On the Ethics of Imaginary Research

John D. Norton*

Department of History and Philosophy of Science
Center for Philosophy of Science
University of Pittsburgh

The purpose of this note is to draw attention to hitherto overlooked ethical problems raised by the now widespread techniques of imaginary research. Three principles are proposed as ethical guides for future research practice.

*April 1, 2010. I am grateful to my imaginary friend for helpful if one-sided discussion during the writing of this note.



Real experimental practice has long been controlled by multiple regulations, designed to prevent a wide range of improper experimental procedures.1 They ensure the dignified treatment of human subjects and the humane treatment of experimental animals. Researchers conducting real experiments routinely submit their protocols for the approval of IRB's, Internal Review Boards, and research establishments maintain appropriate registration.

Imaginary experimentation and research has so far been exempt from similar regulatory attention. This lapse results from the dismissal as merely imaginary of the harms that may arise from imaginary experimentation. In the broader philosophy of science and philosophy of mathematics literature, realism remains a hotly contested viewpoint. Yet the realist prejudice that strictly partitions everything into real and imaginary is all that sustains this dismissal of imaginary harms.

1See for example, D. Harms, "The Belmont Report: Ethical Principles and Guidelines for the Protection of Human Subjects of Research," http://www.hhs.gov/ohrp/humansubjects/guidance/belmont.htm


It has long been apparent that a more complex analysis of the real and imaginary is needed. What has prevented a full recognition of the scope of the imaginary harm wrought by imaginary experimentation is a lack of adequate documentation. One can scarcely doubt the irresponsibility of Galileo's Sagredo when he cast canon balls "weighing one or two hundred pounds or more" from the heights.2

Since no reports are given, we can only imagine the peril this experiment brought to innocent passers-by. Indeed we can now see the utter moral vacuity of Salviati in his excoriation of Aristotle for not imagining just such an experiment.

One may argue that the vague description of the location of Sagredo's experiment means, at best, that no one person specifically was endangered, or, at worst, that the dangers were widespread and thus diluted. Unfortunately later implementations of Sagredo's experiment have been more specific in their placement and thereby localized the imaginary harms more narrowly.

2G. Galilei, Dialogues Concerning Two New Sciences. New York: Dover, 1954, p.62.



The distinctive leaning tower of the figure and the period dress locate the experiment in sixteenth century Pisa, just as it illustrates the willful negligence of the experimenter.

One would like to think that the tradition of thought experimentation initiated by Galileo evolved towards more benign experiments in the hands of later, more enlightened practitioners. Unfortunately, with greater theoretical power comes the ability to wreak greater harm. We need only look to Newton's famous imaginary experiment of a stone projected from a mountain into dangerously low earth orbits.3

Newton's mountain

The figure is too coarse for us to see the dangers to inhabitants of many continents. We can only imagine them.

3I. Newton, Mathematical Principles of Natural Philosophy: Vol. II: The System of the World. Berkely: University of California Press, 1962, p. 551.


This reckless experimental practice must submit to the discipline of principles and regulation. In the following, I will articulate three ethical principles to which, I urge, all practitioners of imaginary research must be held.

First Principle

Non-existence does not deprive a living being of rights.

We have long accorded rights to non-existents. A familiar example arises in the case of someone discharging a firearm in a crowd. Even if the discharge harms no real person, the perpetrator is still guilty of gross negligence and is subject to prosecution and punishment. We justify this treatment by imagining what could have happened. What if someone were in the line of fire? The outcome could have been deadly.

That is, we are quite willing to exact severe legal penalties in defense of the rights of an imaginary person whose harms are only imagined.

A related matter that will not be pursued here in the detail it deserves is the mistreatment of animals. It is, sadly, routine in the practice of imaginary research. Only the most egregious of cases have received popular attention. Who has not been horrified at the sadistic torture by Erwin Schoedinger of his cat? Yet it is tolerated merely because the poor beast has the additional misfortune of non-existence.

Schroedinger cat


Second Principle

All imaginary experimentation must be subject to cost-benefit analysis.

This principle is no different from the demand placed on real experimentation. Medical research that proves lethal to an experimental subject may yield results that save future lives. Yet the cost of the lives assuredly lost is universally judged greater than the benefit.

In the case of thought experimentation, one must always ask if the experimental goals might be achieved better by a real experiment. One might think that an imaginary experiment is always to be preferred to a real one in a cost benefit analysis. For the real harms brought by a real experiment are most likely vastly worse than the imaginary harms of an imaginary experiment.

While that relative assessment of the magnitude of possible harms of real and imagined experiments is likely correct, it does not take into account the frequency with which the experiments are conducted. Real experiments are rarely conducted more than once or twice; they are too hard and too expensive. An experiment in thought, however, once it is created, can be replicated at will. If it is well-crafted, just this is likely to happen.

Even if the harm wrought with each replication of the imaginary experiment is minuscule, the accumulated harm of very many repetitions can readily exceed the harm of a single execution of a real experiment, especially if the latter is well-controlled.

If it is determined that a thought experiment is unavoidable, a corollary is that the least harmful design should be sought, compatible with the successful functioning of the experiment. Schroedinger did not need to torture a cat to achieve his result in quantum theory. A mouse or even a cockroach would have sufficed. But an amoeba would not.


Third Principle

Thought experimentation must proceed under informed consent.

The author of a thought experiment carries a special burden not placed on creators of real experiments. For all readers become the author's research assistants, actively participating in the execution of the experiment. Imaginary scenarios can be created in text quite rapidly and so fast that the reader can become an unwitting accomplice in the creation of imaginary harms. What if you are a brutal dictator of a small nation who oversees the mild reprimand of a few innocent citizens. The harm has been imagined before the reader has a chance to consent to participation in its creation.

In this case, the damage was small. It was merely misplaced reprimands. But what if it had been a genocide?

More difficult issues are raised by the choice of language. What if a thought experiment involving great harm is written in a language a reader does not know. Do we discount the harm merely because the reader did not understand the language read? We are loath to issue complete exculpation for speech acts on these grounds. A person inciting a mob to violence by calling "kill him" is not excused if we learn that the person calling does not understand English. Should it be any different with imaginary experiments? What if the experiment is read aloud?



One may doubt whether these three principles will survive longer term philosophical scrutiny. That remains to be seen. However one should not allow those doubts concerning the future to prevent action in the present. As long as we are unsure of the severity and extent of the harms created by imaginary research, we should act prudently to limit the possible harms. We cannot assume that realists' hesitations concerning imaginary harms will prevail over the well-reasoned fears of the anti-realists. We cannot play "wait and see," while risking the multiplication of harms that may be irreparable.

How does one heal the concussion of an imaginary passer-by or the poisoning of a fictitious cat? Is the cat I bring back to life in my imaginary medical experiment the same one that died in your experiment? Or have my efforts to revive your casualty merely duplicated the harm?

As long as these doubts remain, I call upon universities and other research centers to implement Internal Review Boards for imaginary experimentation and that these IRBs be commissioned to hold their researchers to the principles enunciated here.