Gendler, T. S. (1998). Galileo and the Indispensability of Scientific Thought Experiment. The British Journal For the Philosophy of Science, 49(3), 397–424.

Summary

In this paper, Gendler argues that thought experiments are important and justified for the use of scientific inquiry, using Galileo’s thought experiment demonstrating that two objects with different masses fall at the same rate. The thought experiment goes, imagine that two objects with different weights are strapped together. Because the lighter object falls slower than the heavier object, it must slow the heavier object down, and thus the speed that they fall together must be somewhere in between the speed of the heavy object alone and the light object alone. But, together, they also have a greater mass, meaning that together they should fall faster than the heavy object alone. Thus, for there to not be a contradiction, the objects must actually fall at the same speed.

Gendler addresses the claim that thought experiments are just another form of deductive/inductive reasoning from a set of explicit premises, and explains what they add over and above pure argumentation. Specifically, she argues against the Elimination Thesis:

The Elimination Thesis: Any conclusion reached by a (successful) scientific thought experiment will also be demonstrable by a non-thought-experimental argument. (pg. 398)

She breaks the Elimination Thesis down into two separate claims:

The Dispensibility Thesis: Any good scientific thought experiment can be replaced, without loss of demonstrative force, by a non-thought-experimental argument. (pg. 401)

and

The Derivativity Thesis: The justificatory force of any good scientific thought experiment can only be explained by the fact that it can be replaced, without loss of demonstrative force, by a non-thought-experimental argument. (pg. 401)

First, Gendler rephrases the thought experiment in terms of a few propositions:

(1) Natural speed is mediative [averaging].
(2) Weight is additive.
(3) Natural speed is not directly proportional to weight.

Next, she goes about disproving the Dispensibility Thesis by giving an example set of alternate premises that an Aristotelian might adhere to, given their existing belief that objects fall at the same weight:

(4) Natural speed is not physically determinate for strapped-bodies.
(5) Weight is not physically determinate for strapped-bodies.
(6) Natural speed and weight are mediative for strapped-bodies that are united (two objects). Natural speed and weight are additive for strapped-bodies that are unified (one object).
(7) Natural speed and weight for strapped-bodies are determined by a degree of connectedness ($C$) such that the speed/weight of $B_1$-strapped-to-$B_2$ where $B_1$ has $w_1$ and $B_2$ has $w_2$ will be: $C(w_1+w_2)+(1-C)(w_1+w_2)/2$

That is, the Aristotelian takes this “degree of connectedness” as being a key relevant property, and thus is able to preserve the idea that speed and weight are proportional. Of course, this claim seems a bit ridiculous, and that is because, as Gendler claims, the thought experiment is revealing to us tacit assumptions that we did not realize we had:

(8) Natural speed and weight are physically determined.
(9) Entification [number of objects/entities that something counts as] is not physically determined.

However, it is important to note that “prior to the thought experiment, the Aristotelian is explicitly committed to the negation of (3), and this background commitment serves as a filter through which apparently contrary evidence will inevitably be reinterpreted” (pg. 409). Thus, it does not necessarily follow that anyone given (1) and (2) would reach the conclusion of (3)—particularly if they already thought (3) was false, they would look for ways of explaining away or disproving (1) and (2) so that (3) would remain false.

Next, Gendler moves on to disproving the Derivativity Thesis by showing that through the thought experiment knowledge has been gained, both in the sense that something new has been learned, and that this new knowledge is justified.

The first part of this argument (that something new has been learned) is a straightforward claim: presumably, altering one’s beliefs so that they negate a piece of knowledge counts as having learned something “new”, as it is not simply a combination of previous beliefs. But even more importantly, the Aristotelian is led to think of speed as a different type of concept entirely: rather than being a derivative of weight, it is something else. As Gendler puts it, “it brings the Aristotelian to recognize the inadequacy of his conceptual framework for dealing with phenomena which—through the contemplation of this imaginary case—he comes to recognize as always having been part of his world.” (pg. 412)

The second part of the argument (that the knowledge is justified) is trickier. Gendler asks, “Why should we think that our pre-theoretical beliefs about the structure of the physical world are reliable?” (pg. 414). She doesn’t really give an answer to this question, but does argue that we do have (implied veridical) knowledge of the world, but it is not accessible by argument alone: it requires something like a thought experiment to tap into it.

Takeaways

This is a somewhat different take on the notion of simulation than what I’ve been thinking about so far. Although Gendler doesn’t refer to thought experiments as simulations per se, they are related in the sense of being a mental reproduction of something in the world.

I have more thoughts about this, but I’m too tired tonight to get them in a coherent enough form to write down. I will include them on my notes for the other papers in this topic (thought experiments) tomorrow.