“If the Third World War is fought with nuclear weapons,
the fourth will be fought with bows and arrows.
”
Lord Louis Mountbatten
There are mainly two important problems in the philosophical discussions of thought experiments where intuitions are involved: First, there is the problem of the informativeness of thought experiments. Thought experiments provide us with new information. But where does such information come from? The problem of the informativeness of thought experiments is most controversially discussed in the debate between James Robert Brown and John D. Norton (p. 89)
Brown’s platonic account holds that thought experimentation has access to a specific type of intuition which is independent of – and irreducible to – empirical observation. These intuitions are about some a priori realm of universals. In contrast to this view, Norton holds that thought experiments are (“mere”) deductive or inductive arguments. Following this path, one should say that thought experiments are informative in the same way arguments are, that is, by making use of empirical data and inference. In this debate, Daniel Dennett stresses that, whatever thought experiments are, they can sometimes “lead to a quick and uncritical jump to a conclusion that is not really warranted” (p. 90).
Brendel is explicitly on Norton’s side and in this paper he examines the legitimacy of the method of thought experiments.
What are thought experiments?
Brendel locates the first use of the expression “thought experiments” in a paper by Hans Christian Ørsted (written in 1811). Nevertheless, he acknowledges Ernst Mach as the first man of science to have investigated this type of experimentation. According to Mach, thought experiments have a “propaedeutic function”, they are some sort of prequels to actual experimentation. Brendel agrees with this conception but stresses that thought experiments are, first and foremost, experiments and we must view them as such because of the characteristics they share with empirical experiments: (1) the study of functional dependency of variables by planned and controlled data, (2) the dependency of the practice on some background theoretical assumptions.
Counterfactuality, Brendel argues, is not a necessary condition for thought experiments. Granted, some thought experiments are such that they cannot, as far as technology stands, be put into practice (Einstein’s observer traveling alongside a beam of light, for instance). But this is not always the case. Many thought experiments can but need not be put into practice. Both the former and the latter are, in the same sense, thought experiments, not empirical ones.
Brendel identifies several functions of thought experiments:
(a) “alethic refuters” their function is to give birth to paradoxes (acc. to Sorensen); or track down hidden contradictions/inconsistencies (acc. to Kuhn), by imagining new and unusual situations, we create awareness around the criteria that govern the use of concepts
(b) thought experiments can be used to provide evidence in support of a questionable theory (Newton’s bucket experiment shows the existence of absolute space)
(c) pedagogical function of illustrating a certain theory/philosophy
(d) detect vagueness and “unstable intuitions” & then explicate
The problem of informativeness
Brown’s Platonic account in The Laboratory of Mind starts from the following definition:
A Platonic thought experiment is a single thought experiment which destroys an old or existing theory and simultaneously generates a new one; it is a priori in that it is not based on new empirical evidence nor is it merely logically derived from old data; and it is an advance in that the resulting theory is better than the predecessor theory.
Brown’s favourite example was Galileo’s “falling balls” experiment. Brendel rejects Brown’s account of the experiment because: (1) “the supposedly intuitive grasp of the conclusion that all bodies fall at the same speed does not at all follow immediately and with no help of other premises from the demonstrated contradiction”, (2) as opposed to the argument-view of thought experiments, “it remains unclear when and why an intuitive grasp of the abstract realm can go wrong”. Brendel refers to Norton’s reconstruction of Galileo’s experiment as a reduction ad absurdum argument and claims that there is no a priori argumentative gap involved.
Therefore, we should reject the view that we gain new information from thought experiments by a special epistemic capacity for intuitively perceiving the laws of nature in a Platonic realm. (p. 95)
However, Brendel does not go all the way to the other (Norton’s) extreme to claim that thought experiments are arguments; that in conducting thought experimentation one conducts argumentation. “Instead, there are some a posteriori acquired “truths” that function as implicit background knowledge, enabling us to come to a relatively quick decision in the evaluation of a thought experiment. But we can always make these premises explicit by reconstructing the thought experiment as an argument” (p. 96). The reconstruction could detect overlooked premises or unwarranted assumptions – since intuition, just like sense perception, is fallible.
Thought experiments can be dismissed because they are based on implausible, incoherent or inconsistent premises or because they involve inconclusive judgements, illogical inferences or other kinds of argumentative shortcomings like a petitio principii. (p. 97)
But aside from their argumentative features, they also “appeal to intuitions”. It is whay Daniel Dennett called them “intuition pumps”.
As Brendel mentioned, thought experiments share with empirical ones the control of data. Thus, he argues, the lack of such control is – just as with normal experiments – a fault. He cites Putnam’s twin earth[1] experiment as an illegitimate intuition pump, guilty of such fault: “since Putnam does not give us any explanation as to why the twin earth is not radically different from our world” (p. 99).
A legitimacy problem is also involved in the counterfactual assumption, which needs to be irrelevant – i.e. it’s inapplicability at the time of the thought experiment, if at all, needs to be a mere incidental property. And it can be the case that one overlooks relevant impossibilities. Brendel illustrates this with two of Einstein’s famous thought experiments: the “traveller” and the “clock in a box”. In the first, since the assumption (i.e. a person travelling at the speed of light alongside a beam of light) is not denied or in any way undermined by Maxwell’s theory, Einstein was warranted to use it in a thought experiment. In the second, the assumption that a mechanism that measures time & energy at the same time is possible is not supported by the general theory of relativity (ironically so!), and therefore it is an unwarranted assumption. (pp. 101-103).
In other words, the legitimacy of the assumption rests in its consistency with the backdrop of the presupposed theoretical framework. (Of course, ideally, this backdrop is explicit, but in reality it is hardly the case. This could be why thought experiments, unlike arguments, need explanations themselves. Even if one understands every concept involved, in order to understand the use of that experiment, one must know the theories it uses and/or refutes).
When the “setting” – i.e. the theoretical knowledge assumed – is very unfamiliar, as it is the case with teletransportation, fission process, brain transplant etc, thought experiments are closer to stipulations rather than discoveries:
In such fictional situations our intuitions are of no great help, because it is very hard to decide in a non-question-begging way what the real criteria for personal identity are. That is why a lot of philosophers are very critical about the method of thought experiments in such applications. Wiggins, for example, points out that by “denaturing the human subject” in these thought experiments, we do not learn anything about the nature of personal identity because the decision whether the concept of identity is legitimately applied is more a question of stipulation and not a matter of discovery
Among such dubitable assumptions, there is also the one that only one notion describes such and such piece of reality (i.e. personal identity) in a coherent way, and that the task of the philosopher is to choose that one from the many competing ones.
So, how to avoid intuition pumps:
(1) make sure the thought experiment is not under-determined
(2) the jump from the particular (imaginary) situation to the general conclusion
(3) the counterfactuality of the initial assumption should be irrelevant
(4) far-fetched, sience-fiction cases should be reduced to ones where our intuitons can make a difference
[1] Putnam’s famous twin earth thought experiment is a typical example of such an intuition pump. Putnam describes the twin earth as a planet which is exactly identical to the earth we live in except for the fact that the liquid in the rivers, lakes and seas of twin earth has the chemical structure XYZ which is different from H2O. But nevertheless, in its surface structure this liquid cannot be distinguished from water on earth. It is further assumed that every person on earth has an exact “molecular copy” on twin earth. Our “twin earth Doppelgängers” also use the word “water” to refer to the liquid in their rivers, lakes etc. But they do not have any knowledge of the concept of H2O. Putnam now argues that although there is no relevant difference between the mental states of our “twin earth Doppelgängers” and our own, the reference of the word “water” is different. Therefore, reference is not determined by psychological states. “Meanings are not in our heads”.
0 comentarii:
Post a Comment