John Harris gives a comprehensive and generally valid defense of
embryo research. Although nearly all his arguments are valid,
one—the argument comparing natural reproduction to embryo
research—is problematic in several important ways. I focus here
on that argument.
Harris begins his provocative paper1
Harris J. Cambridge Quarterly of Healthcare Ethics
2003;12(4):353–71.
in this way:
I make two appeals to consistency, or to “parity of
reasoning,” that I believe show that no one who either has used
or intends to use sexual reproduction as their means of procreation,
nor indeed anyone who has unprotected heterosexual intercourse, nor
anyone who finds in vitro fertilization (IVF) acceptable, nor anyone
who believes that abortion is ever permissible can consistently object
on principle to human embryo research nor to the use of embryonic stem
cells for research or therapy. (p. 353, and see notes cited
therein)
Harris is concerned with “consistency [of treatment]
of stem cell research with what is regarded as acceptable and ethical
with respect to normal sexual reproduction (p. 361)” Harris makes
four (not two, as he claims) appeals to parity of reasoning about
embryo research and
- engaging in sexual reproduction
- intending to engage in sexual reproduction
- IVF, and
- abortion.
Although his arguments for the third and fourth are successful, his
arguments for the first and second fail. I will argue that the lessons
from natural reproduction are different from those that Harris wishes
to draw.
The following is Harris's argument in full.
Let us start with the free and completely unfettered liberty to
establish a pregnancy by sexual reproduction without any
“medical” assistance. What are people and societies who
accept this free and unfettered liberty committing themselves to? What
has a God who has ordained natural procreation committed herself
to?
We now know that for every successful pregnancy that results in
a live birth many, perhaps as many as five, [note 47: Robert Winston gave
the figure of five embryos for every live birth some years ago in a
personal communication. Anecdotal evidence to me from a number of
sources confirms this high figure, but the literature is rather more
conservative, making more probable a figure of three embryos lost for
every live birth. See: Boklage CE. Survival probability of human
conceptions from fertilization to term. International Journal of
Fertility 1990;35(2)75–94. See also: Leridon H. Human
Fertility: The Basic Components. Chicago: University of Chicago
Press; 1977. Again, in a recent personal communication, Henri Leridon
confirmed that a figure of three lost embryos for every live birth is a
reasonable conservative figure.] early embryos will be lost or
“miscarry” …
How are we to think of the decision to attempt to have a child
in the light of these facts? One obvious and inescapable conclusion is that
God and/or nature has ordained that “spare” embryos be
produced for almost every pregnancy and that most of these will have to
die in order that a sibling embryo can come to birth. Thus, the
sacrifice of embryos seems to be an inescapable and inevitable part of
the process of procreation. It may not be intentional sacrifice, and it
may not attend every pregnancy, but the loss of many embryos is the
inevitable consequence of the vast majority (perhaps all) pregnancies.
For everyone who knows the facts, it is a conscious, knowing, and
therefore deliberate sacrifice; and for everyone, regardless of
“guilty” knowledge, it is part of the true description of
what they do in having or attempting to have children.
We may conclude that the production of spare embryos, some of
which will be sacrificed, is not unique to assisted reproduction technologies
(ART); it is an inevitable, (and presumably acceptable, or at least
tolerable?) part of all reproduction.
Both natural procreation and ART involve a process in which
embryos, additional to those that will actually become children, are created
only to die. I will continue to call these “spare” embryos
in each case. If either of these processes is justified it is because
the objective of producing a live healthy child is judged worth this
particular cost. (p. 362)
Harris argues that there is no difference between intending to bring
about an outcome and merely foreseeing it. One example he gives is that
it would be wrong to use a mobile phone near a pregnant woman if using
that phone exposed the woman to a high chance of having a
miscarriage.
For Harris, if
- an act that has the foreseeable consequences of bringing into
existence a new life and causing the death of four embryos is
permissible, then
- an act that has the foreseeable consequences of saving an existing
life and causing the death of four embryos is permissible.
Harris assumes, plausibly, that we have as much reason to save an
existing life as we do to bring a new life into existence. He also
assumes, again plausibly, that embryo research will result in the
saving of existing lives.
Where This Argument Fails
Nonmaleficence
Harris's mobile phone example is important. The mobile phone
user in his example violates a principle of nonmaleficence.
Nonmaleficence is usually defined as “not harming.” But we
should clarify or extend this definition to include not exposing people
to unreasonable risks of harm. The moral of the mobile phone story is
that we should not expose people to unreasonable risks.
Does natural reproduction expose the embryo to unreasonable risks? If
the embryo is not a person with rights, the question is redundant. But
Harris, although he believes the embryo is not a person, attempts to
set this issue aside. He is trying to get leverage on those who may
believe the embryo is a person but nonetheless believe natural
reproduction is permissible despite the high embryo wastage it
entails.
It is debatable whether natural reproduction imposes any
risk at all on any embryo. The risk of dying (let us assume it is 80%)
is inherent to the embryo's nature. It is not as if the same
embryo could have been created without that chance of dying.
Let us assume for argument's sake, however, that natural
reproduction can be said in some sense to have imposed an 80% chance of
dying on the embryo. This is very high. If I gave you a drug that had
an 80% chance of killing you, I would be acting very wrongly. But in
deciding whether this risk is reasonable to impose on the embryo, we
must ask what the alternative is. In the embryo's case, the
alternative is nonexistence. A 20% chance of a good life is better than
nonexistence. So, the imposed risk, if there is one, is reasonable.
Consider the sick child analogy. Imagine your child is certain to die
of a disease in a remote village in Africa (option A). Instead of
certain death, you could undertake a journey in which there is a
four-fifths chance of being killed by natives. But if you make it to
the city, there is a hospital where your child can be cured (option B).
Option B is better than option A. Taking the journey imposes an 80%
chance of being killed, but this risk is reasonable compared to certain
death.
If the risk is reasonable when the alternative is death, then it is
also reasonable when the alternative is nonexistence. Some might object
to this. Consider the genetic disorder analogy. Both members of a
couple have a rare genetic disorder such that any child of theirs is
certain to be born with a genetic condition. The child will develop in
utero, but this condition has an 80% mortality immediately after birth.
However, if the child survives, it will likely be normal.
Some people believe that it is wrong to knowingly conceive a child
with a high chance of having a severe genetic condition. Before the
advent of prenatal diagnosis, some couples, having discovered they are
at risk of having a child with a genetic disorder, chose to refrain
from having more children. These couples believed it was better not to
have a child than to have a child with a severe genetic disorder.
It is not clear whether these couples were rational or right in
choosing to remain childless rather than risk having a child with a
disorder. However, even if it is permissible, right, or rational to do
as these couples did, this does not bear on the genetic disorder
analogy or natural reproduction for two reasons. First, they may have
been right because it was in their interests not to have a child with
such a disorder. However, if a couple wish to have a child with genetic
disorder, it may be in their interests to have such a child. Second,
the child in these scenarios was likely to live with a severe genetic
condition and severely impaired quality of life. In both the genetic
disorder analogy and natural reproduction, the condition is possibly
lethal but does not affect quality of life if the human survives. It
may be wrong to bring into existence a child who will suffer
enormously. It is not clearly wrong to bring into existence a child who
merely has a low chance of surviving but will likely otherwise have a
very good life.
Natural reproduction gives every embryo the greatest chance it could
have of becoming a baby. And that baby has a high chance of having a
good life. The principle that drives intuitions about the wrongness of
mobile phone use in Harris's example does not bear on natural
reproduction at all. Although the risk of an embryo dying in natural
reproduction is high, perhaps as high as 80%, it is the lowest it could
be.
This is quite different from embryo research. Embryo research imposes
new risks on embryos either created as part of infertility programs or
specifically for research. I will next examine whether this risk is
reasonable or unreasonable.
Instrumentalization
Harris claims that Catholics and prolife advocates “are
treating the 1–4 embryos that must be sacrificed in natural
reproduction as a conscious (though not intended) means to have a live
birth” (p. 364).
This may be so. It is important to distinguish, however, between
being treated as a means and as a mere means (as Harris himself has
done in many of his other articles). Kant's dictum is that one
should treat human beings always as an end, never merely as a
means. It is only the latter that Kantians object to and that
represents instrumentalization.
Here is the difference between being treated as a means (and also as
an end) and being treated as a mere means:
- A couple want to have a child, so they create an embryo. They do
everything they can to help that embryo become a baby, but, alas, it
dies from natural causes. They treated it as a means to having a child
but also did everything they could to help it to survive, and so also
treated it as an end in itself.
- A couple have a sick child who will die without a bone marrow
transplant. So, they create an embryo for the purpose of destroying it
for its stem cells. This treats the embryo as a mere means to securing
the health of their existing child (assuming, as I have been, that the
embryo has rights and interests).
Creating embryos by natural reproduction does not instrumentalize
them as mere means even if they are highly likely to die.
Harris claims, “Given that decisions to attempt to have
children using sexual reproduction as the method (or even decisions to
have unprotected intercourse) inevitably create embryos that must die
…” (p. 363). This is false. Some people fall pregnant the
first time they try to have a child. These acts of sexual reproduction
do not create embryos that must die. So, it is not inevitable
that engaging in sexual reproduction results in embryos that must die,
even if it is highly likely.
Harris gives an example of what he believes is instrumentalization:
A woman has two fertilized eggs and is told it is certain that
if she implants both only one will survive but that if she implants
only one it will not survive. Would she be wrong to implant two embryos
to ensure a successful singleton pregnancy? (p. 364)
This is not instrumentalization. What is characteristic about
instrumentalization is that it reduces the chance of an embryo
surviving for some other purpose. In this case, however, there is no
such reduction in risk. Both embryos have the greatest chance of
surviving that they could have. Each is in this way treated as an end
in itself, even if this is a means to increasing the chances of a
couple having a baby.
Harris gives another example of instrumentalization:
Suppose that for some biological reason there was a condition
that required that, for one embryo to implant, it was necessary to
introduce a companion embryo that would not, and we could tell in
advance which would be which. (p. 365)
This is not instrumentalization, either. Again, each embryo has the
greatest chance it could have of surviving. The fact that there is
another benefit—increasing the chance of another embryo surviving
or the couple having a baby—does not compromise this.
To put the issue in the plainest possible terms, we need to
distinguish two ways in which embryos perish in order that a live child
is born:
- Each embryo has a one in five chance of becoming a child.
- An embryo, A, would not form a child normally. However, if another
four embryos, B–E, were destroyed and cells taken from these and
inserted into the first, then the first would definitely survive. (This
mirrors the situation of multifetal reduction in which a woman has
quintuplets. All five babies are likely to die. But if four are killed,
the remaining one has a high chance of surviving.)
The first does not involve instrumentalization; the second does. But
the second is not what happens in nature.
Consider a case of true instrumentalization. Imagine that embryos
have a 20% chance of surviving. Imagine that, by splitting an embryo,
we give each embryo a 15% chance of becoming a baby. If both are
implanted, the chance of having at least one live born child is 27.75%.
This increases the chance of having a baby. But A is destroyed and
replaced by two qualitatively identical embryos B and C, each of which
has a lower individual chance of producing a baby. This
instrumentalizes embryo A for the couple's purpose of having a
baby. Whether it is wrong is another question. But natural reproduction
does not involve any such instrumentalization.
Harris frequently uses the term “sacrifice”—but no
embryos are sacrificed in natural reproduction. Each gets the greatest
chance it could have of living. Sacrifice implies
instrumentalization—sacrifice for some purpose.
“Both natural procreation and ART involve a process in which
embryos, additional to those that will actually become children, are
created only to die” (p. 362). “Only to die” here is
ambiguous—it could mean “but do die.” Or it could be
used in the sense “only for the purpose of dying.” Talk of
embryos dying for other embryos and sacrifice suggests the latter
interpretation, which is false. In the case of natural reproduction,
these embryos are not created for the purposes of giving other embryos
a greater chance of life—every embryo gets the best and equal
chance of being a baby.
Harris's claim that embryos in assisted reproduction are
instrumentalized (as mere means) is also not entirely correct. In
assisted reproduction, a batch of embryos is created at one time. Then
each embryo in that batch has some chance of surviving, albeit lower
than if fewer embryos had been created or if it had be conceived
naturally. However, it is not created only to die. It is a lottery that
embryos do die.
Existing versus New Lives
Let us return to Harris's central argument. Recall that, for
Harris, if
- an act has the foreseeable consequences of bringing into existence
a new life and causing the death of four embryos is permissible,
then
- an act that has the foreseeable consequences of saving an existing
life and causing the death of four embryos is permissible.
This argument fails because Harris fails to consider the alternative
courses of action. Comparing saving with bringing into existence:
- We have as much reason to save existing lives as we have to bring
new lives into existence.
- It is reasonable to act so that four embryos die to bring one new
life into existence, when the alternative is having no children at
all.
- It is reasonable to act so that four embryos die to save one life,
when the alternative is that none live.
- It is reasonable to act so that four embryos die to save one life,
when the alternative is that one person dies and each of four embryos
has a 20% chance of surviving.
The third may follow from the first and second. However, the fourth
does not follow from the first and second or from the first, second,
and third.
Consider another thought experiment. Imagine that infant mortality is
very high—80% of infants born die of natural causes.
Harris's argument would also apply to this case, if we accept a
prolife view of the embryo's moral status:
- Natural reproduction entails that it is acceptable for four infants
to die to bring a new life into existence. (P1)
- There is at least as strong a moral reason to save existing lives
as there is to bring new lives into existence. (P2)
- If it is permissible for four infants to die to bring one new life
into existence, then it is permissible for four infants to die to save
one existing life. (conclusion)
It would not be permissible to kill four healthy infants to save one
existing life—so the argument is invalid. Harris has assumed, for
argument's sake, that the embryo has a right to life. If that is
so, the healthy embryo or infant has the same right to life as anyone
else. So, it cannot be right to kill embryos for research.
Alternatively, Harris has merely assumed that the embryo does not have
a full moral status, in which case he is begging the question (or
assuming what he set out to prove). In that case, his argument would
prove nothing.
Embryonic Stem Cell Research on Spare
Embryos
I am told that men are often blind to the faults of a beautiful
woman. Harris is overly attracted by the sexiness of the analogy with
natural reproduction. However, Harris's approach of appealing to
consistency can be resuscitated and applied to spare embryos.
What are spare embryos?
Ten percent of couples are infertile, and IVF has offered these
people the opportunity to have a child. It is an accepted part of IVF
that more embryos are produced than a couple may need. These embryos
are then frozen to reduce risks to the woman of numerous attempts to
obtain eggs from her body, which can cause illness or death. Prior to
embryo freezing, several embryos are transferred to the woman at one
time, which leads to the high rates of mortality and disability
associated with triplets or quadruplets. The consequences of not
freezing embryos would be that women would be harmed and children born
by IVF would be worse off.
That is why we have about 70,000 frozen embryos in Australia and more
in countries like the United Kingdom.
Society has made the decision that it is acceptable to reduce the
risk to existing women by creating excess embryos. If embryos can be
created and destroyed to reduce the risks to women who are infertile,
then it would be reasonable to create and destroy embryos to reduce the
risk to the health of other people—that is, those who may benefit
from stem cell therapies.
If we accept IVF and the production of spare embryos as a part of
infertility treatment, we should accept the creation of new embryos as
part of the research into embryonic stem cells.
Embryo Research and Unreasonable Risk
What are the true lessons from natural reproduction? The main lesson
is that imposing risks on the embryo can be acceptable if it is the
best strategy to improve the chances of that embryo having a healthy,
happy life as a person. I have argued elsewhere that in certain
conditions, some destructive embryo research is the best strategy to
maximize the chances of a population of embryos surviving to live a
reasonable life.
Does destructive research impose unreasonable risk on the embryo?
There are two ways in which we can avoid violating a principle of
nonmaleficence. The first is to argue that the embryo is not the kind
of being that can be harmed by being killed. This is to deny that the
embryo has a moral status.
The second is to argue that the embryo that is killed has not been
exposed to an unreasonable risk. I have argued, after Harris,2
Harris J. Survival lottery. In: Harris J, ed.
Bioethics. New York: Oxford University Press;
2001:300–15.
that if the embryo is a part of a
population of embryos that will benefit from the practice of killing
some, then killing may not be wrong.
3 Savulescu J. The embryonic stem cell lottery and the
cannibalization of human beings. Bioethics
2002;16(6):508–29.
Whether it is wrong turns on the
fraction of ART embryos that are killed and what the benefits are to
that population. If stem cell therapies will benefit that population in
significant ways, then this may be enough to justify killing some.
Unreasonable Risks
I have so far been considering what constitutes a reasonable risk to
impose on an embryo from the embryo's own perspective (assuming it
has one). In fact, society imposes very unreasonable risks on embryos
all the time. All preimplantation genetic diagnosis and virtually all
prenatal testing work against the interests of the embryo tested. Not
only is there some risk of harm to the embryo directly, but the only
decision that can be made on the basis of these tests that will affect
the outcome of the pregnancy is to kill the embryo. Genetic testing is
not done for the embryo's sake but for the couple's sake.
Indeed, couples are, in practice, allowed to authorize the killing of
embryos and fetuses for many reasons. The reason these practices are
permissible is that the embryo has no perspective, and no rights or
interests, because it is not a person. Not only are we permitted to
expose embryos to reasonable risks, but, if we are consistent with
other social practices, we can expose embryos to any risk in
research, provided they are not allowed to live on in a damaged state.
If we are consistent, as Harris urges us to be, any kind of destructive
embryo research—including creating new embryos by IVF or
cloning—is permissible.
