Science in Society Archive

The Precautionary Principle is Coherent

Peter Saunders and Mae-Wan Ho argue: The precautionary principle is not an algorithm for making decisions, but a principle for making decisions based on available evidence. So let's look at the evidence.

In a recent article, Comstock claims that the precautionary principle commits us at the same time to two contradictory courses of action: that we should develop GM crops and that we should not (1). He concludes that the principle is therefore 'incoherent' and challenges defenders of the principle to explain why it is not.

Like so many other opponents of the precautionary principle, Comstock misunderstands its role. He assumes it is an algorithm for making decisions, which is why he writes of the principle as committing us to one or another course of action.

We want to emphasise that the precautionary principle is not an algorithm for making decisions. It does not make decisions for us, but it is a principle on which to base decisions. It is a principle for assigning the burden of proof, in much the same way that the defendant in a criminal court is assumed innocent until proven guilty 'beyond reasonable doubt' (see "Use and Abuse of the Precautionary Principle" by Peter Saunders, ISIS News#6 This important rule reflects society's view that convicting the innocent is far worse than acquitting the guilty. It has a profound effect on the outcome of many trials, but it still leaves the jury with a lot to do. They still have to weigh up the evidence, and they have to decide for themselves what constitutes 'reasonable doubt'.

In the same way, the precautionary principle requires us to assign the burden of proof to those who want to introduce a new technology, particularly in cases where there is little or no established need or benefit and where the hazards are serious and irreversible. It is up to the perpetrators to prove that the technology is safe 'beyond reasonable doubt' . We cannot expect the precautionary principle by itself to tell us what to do about GM crops or any other new technology. Like a jury, we have to weigh up the evidence, and like a jury we have to come to a decision.

This is a good time to consider the evidence as there is a public hearing going on in the UK on Aventis' Chardon LL GM maize, which the UK government is proposing to put on the National List. The transcripts of witness statements from citizens and scientists are all posted on the website of the UK Ministry of Agriculture, Fisheries and Food (

So, what is the evidence on GM crops? There is practically no evidence that they are safe, of the kind that could stand up in a court of law. There is less than a handful of papers on the subject of safety assessment published in peer-reviewed scientific journals. The vast majority consists of unpublished reports submitted to regulatory bodies for product approval, and these, far from supporting claims of safety, actually provides evidence to the contrary (see transcript of Dr. Arpad Pusztai's statement Chardon LL hearing, London, October 24, 2000). The published papers from the industry are no better. For example, Monsanto's study on Roundup Ready soya was seriously flawed. The two papers (Hammond et al, 1996, Journal of Nutrition 126, 717-26; and Padgette et al, 1996, Journal of Nutrition 126, 702-16) showed, among other things, significant increases in milk fat in cows and lower weight gains in male rats fed GM soya. There was also a 26.7% increase in a major allergen and growth inhibitor, ?-antitrypsin in the GM soya. Monsanto had failed to submit even more damning data indicating that another allergen, a soya lectin, was increased by 100% in retoasted soya beans ("Buried data in Monsanto's study on Roundup Ready soybeans" by Barbara Keeler (Ericka))

On the other hand, there is already plenty of evidence of actual and suspected hazards from findings reported in the scientific literature. Readers can see for themselves from all the transcripts of the Chardon LL hearing.

We summarise some of the findings which have been presented at the hearing (see Dr. Mae-Wan Ho's statement, October 26, 2000 and elsewhere (see World Scientists' Open Letter and other papers, .

  • GM genes such as those coding for bt toxins are harmful to beneficial and endangered insect species. Several of the toxins are also known to be actual or potential allergens for human beings.
  • New, unexpected toxins and allergens have arisen from the inherently random, uncontrollable nature of the process whereby GMOs are made.
  • GM constructs in GM plants have spread to related species by cross pollination, and weeds and superweeds resistant to multiple herbicides have appeared.
  • GM constructs containing antibiotic resistance genes have spread to bacteria in the soil and in the gut of bees. These bacteria constitute a reservoir of antibiotic resistance genes, which may be passed on to pathogenic bacteria, making infections very difficult to treat.
  • DNA is found not to be readily broken down by most commercial processing or in the gut of mammals.
  • The gut of livestock and human beings contain bacteria that can take up foreign DNA containing antibiotic resistance genes, making infections untreatable.
  • Horizontal transfer of GM genes to unrelated species has been found to create new viruses in many GM plants with viral genes in the GM construct.
  • GM constructs may spread more readily by horizontal gene transfer, if only because GM constructs are specifically designed to cross species barriers and to invade genomes.
  • All cells, including those of human beings, can readily take up GM constructs, and in some cases incorporate them into the cell's genome.
  • The cauliflower mosaic virus 35S promoter (CaMV 35S promoter) widely used to make GM genes over-express continuously, is active in practically all living species, plants, bacteria, algae, fungi, amphibian and human beings. Hence, any gene linked to it will be active in all species to which the GM construct is transferred. The CaMV 35S promoter can also make other host genes over-express. Over-expressing of certain cellular genes is associated with cancer in animals and human beings.
  • The CaMV 35S promoter has a recombination hotspot where it is prone to break and join up with other DNA, thus increasing the likelihood for horizontal gene transfer.
  • The CaMV 35S promoter is known to substitute in part or in whole for promoters of other viruses to give infectious viruses. That means it has the potential to wake up dormant viruses that have now been found in all genomes, plants and animals included.
  • GM lines are notorious unstable, do not breed true, and do not perform consistently in the field. Evidence is emerging on yield drag, increased use of herbicides, susceptibility to disease, and other failures.

Given the weight of evidence, it seems obvious to us that no GM crops should be planted in open fields, unless and until we can be convinced, by counter-evidence, that the risks are minimal. It also suggests that we should be far more cautious about genetic engineering altogether than we are at the present.

But by being cautious about this new technology, are we, as Comstock and others claim, running equal risks in the other direction? Do we risk losing potential benefits, or being able to deal with needs that may appear in 50 years time? Even if we were, that would not show that the precautionary principle is incoherent, only that we have to strike an appropriate balance. This might well be difficult in some cases, but not in this one.

The biotech companies and their supporters say we need GM crops to increase yield to feed a growing world population. Norman Borlaug, father of the green revolution and prominent supporter of agricultural biotechnology, claims GM crops are needed to feed a projected 10 billion. Let us look at the best available evidence. There is no scientific report documenting that yield has been increased in any GMO compared to non GMOs; quite the contrary is the case, as mentioned earlier, yield drags are frequently reported. What about population increase? The estimate produced by the United Nations Population Division is that world population growth had been slowing down since the 1960s. The total world population will peak at 7.7 billion in 2040, then go into long term decline to 3.6 billion by 2150, less than two-third of today's number. Similarly, the recent FAO report (Agriculture: Towards 2015/30, FAO Global Perspectives Studies Unit, July 2000) concludes that existing technologies, not counting GMOs, will produce enough and more than enough food to meet population growth. The real problem is one of distribution, as generally acknowledged. People are starving in the midst of plenty.

What about the possibility that at some time in the future we may have to make changes in the crops we grow and that genetic engineering may be needed? Or that with more research, gene biotechnologists will be able to produce new varieties that are indeed better and safer than the present ones? Even allowing for those possibilities does not mean we have to rush ahead with the present inadequately researched and tested technology. Nor does it mean we have to accept unsubstantiated promises that GM crops will provide the answer. And we certainly are not going to help our chances of developing appropriate new varieties by gene biotechnology or other methods, if we too hastily replace the diversity of crops that we still have by a few standard commercial GM varieties of unproven agronomic performance.

Looking at all the evidence and taking seriously the precautionary principle thus lead to the following conclusion. We should continue doing basic research in molecular genetics, including research relevant to the safety of GM constructs as well as making GM plants; for example, on how to modify existing genes precisely and safely, rather than to transfer in GM constructs haphazardly. But all that should be done in the laboratory and in the greenhouse under carefully contained conditions.

There should also be major effort devoted to developing better varieties of crops by conventional breeding and to research on organic, low-input farming methods. Agroecological farming methods which use crops and knowledge adapted to local conditions have been increasing yields two, three-fold since the 1980s. They provide social, environmental and health benefits in Latin America, Africa and Asia. There are good reasons to encourage farmers to grow and sell locally crops that are adapted to local conditions, and not to pressure them into growing national or international varieties for export. Export industrial agriculture is responsible for a great proportion of the fossil fuel consumption that contributes to climate change. Furthermore, there is incalculable health bonus to be gained in phasing out agrochemicals that are known to be linked to cancers and many other illnesses.

In that way, we can be confident about feeding the world today and for the foreseeable future, and we will still stand to gain from whatever benefits GM technology may bring. The only losers will be the biotech industry, because they cannot afford to wait. The rest of us can.

Prof. Peter Saunders Department of Mathematics
Dr. Mae-Wan Ho Institute of Science in Society

(1) Comstock's note, based on his talk at the meeting on Biotechnology held in Cambridge, MA, 22-23 September 2000, is available on

Article first published 31/10/00

Got something to say about this page? Comment

Comment on this article

Comments may be published. All comments are moderated. Name and email details are required.

Email address:
Your comments:
Anti spam question:
How many legs on a cat?