200 • Impacts of Applied Genetics— Micro-Organisms, Plants, and Animals 
Discussion of this kind of harm is hindered by 
the difficulty not only of quantifying the prob- 
ability of an occurrence but also of predicting 
the type of damage that might occur. The differ- 
ent types of damage that can be conjured up are 
limited only by imagination. The scenarios have 
included epidemics of cancer, the spread of oil- 
eating bacteria, the uncontrolled proliferation 
of new plant life, and infection with hormone- 
producing bacteria. 
The risk of harm refers to the chance of harm 
actually occurring. In the present controversy, 
it has been difficult to distinguish the possible 
from the probable. It is, for instance, possible 
that an individual will be killed by a meteor fall- 
ing to the ground, but it is not probable. Analog- 
ous situations exist in genetic engineering. It is 
in this analysis that debate over genetic engi- 
neering has some special elements: the uncer- 
tainty of what kind of harm could occur, the un- 
certainty about the magnitude of risk, and the 
problem of the perception of risk. 
Identification of possible harm 
The first step in estimating risk is identifying 
the potential harm. It is not very meaningful to 
ask: How much risk does rDNA pose? The con- 
cept of risk takes on meaning only when harm is 
identified. The question should be: What is the 
likelihood that rDNA will cause a specific dis- 
ease such as in a single individual or in an entire 
population? The magnitude of the possible harm 
is incorporated in the question of risk, but dif- 
fers in the two cases. A statement about the risk 
of death to one person is different than one 
about the risk of death to a thousand. The right 
questions must be asked about a specific harm. 
Since no dangerous accidents are known to 
have occurred, their types remain conjectural. 
Identifying potential harm rests on intuition and 
arguments based on analogy. Even a so-called 
risk experiment is an approximation of subse- 
quent genetic manipulations. That is why ex- 
perts disagree. No uncontestable “scientific 
method” dictates which analogy is useful or ac- 
ceptable. By their very nature, all analogies 
share some characteristics with the event under 
consideration but differ in others. The goal is to 
discover the one that is most similar and to 
observe it often. This process then forms the 
basis for extrapolation. 
For example, it has been argued that ecologi- 
cal damage can be caused by the introduction of 
plants, animals, and micro-organisms into new 
environments. Scores of examples from histoi’v 
support this conclusion. I’he introduction to the 
United States of the Brazilian water hyacinth in 
the late 19th century has led to an infestation of 
the Southern waterways. Unconti’olled spread 
of English sparrows originally imported to con- 
trol insects has made eradication programs nec- 
essary. Countless other examines are confirma- 
tion that biological organisms may, at times, 
cause ecological damage when introduced into a 
new environment. Yet there is no agreement on 
whether such analogies are particularly reltv 
vant to assessing potential dangei's from genet- 
ically engineered organisms. It could he ar- 
gued— e.g., that a genetically engintHM'tuI orga- 
nism (carrying less than 1 pei’cent new genes) is 
still over 99 percent the same as the original, 
and is therefore not analogous to lh(> "totally 
new" organism introduced into an (‘cosystem. 
Some experts emphasize? the difhM’ences be- 
tween the situations; othei's emphasize? the? simi- 
larities. 
Other analejgies ha\e? he?e?n raise'el. Ne’w 
strains of influenza \ irus arise* re'geilarly. Some* 
can cause epielemics he?e'ause* the* pe)|)ulalie)n, 
never before expe)seel te) them, e'arrie*s no pro- 
tective antibodies. \'e?t e?an this analog_v sugge*sl 
that relatively harmless strains of E. roli might 
be transformed inte) e?[)iele?mie? |)athoge*ns? I he*r(* 
is disagreement, anel eie?hale*s e-onlinue* about 
what "could happen" e>r what is e*\(*n logie'ally 
possible. 
Estimates of harm: risk 
Assuming that agre?e?me*nt has he*e*n reae he*el 
on the possibility of a spe?e*ifie' harm, w hat can he 
done to ascei’tain the? probiibility'^ W hat is the* 
likelihood that elamage? will oe'cur? 
Damage invariably oe?e urs as the* result ol .i 
series of events, e?ach e)f whie h has its own par 
ticular chane;e e)f e)ce'uri ing. flow charts h.we 
been prepare?el te> iele?ntifv the*se* ste*|)s. \ t\pie-.il 
