is unbiased? No, of course it can’t. But this sort of
reasoning is being used to prove that GM technology is safe.
The fallacy, and it is a fallacy, comes about through either a
misunderstanding of statistics or a total neglect of the precautionary
principle – or, more likely, both. In brief, people are claiming to have
proven that something is safe when what they have actually done is to fail
to prove that it is unsafe. It’s the mathematical way of claiming that
absence of evidence is the same as evidence of absence.
To see how this comes about, we have to appreciate the difference between
biological and other kinds of scientific evidence. Most experiments in
physics and chemistry are relatively clear cut. If we want to know what
will happen if we mix copper and sulphuric acid, we really only have to try
it once. We may repeat the experiment to make sure it worked properly, but
we expect to get the same result, even to the amount of hydrogen that is
produced from a given amount of copper and acid.
Organisms, however, vary considerably and don’t behave in closely
predictable ways. If we spread fertiliser on a field, not every plant will
increase its growth by the same amount, and if we cross two lines of maize,
not all the resulting seeds will be the same. We often have to use some
sort of statistical argument to tell us whether what we have observed
represents a real effect or is merely due to chance.
The details of the argument will vary depending on exactly what it is we
want to establish, but the standard ones follow a similar pattern.
Suppose that plant breeders have come up with a new variety of maize and we
want to know if it gives a better yield than the old one. We plant one
field with each of them, and we find that the new variety does actually
produce more maize.
That’s encouraging, but it doesn’t prove anything. After all, even if we
had planted both fields with the old strain, we wouldn’t have expected to
get exactly the same yield in both. The apparent improvement might be just
a chance fluctuation.
To help us decide whether the observed effect is real, we carry out the
following calculation. We suppose that the new strain is actually no better
than the old one. This is called the “null hypothesis” because we assume
that nothing has changed. We then estimate as best we can the probability
that the new strain would perform as well as it did simply on account of
chance. We call this probability the p-value.
Obviously, the smaller the p-value the more likely it is that the new
strain really is better, though we can never be absolutely certain. What
counts as a small enough value of p is arbitrary, but over the years
statisticians have adopted the convention that if p is less than 5% we
should reject the null hypothesis, i.e. we may infer that the new strain is
better. Another way of saying this is that the increase in yields is
‘significant’.
Why have statisticians fastened on such a small value? Wouldn’t it be
reasonable to say that if there is less than an even chance (i.e. p=0.5) of
such a large increase then we should infer that the new strain is better?
No, and the reason why not is simple. It’s a question of the burden of
proof. Remember that statistics is about taking decisions in the face of
uncertainty. It is a serious business advising a company to change the
variety of seed it produces or a farmer to switch from one he has grown for
years. There could be a lot to lose if we are wrong. We want to be sure
beyond reasonable doubt that we are right, and that’s usually taken to mean
a p=value of 0.05 or less.
Suppose we obtain a p-value of greater than 0.05. What then? We have failed
to prove that the new strain is better. We have not, however, proved that
it is no better, any more than by finding a defendant not guilty we have
proved that he is innocent.
In the example of the antique coin, the null hypothesis was that the coin
was fair. If that were the case, then the probability of a head on any one
throw would be 1/2, so the probability of three heads in a row would be
(1/2)3=0.125. This is greater than 0.05, so we cannot reject the null
hypothesis. Thus we cannot claim that our experiment has shown the coin to
be biased.
Up to that point, the reasoning was correct. Where it went wrong was in the
claim that the experiment has shown the coin to be fair. It did no such
thing.
Yet that is precisely the sort of argument that we see in scientific papers
defending genetic engineering. A recent report “Absence of toxicity of
Bacillus thuringiensis pollen to black swallowtails under field conditions”
(3) claims by its title to have shown that there is no harmful effect. In
the discussion however, the authors state only that there were “no
significant weight differences among larvae as a function of distance from
the corn field or pollen level.” In other words, they have only failed to
demonstrate that there is a harmful effect. They have not proven that there
is none.
A second paper (4) claims to show that transgenes in wheat are stably
inherited. The evidence for this is that the “transmission ratios were
shown to be Mendelian in 8 out of 12 lines.” In the accompanying table,
however, six of the p-values are less than 0.5 and one is 0.1. That is not
sufficient to prove that the genes are unstable and so inherited in a
non-Mendelian way. But it does not prove they are, which is what was
claimed. (5)
The way to decide if the antique coin is biased is to toss it more times
and see what happens. In the case of the safety and stability of GM crops,
more and better experiments should be carried out.
The Anti-Precautionary Principle
The precautionary principle is so obviously common sense that we might
expect it to be universally adopted. That would still leave room for debate
about how big the risks and benefits are likely to be, especially when
those who stand to gain if things go right and those who stand to lose if
they do not are not the same. It is significant that the corporations are
implacably opposed to proposals that they should be liable for any damage
caused by the products of GM technology. They are demanding a one-way bet:
they pocket any gains and someone else pays for any losses. It also gives
us an idea of how confident they are about the safety of the technology.
What is harder to understand is why our regulators are still so reluctant
to adopt the precautionary principle. They tend to rely instead on what we
might call the anti-precautionary principle: When a new technology is
proposed, it must be approved unless it can be shown conclusively to be
dangerous. The burden of proof is not on the innovator; it is on the rest
of us.
The most enthusiastic supporter of the anti-precautionary principle is the
World Trade Organisation (WTO), the international body whose task it is to
promote free trade. A country that wants to restrict or prohibit imports on
grounds of safety has to provide definite proof of hazard, or else be
accused of erecting artificial trade barriers. A recent example is the
WTO’s judgement that the European Union’s ban on US growth-hormone injected
beef is illegal.
By applying the anti-precautionary principle in the past, we have allowed
corporations to damage our health and our environment through cigarette
smoking, lead in petrol, and high levels of toxic and radioactive wastes
that include hormone disrupters, carcinogens and mutagens. The costs in
human suffering and environmental degradation and in resources to attempt
to put these right have been very high indeed. Politicians should bear this
in mind.
Conclusion
There is nothing difficult or arcane about the precautionary principle. It
is the same reasoning that is used every day in the courts and in
statistics. More than that, it is just common sense. If we have genuine
doubts about whether something is safe, then we should not use it until we
are convinced it is. And how convinced we have to be depends on how much we
really need it.
As far as GM crops are concerned, the situation is clear. The world is not
short of food. Where people are going hungry it is because of poverty.
Hardly anyone believes that there will be a real shortage within 25 years,
and a recent FAO report predicts that improvements in conventional
agriculture and reductions in the rate of increase of the world’s
population will mean we will continue to be able to feed ourselves
indefinitely.
On the other side, there is both direct and indirect evidence that gene
biotechnology may not be safe for health and the environment. The benefits
of GM agriculture remain hypothetical.
We can easily afford a five-year moratorium to support further research
into improving the safety of gene biotechnology and making it more precise
and more effective. We should also use the time to develop better methods
of sustainable farming, organic or low-input, which do not have the same
potentially disastrous risks.
PTS 21/08/00
(1) See, for example, S. Holm and J. Harris (Nature, 400 (1999) 398).
Compare C.V. Howard & P.T. Saunders (Nature 401 (1999) 207) and C.
Rafffensburger et al. (Nature 401 (1999) 207-208).
(2) We are now told that in the case of tobacco and lead, many in the
industry knew about the hazards long before the public did. It is not
always wise to accept broad and unsupported assurances about safety from
those who have a very strong interest in continuing the technology.
(3) A.R. Wraight et al (2000), Proceedings of the National Academy of
Sciences (early edition). Quite apart from the use of statistics, it
generally requires considerable skill to design and carry out an experiment
to provide a convincing demonstration that an effect does not occur. It is
all too easy to fail to find something even when it is there.
(4) M.E. Cannell et al. Theoretical and applied Genetics 99 (1999) 772-784.
(5) In fact, an alternative statistical analysis of the data indicates
considerably more instability than the authors found.