technology introductions at the national level, this social dialogue

is very complex, involving both scientific debates and conflicts

between competing societal goals. At the international level,

complexity increases, as different countries, with different cultures

and at different stages of technological and economic development,

tend to have very different perspectives on how to strike the right

era of unprecedented economic progress, driven largely by

technological innovation. The pace of change is not slowing; indeed,

it may be accelerating, in response to burgeoning global market

forces and exponential growth in scientific information. One of the

strongest areas of science and investment as we open the 21st century

the past 50 years, many of which involved advances in chemistry, have

improved life immeasurably for most people, at least in the developed

countries. We live much longer, and the quality of our lives is

better in countless ways. Most people would agree that the benefits

to humanity of this half-century of technological progress probably

vastly outweigh whatever harm has been done to human health and the

that some new chemicals, and the ways we produce and use them, have

contributed to human diseases and have adversely affected ecosystems,

sometimes in unarguably harmful ways, and perhaps more often, in

subtler ways about whose ultimate consequences we are not yet fully

certain. Often, these adverse effects were not foreseen when the

technology was introduced. As the scientific evidence developed

documenting risks, intense controversies ensued, as those concerned

mostly about risks were challenged by those with an economic interest

effects of new chemicals over the past half-century has led to

growing support for application of the so-called "Precautionary

Principle." The precautionary approach calls for developing better

mechanisms for anticipating adverse side-effects of new technologies,

and for reviewing technologies more thoroughly, exploring alternative

ways for reaping benefits while minimizing adverse collateral

such decisions have been left to market forces. Innovators develop

and offer for sale products they believe are better in some important

way, and if capital investors and buyers agree, the technology

spreads. Governments oversee the process and may regulate the

introduction of technologies to some extent, but in free-market

societies, most legal and regulatory institutions are heavily biased

in the direction of permitting new products to enter the market,

assessing risks retrospectively, and demanding persuasive scientific

also rewarded ignorance. According to a study done several years ago

by the U.S. National Research Council, about 80,000 high-volume

chemicals are produced in the U.S., and adequate toxicological data

exist for less than 5 percent of them. Today, a definition of

"adequate" data would undoubtedly include an assessment of effects on

the developing nervous, endocrine and immune systems, and very few

chemicals have been tested for these effects. Data on effects of

single substances cannot predict the interactive effects of the

multiple chemicals to which people are routinely exposed. Efforts to

assess the potential effects of chemicals on critical human

developmental processes are thus constrained severely by limitations

of science. Assessment methods for other hazards associated with

foods, such as microbiological contaminants and genetically modified

crops, are less robustly developed than those for chemicals. It is

therefore sometimes impossible, using available risk assessment

regulatory philosophies in most Western, developed countries have

traditionally had a large, inherently precautionary component. Most

national governments, for example, require toxicity testing of food

additives to ensure that they are safe, before they are permitted to

be used in foods. Food plays a central role in our health and in our

cultures, and most societies agree that more caution should be

heavily on the use of risk assessment and on decisions based on

"science." A consensus has grown that this approach has sometimes

failed to adequately prevent serious food safety problems. The

bovine spongiform encephalopathy (BSE) outbreak in the United

Kingdom, and contamination of foods with dioxins, a problem in many

countries, are recent examples of failures of the

risk-assessment-based food safety paradigm. Such problems, coupled

with growing awareness in all sectors of some inherent limitations of

risk assessment, have stimulated interest in applying the

"Precautionary Principle" more explicitly in food safety risk

to scientific risk assessment, but rather an extension and expansion

of science in risk assessment. I have coined a term, precautionary

risk assessment, to emphasize the strong link between precaution and

science. The essence of precautionary risk assess-ment is to

treat scientific questions scientifically. Often, in food

safety risk analysis, science is used politically. Risk assessments

are narrowly focused on questions where ample scientific data exist,

and seem designed to show that risks are "acceptable." A

precautionary risk assessment takes a broader approach,

defining a full array of risk-related questions needing answers. The

assessment then looks rigorously at such issues as how much data

exist about a given risk, which questions cannot be adequately

answered with the available data, the possible consequences of being

unable to answer certain questions (i.e., the risks of ignorance),

the knowledge gaps that need to be filled to get better answers, and

whether available scientific methods can answer all the important

provides a better basis for decisions as to whether we should proceed

to adopt a technology that has risks of unknown magnitude, or whether

we should take more time and try to find alternative ways to benefit

from the technology while avoiding the risks, before a new technology

chemicals that are deliberately added to foods generally are

rigorously assessed for safety, many other economically important

chemicals are dispersed in the environment, and contaminate our

foods, often at very low levels. A British journalist named Lucy

Johnston received wide publicity earlier this year, when she had a

sample of her body fat analyzed for chemicals, and wrote about what

the tests found. Hundreds of different pollutants, most of them

pesticides, were detected in her tissue. Analysts who have tested

larger populations this way report that more than 500 industrial

chemicals are commonly found in the average person's body fat. And

those represent only those chemicals that are readily detected by

available analytical screening methods; they are but the tip of the

contaminants. The cumulative health risks posed by the chemicals in

our foods are largely unknown, and perhaps unknowable. Most dietary

residues have not been tested adequately for toxic effects,

especially for effects on developmental processes that many

toxicologists now think are most sensitive low-dose damage. If our

health is being adversely affected by chemical contaminants, it will

be very difficult to measure the effects with existing scientific

methods. Animal test data on single substances can't replicate

uneven exposure to mixtures of chemicals that people encounter.

Studies of human disease patterns can't sort out all the confounding

variables, and there are no suitable unexposed populations to serve

as "controls" for such studies. In short, in trying to learn whether

pollutants in our foods are harming us, we are limited by the

magnitude, many people respond with what psychologists call "denial,"

or refusal to acknowledge the problem. One classic form of denial is

to assert that no harm is occurring, since none has been

scientifically demonstrated. Another popular response is to claim

that exposures to chemicals known to be harmful at high doses are

invariably safe at low doses. Both of these commonplace assertions

rest on heuristics-that is, familiar, everyday strategies for

scientists as well as non-scientists are often colored by a

deep-seated belief that something that our senses cannot detect could

not be harmful. Many people, including many scientists, seem

convinced that "low-level" exposure to chemicals must be safe, and

define "low level" approximately by the concentrations that are

quantifiable with current analytical methods. Over the years, I have

been assured by many sincere people that one part per million of this

or one part per billion of that cannot possibly be harmful, it's just

non-scientific; it is based on intuition and perhaps faith, but not

on rigorous examination of scientific data. A scientific perspective

on this question might consider the fact that the appropriate unit of

chemical exposure, in terms of biological activity, is a molecule,

rather than a part per billion. (Toxic effects occur, after all, at

look at an example. Baby bottles made of polycarbonate plastic can

release traces of bisphenol-A, a chemical with estrogen-like effects,

into liquids they contain. If the bottle holds infant formula, a

baby might be continuously exposed to a hormonally active agent at

polycarbonate bottles) have asserted that 1 ppb of bisphenol-A is too

low an exposure to have biological effects. But what is the science

behind this statement? There are no data on effects of bisphenol-A

on human babies. Some animal tests have reported adverse effects of

fetal exposure on the developing reproductive system, but the data

are not definitive yet, and have been hotly disputed by the

the question. Using basic, undisputed facts-the molecular weight of

bisphenol-A, Avogadro's number, and the volume of a baby bottle-one

can easily calculate that a 200 ml bottle of fluid contaminated with

1 ppb of bisphenol-A contains roughly 500 trillion (that is,

two perspectives. One, based on firm conviction but no data, asserts

that there is no effect of bisphenol-A in baby bottles, because none

has been observed scientifically and because one part per billion of

BPA is "too low" an exposure level to have biological effects. The

other, based on simple, undisputed scientific facts, notes that

polycarbonate bottles can expose babies to unimaginably large numbers

of molecules of an estrogen-like chemical, several times a day. We

must ask, on what basis can we presume that such exposure has no

biological effects? What if "low-level" exposure is not

intrinsically "safe;" what if, instead, our inability to measure

this case would emphasize not the lack of concrete data showing harm

in babies exposed to 1 ppb of BPA in their formula, but rather would

recognize that 1 ppb is not necessarily a "low" exposure. It would

assess the difficulties of knowing whether or not the quadrillions of

molecules a baby ingests daily have any harmful effects on the tiny

"biotechnology revolution" will probably expand and accelerate over

coming decades, especially if the developers of GM crops can begin to

deliver the promised but to date largely unachieved benefits of

genetic engineering, in terms of more abundant and sustainable food

including many in the consumer movement, are deeply concerned that

risks associated with biotechnology are not adequately understood,

and that the rush to commercialize GM crops almost guarantees that if

evidence should ultimately emerge of damage done by these crops, it

to enhance crops' resistance to pests seem likely to pose risks

similar in many ways to those of chemical pesticides. The benefits

of chemical pest control sped widespread adoption of the new

technology, while the negative side-human health hazards and

ecological effects that made pest control more difficult and harmed

non-pest wildlife-were documented only gradually, over decades. Many

who recall that experience would prefer not to see it repeated with

yet know enough about the nature of genetically modified organisms to

be sure how they will behave in natural ecosystems. Nor do we

understand how ecosystems work fully enough to be sure how

introduction of genetically modified organisms, and over time,

combinations of many modified organisms, will affect the health and

the stability of our natural life-support systems. Consumers, who up

to this point have derived few direct benefits from genetically

modified crops, are generally open-minded about the technology,

anticipating benefits, but also interested in ensuring that safety

an opportunity to learn from experience with chemicals, and to take

another approach with genetically modified crops. The alternative

approach would exercise more caution at the outset, take a more

precautionary approach to risk assessment, and give more weight to

the need to avoid future harm, without abandoning the many potential

of the day cannot be resolved by relying on scientific data and

traditional risk assessment methods. As our understanding of risk

advances, we have learned that many questions about food-related

risks cannot be answered with current knowledge. Precaution,

sensibly applied, is one useful tool for making decisions of this

science and risk assessment. More accurately, it requires an even

more rigorous use of science. It pays greater attention to what

science does not know, and to the possible consequences of

a shift in philosophy on some long-standing conflicts in societal

values. One involves the concept of "burden of proof." For

decades, new technologies have been presumed safe until proven

harmful. Today, there is a growing tendency to place a greater

burden on proponents of a new technology, to demand that risk

questions be better identified and addressed, before innovations are

widely adopted. This reflects social learning from past mistakes,

and a greater sense of equity-an assertion that consumers, and future

generations, have the right not to have risks imposed upon them

without more discussion of who is benefiting, and of how much risk is

greater role for government, and less reliance on unbridled market

forces, to chart the course of technology. It requires a conscious

effort to look for alternative solutions to food-related technical

problems, and to choose options with the best risk/benefit ratios.

Attempts to "control" technology risk stifling innovation, and

governments will proceed cautiously, seeking the right balance. But

a better balance must be found than has prevailed for the past 50

moving towards expanding use of precaution. I believe that this is

the right direction, and that the trend will continue. I think most

consumers would like to avoid repeating the history of chemical

innovations, as the biotechnology revolution unfurls. We need to

learn from our experience with chemicals, by reflecting objectively

on what we still do not know about potential harm to our health and

our planet from the myriad contaminants in our foods and our bodies.

It is too late to "call back" dispersed chemicals, but not too late

to prevent the release of potentially harmful transgenic organisms.

Undoubtedly, we can do a better job of understanding what we need to

know, and of gathering data to inform our decisions, before we

assessment will not be easy. In the Codex debate, simply defining

the terms clearly enough to permit a trans-national, trans-cultural

dialogue on this topic has proven exceedingly difficult. Sorting out

science and value trade-offs is complicated enough, and in the

international arena, national interests in promoting trade, private

sector resistance to restrictions on markets, and other political

factors have all confounded efforts to improve food safety risk

point will differ for different societies, and that a developing

country may choose to pursue the benefits of rapid economic growth,

and be less precautionary about risks than a wealthy nation with

mature technologies might prefer. An international consensus on the

precautionary approach to risk assessment is essential to meet the

food safety challenges of the 21st century, and I have faith that all