that they are all uncertain, let alone that they are uniformly uncertain. Examining
how they were constructed, and identifying the elements that contributed and
the means by which they were hybridized, can enable scholars, policy-makers
and citizens to make well-informed judgements about their relative reliability
(van Zwanenberg and Millstone, 2000).
One scientific field that is characterized by profound uncertainties concerns the
implications of studies using laboratory animals to try to identify the potential adverse
effects of chemicals on public and environmental health. In 1984, a leading US
toxicologist explained that: ‘Given all the uncertainties, it may seem that the very
laborious and expensive tests for assessing safety may be no better than throwing
darts at a board full of numbers’ (Wodicka, 1984). In the same year I asked the
UK’s then leading professor of toxicology what was the relevance of tests of rats
and mice to humans, and his reply was the one I least expected, namely: ‘Your
guess is as good as mine’ (pers. comm. Guildford, May 1984).
While extrapolations from models may be uncertain, estimates of the magnitude
of the attendant uncertainties are rarely provided. In the late 1970s a valiant
innovative attempt was made by a US National Academy of Sciences panel to
estimate the upper and lower bounds of the carcinogenic risk that the artificial
sweetener saccharin could pose to the US population. The panel estimated those
bounds by comparing the most pessimistic interpretation of the most worrying
evidence against the most optimistic interpretation of the most reassuring study.
The resulting estimate was that if, on average, US residents were to continue to
ingest ~120 milligrams of saccharin daily for a period of 70 years (which then
corresponded to the average level of consumption in the USA) it was unlikely that
fewer than 0.22 extra deaths from cancer would occur over that period, while it
was unlikely that more than 1,144,000 extra deaths would occur (US NAS, 1978).
In other words, estimates of the potential carcinogenicity of saccharin for humans
were characterized by uncertainty of six orders of magnitude! A remarkable fact about
that analysis has been the conspicuous reluctance of toxicology professionals or
policy-makers to ever cite that conclusion or acknowledge its implications.
Bizarrely, expert advisory committees in the USA, the UK, the European Commis -
sion (EC)/ European Union (EU) and at the WHO have subsequently all chosen
to assign numerical values to an ‘acceptable daily intake’ for saccharin in the form
of single point estimates, some of which were even specified to one decimal place,
rather than as ranges (CEC, 1977; CEC, 1985; MAFF, 1990; WHO, l993; US
FDA, 2013). The practice of ignoring, neglecting and/or concealing uncertainties,
and providing single point estimates is unfortunately widespread.
One of the reasons why expert advisers too frequently ignore and/or conceal
uncertainties is because there is a ‘market’ for their misrepresentations. Policy-makers,
such as government ministers and European Commissioners, often prefer to use
their advisory committees to shield them from having to take responsibility for
contestable and contested decisions. When expert advisory committees or panels
highlight the existence, and even the magnitude, of policy-relevant uncertainties,
policy-makers are obliged to take some responsibility for decision-making in
Invoking ‘science’ in green transformations 45
how they were constructed, and identifying the elements that contributed and
the means by which they were hybridized, can enable scholars, policy-makers
and citizens to make well-informed judgements about their relative reliability
(van Zwanenberg and Millstone, 2000).
One scientific field that is characterized by profound uncertainties concerns the
implications of studies using laboratory animals to try to identify the potential adverse
effects of chemicals on public and environmental health. In 1984, a leading US
toxicologist explained that: ‘Given all the uncertainties, it may seem that the very
laborious and expensive tests for assessing safety may be no better than throwing
darts at a board full of numbers’ (Wodicka, 1984). In the same year I asked the
UK’s then leading professor of toxicology what was the relevance of tests of rats
and mice to humans, and his reply was the one I least expected, namely: ‘Your
guess is as good as mine’ (pers. comm. Guildford, May 1984).
While extrapolations from models may be uncertain, estimates of the magnitude
of the attendant uncertainties are rarely provided. In the late 1970s a valiant
innovative attempt was made by a US National Academy of Sciences panel to
estimate the upper and lower bounds of the carcinogenic risk that the artificial
sweetener saccharin could pose to the US population. The panel estimated those
bounds by comparing the most pessimistic interpretation of the most worrying
evidence against the most optimistic interpretation of the most reassuring study.
The resulting estimate was that if, on average, US residents were to continue to
ingest ~120 milligrams of saccharin daily for a period of 70 years (which then
corresponded to the average level of consumption in the USA) it was unlikely that
fewer than 0.22 extra deaths from cancer would occur over that period, while it
was unlikely that more than 1,144,000 extra deaths would occur (US NAS, 1978).
In other words, estimates of the potential carcinogenicity of saccharin for humans
were characterized by uncertainty of six orders of magnitude! A remarkable fact about
that analysis has been the conspicuous reluctance of toxicology professionals or
policy-makers to ever cite that conclusion or acknowledge its implications.
Bizarrely, expert advisory committees in the USA, the UK, the European Commis -
sion (EC)/ European Union (EU) and at the WHO have subsequently all chosen
to assign numerical values to an ‘acceptable daily intake’ for saccharin in the form
of single point estimates, some of which were even specified to one decimal place,
rather than as ranges (CEC, 1977; CEC, 1985; MAFF, 1990; WHO, l993; US
FDA, 2013). The practice of ignoring, neglecting and/or concealing uncertainties,
and providing single point estimates is unfortunately widespread.
One of the reasons why expert advisers too frequently ignore and/or conceal
uncertainties is because there is a ‘market’ for their misrepresentations. Policy-makers,
such as government ministers and European Commissioners, often prefer to use
their advisory committees to shield them from having to take responsibility for
contestable and contested decisions. When expert advisory committees or panels
highlight the existence, and even the magnitude, of policy-relevant uncertainties,
policy-makers are obliged to take some responsibility for decision-making in
Invoking ‘science’ in green transformations 45