Nathan asked:

When we carry out a thought experiment, we can’t test the underlying philosophical hypothesis with any empirical data. So, besides logical flaws, what are the criteria for evaluating a philosophical hypothesis? And how can we benefit from thought experiments in our daily lives?

Answer by Massimo Pigliucci

Thought experiments are a type of argument, inviting us to consider the logical, and sometimes the empirical, implications of certain premises. They tend to invoke vivid images, and are therefore often memorable, but they are certainly not limited to philosophy, since scientists have also proposed a number of them, some of which have become famous instances of the genre.

Consider, for example, Galileo’s imagining what would happen if we dropped a heavy object from the height of a tower, tethered to a less heavy one. According to the then dominant Aristotelian physics, the heavier object – on its own – should fall more rapidly than the lighter one. But once the two are connected, Aristotelian physics predicts both that they should go faster (the combined objects are heavier than either of the two considered separately) and that they should go slower (the lighter one acting as a drag on the heavier one). This engenders a logical contradiction, showing that there must be something seriously wrong with Aristotelian physics (indeed, there was! And Galileo’s efforts paved the way for Newtonian mechanics).

Or think of Einstein’s famous thought experiment in which he imagined himself riding a wave of light, which led to his formulation of the theory of Special Relativity. And then, of course, there is quantum mechanics and Schrodinger’s cat, who is neither dead nor alive until an observer interferes with the experimental apparatus. And so forth.

While scientific thought experiments are meant to bring about insights ultimately concerning empirical matters (the behavior of falling bodies, the speed of light, or what matter does at very small scales), philosophical thought experiments are more concerned with conceptual issues, and are therefore constrained more by coherence than by empirical evidence. (This, incidentally, is a major distinction between science and philosophy.)

For instance, John Searle proposed the famous Chinese Room thought experiment, in which he invites us to reflect on what exactly produces meaning, suggesting that there must be more to it than the manipulation of symbols of which a computer – as currently understood – is capable. Then there is the Experience Machine thought up by Robert Nozick, meant to question the idea of ethical hedonism (if all that matters in life is pleasure, would it be all right to just spend all your existence inside a machine that gives you pleasure?). We also have Hilary Putnam’s ‘twin earth’ experiment, which is concerned with the relationship between meaning and reference in philosophy of language, and John Rawls’ idea of a ‘veil of ignorance’ characterizing an imaginary original position from where we get to debate how to structure a just society.

How do we test a philosophical thought experiment? We don’t, because they are not meant to be equivalent to scientific hypotheses. They are, as I said, arguments, that is examples of reasoning from certain premises to certain conclusions. Arguments can be accepted or rejected following two criteria: the soundness of their premises, and the validity of their structure. With respect to the first, a premise may be unsound, meaning that it is false. If so, even if the argument is well constructed (i.e., it is formally valid), its conclusions have to be rejected. With respect to the second, one can have an invalid argument (i.e., its form is incorrect) based on sound premises; also in this case the conclusion ought to be rejected.

Even when a thought experiment succeeds, because its premises are sound and its structure is valid, we may still not buy into its conclusion. That is because philosophical thought experiments are concerned with conceptual (or logical) coherence, and there often is more than one conceptually coherent way of describing reality. If so, only empirical data can tell us which of a number of logical scenarios actually happens to be true. (That said, it is still instructive to consider logically coherent alternatives, because they tell us how things could have been.)

As for relevance to daily life, that depends. Einstein’s thought experiment, or Schrodinger’s cat, are hardly going to be helpful in paying our bills or deciding whether to take a new job or not. But they are still crucial to enhance our understanding of the world. Some philosophical thought experiments are even more remote from daily preoccupations, for instance when they deal with issues of metaphysics. But a number of them concern very practical matters indeed, such as ethics, justice, and the sort of life we ultimately want to live.


Answer by Craig Skinner

I enjoyed Massimo Pigliucci’s stimulating and informative answer, and agree with nearly everything he says.

But I have one grumble.

As with nearly all commentators since Galileo, he gives Aristotle’s physics an unfairly bad press.

Speaking of Galileo’s thought experiment of the fall of a heavy object tethered to a lighter one, he says:

‘Aristotelian physics predicts both that they should go faster and… slower… a logical contradiction….something seriously wrong with Aristotelian physics’.

Not quite.

Aristotle’s physics is a coherent system of fluid mechanics within the then accepted frame of reference (movement in a system of concentric spheres of increasing density towards the centre). Thus, earth moves down in air and water; water moves down in air, air moves up in water, intermediately dense things move up in one medium and down in another eg wood falls in air, rises in water, rests on earth. A bigger object will fall faster than a smaller of equal density (and shape) because the smaller one’s greater surface/volume ratio produces more drag, impeding its velocity. He had a go at quantifying all this but maths was not his strong point.

And so, were Aristotle to return and hear of Galileo’s thought experiment, he would, on learning also that we had invented flying machines, invite us to simultaneously drop, from a height of 50 stadions, a big metal sphere and a smaller one. He predicts that the bigger one hits the ground first. And he is right. And if the two spheres were tethered together, but stayed separate objects, each subject to the drag effect of the air, the smaller one would indeed retard the bigger. Aristotle also predicts that an object falling through air or water will reach a maximum speed, and indeed things dropped from a great height do just this. It’s no good Galileo complaining that he meant falling in a vacuum. Aristotle simply replies that his physics was about the real world where things fall through air or water. Also, he would add, the idea of a vacuum is incoherent because, in it, a falling object would reach infinite speed. Right again — an object accelerating in vacuo in a gravitational field indefinitely would reach arbitrarily high speed. It’s just that in the empirical world, as opposed to the idealized situation, the object must ultimately hit the body which is gravitationally attracting it.

Aristotle’s physics lasted 2000 years, not because of dogma slowing science’s progress, but because it was a good theory, until Kepler, Galileo and Newton came up with better. Their framework, absolute space and time, in turn, eventually gave way to Einstein’s relativistic spacetime. And Einstein’s theory doesnt work below the Planck scale, and will give way to yet better theory, stringy, loopy or otherwise, in due course.

Aristotle has a place in the line of succession of great physicists. Just read Galileo’s Dialogues Concerning Two Sciences to see how deeply indebted Galileo is to him. And for a modern physicist’s view try Rovelli C (2014) Aristotle’s Physics: a Physicist’s Look. PhilSci Archive (on line at