Monday, February 19, 2018

Fetuses don’t play the violin

Abortion is one of the most controversial topics in America today, but among philosophers what is not controversial is that Judith Jarvis Thomson’s “A Defense of Abortion” is one of the most important and influential articles on the topic. Anthologized countless times, whichever position you take in the abortion debate you cannot claim to have an informed opinion without considering Thomson. Principally driven by her creative and memorable ‘violinist thought experiment,’ Thomson changed the game by granting the central pro-life assumption—namely that the fetus is a person with full moral status—and then arguing that abortion is permissible anyway.

For the uninitiated, Thomson asks us to imagine ourselves kidnapped by the ‘society for music lovers’ and hooked up the titular (innocent, unconscious) ‘violinist’ who is using our kidneys to stay alive. If you unhook yourself the violinist will die. Does the violinist’s ‘right to life’ oblige you to stay hooked up to him for 9 months? Thomson says no: even if the violinist/fetus is a person, no person has the right to use another person’s body against their will.

My concern here is not the conclusion itself as much as it is the methodology Thomson employs. My skepticism towards thought experiments has been a theme for my posts here at The Dance of Reason (see here and here) and Thomson’s are no exception. The violinist was the first of several thought experiments Thomson employed in her article, and part of what makes it such an instructive example of philosophical methodology gone wrong is how seamlessly Thomson crafts each new experiment to counter a successive series of objections. Thomson asks us to imagine diseases that can be cured by the touch of a stranger’s hand; a baby expanding like a balloon and crushing a person to death; humans procreating via ‘people seeds’ that implant in unsuspecting stranger’s carpets. Each thought experiment has a rationally comprehensible connection to the objection that comes before it, yet a common response to this whirlwind of imaginative scenarios is that we’ve lost sight of something crucially important: the biological realities of sex, pregnancy, gestation and birth.

Thomson’s experiments are an excellent example of what happens when philosophers find themselves accountable to nothing other than their imaginations. The relationship between philosophy and science fiction is so strong in no small part because both thrive on imagining what happens when we push past our current limitations. But just as bad science fiction is often born out of concepts that outstrip their connect to the world in which we live, bad philosophy is often born out of imagination untethered to the subject attempting to be illuminated.

To be blunt: I have no idea at all what morality demands of people in a world in which human beings procreate via people seeds that gestate in carpets. Moreover, I’m not sure anyone else does, either. What kind of moral values would be most important if babies could expand rapidly enough to crush a person in a house? How much stock should we put in our intuitions about whether or not it is permissible to unhook ourselves from a stranger using our kidneys to filter his blood? That’s not how biology works, that’s not how ethics works, that’s not how any of this works.

Our moral intuitions are delicate things. They are rooted in our evolutionary history, honed by our cultural upbringing and exist to help us navigate moral decisions in the actual world. When you strip the biological and cultural context out and throw those intuitions into fantastical scenarios they quickly loose perspective. Such ‘intuition pumps’ are easily abused in the hands of a creative philosopher who wants to lead their audience by the nose to their preordained conclusion.

None of this is to say that Thomson’s position is incorrect. To the contrary, I think her basic conclusion is true: regardless of their moral status, fetuses have no right to use a woman’s body against her will. And I understand the impulse to invoke a thought experiment, as pregnancy is such a biologically unique phenomena that it’s hard to find ways to triangulate our intuitions about it. But we don’t need to appeal to fantastic, intuition-pumping hypotheticals to establish Thomson’s point. There are a set of rare, but real-world cases that suffice, namely the phenomena of ‘sleepwalking rape.’

The broader phenomena of ‘sexsomnia’ (people initiating sex while neurologically asleep) is well documented by psychologists. There are several cases in which men accused of rape have, with the help of expert clinical testimony, successfully exonerated themselves in court by arguing they were asleep at the time of the attack.[1]

So let’s ask the following question: would a person being raped by a sleepwalking rapist be justified in killing their rapist in self defense? Like both Thomson’s violinist and a fetus[2] the sleepwalking rapist is both a person with full moral status and innocent of any crime. If a woman killed her sleepwalking rapist would we call that an unjustified killing? I think people could have an honest disagreement here, but I suspect that most people (even pro-life people) will say that it would be justified.

The exact implications this has for abortion is, of course, open ended. There are all sorts of ripostes that a pro-life person could give in an attempt to allow killing a sleepwalking rapist while still maintaining that abortion is wrong. I don’t imagine any single argument can settle this highly controversial issue once and for all. My point here is simply to say that philosophers would be well advised to avoid highly unrealistic thought experiments when actual cases from the real world work just as well, if not better.

[1] And of course, others attempting that defense have failed and been sentenced accordingly.
[2] Like Thomson I am assuming for the sake of argument here that the fetus is a person in the moral sense with full moral status, even though also like Thomson I do not believe this to be the case.

Garret Merriam
Philosophy Department
Sacramento State


  1. Thank you for this, Garret. Too often I question the validity of thought experiments that take us out of realistic context. While I am a fan of fantasy and sci-fi, unrelateable metaphors seem to do philosophy a great disservice in giving us food for thought. Achilles never being able to pass the damned tortoise is one that comes to mind because most of us can't get past the fact that it is not something that would actually happen in reality.

    1. Thank you for saying so, Kris. I'm glad you agree. When I was younger I really loved thought experiments and took them way too far. So this is kind of a course correction for me.

  2. I generally agree with you, Garret, thought experiments have very little bearing on how we ought to act in the real world. When you stipulate away certain features of real life, any conclusion from these thought experiments would apply only in a parallel world where the stipulated facts are true. Not in the real world. I and many others before me have had similar suspicions about Rawls’s original position, a purely hypothetical situation in which rational agents are under a veil of ignorance that prevents them from knowing the circumstances of their lives, such as their religion and ethnicity, and then deliberate on the basic principles of justice. There is much to praise about Rawls’s principles of justice as fairness, but one wonders whether this conception of justice is for an ideal community of rational automatons, not the real world with historical and structural injustices, wide disparities of power and resources, and people with deep cultural commitments (not all of which are consistent with justice as fairness) and, as a result, deep disagreements on almost every political issue. So I generally agree; many philosophers are guilty of this sort of armchair philosophy.

    But I wonder whether the problem is not the use of real or imaginary examples, but the use of relevant and irrelevant ones. Relevance usually refers to a similarity relation, something to this effect:

    X is relevant to Y if X and Y have similar features s1, s2,…sn.

    Relevance is also a degreed relation, such that the more similar features X and Y share, the more relevant X is to Y.

    If the problem has more to do with relevance, rather than whether the example is real or hypothetical, your example is no better than the famous unconscious violinist. Some of the same objections raised against the famous unconscious violinist would apply to your example of a sleepwalking rapist.

    For example, the sleepwalking rapist lacks the requisite mens rea for the crime, but he has still done something wrong. Rather than the team of unconscious violinist and the Society of Music Lovers, you have a single agent, the sleepwalking rapist. Between the sleepwalking rapist and the victim, the rapist has done something wrong (and should bear the costs associated with avoiding the risks of such wrongs). He attacked and violated the victim. One cannot similarly say of the unborn child that he has done something wrong; he attacked and violated the mother.

    What the unconscious violinist has over your sleepwalking rapist are the features of being biologically connected, needing the host’s biological resources, having done nothing wrong (this is true of at least half the team of unconscious violinist/Society of Music Lovers), and being subject to impending death upon being disconnected. The imaginary example has more similar features in common with the pregnancy situation. I think these similar features make it powerfully effective in challenging our intuitions. It is far from perfect, but if relevance is what matters here, an unconscious violinist may be better than a sleepwalking rapist.

    I’m not saying that the unconscious violinist is a good thought experiment. Relevance probably more precisely should include a further qualification.

    X is relevant to Y if X and Y have similar features s1, s2,…sn that are not undermined by important differences d1, d2,…dn.

    I’m just saying that a sleepwalking rapist may not be a better example because it seems less relevant.

    1. Thanks for the comment, Chong. I like the idea of relevance in the abstract, but I'm concerned about how we judge what is and isn't relevant. I think Thomson would argue that her thought experiments capture many morally relevant features, and I sort of agree; the problem is that there's a lot of morally irrelevant features, 'noise' if you will, and it can be really hard to parse the signal from the noise. As a general rule, I think real-world case studies are more likely to be relevant in virtue of the fact that the real world imposes many of the same restrictions across cases that simply aren't present in imaginary cases. That's not to say that all real-world cases are relevant to all other real-world cases, but ceteris paribus, I trust real-world case studies more than imagined ones.

      To your comments on the sleepwalking rapist: I find it odd to say 'the sleepwalking rapist attacked and violated the victim.' That phrasing seems to imply awareness, if not intent. If I'm on top of a building and a strong gust of wind throws me off the roof and I land on Russell, killing him, it would strike me as very strange to say 'Garret killed Russell.' I am a 'victim' of forces outside of my control. Ditto for the sleepwalking rapist. If we assume he wasn't aware of this problem (obviously if he knew he might do something like this he would be required to take proactive measures to stop it) then I don't think it makes sense to say 'HE attacked her.' (You might say something like 'His body attacked her' or some such, but things can get messy going in that direction.)

      One way or another, pregnancy is so sui generis that it it hard to make analogies to it. Maybe we should throw out all such comparisons and just deal with abortion head on.