Earlier this week I gave you a section of the current chapter on that Irvine and Mallory and their supposed ascent to the summit of SagarmÄthÄ. Here’s the most recent version for those who might want to compare and contrast.
When an epistemic agent, like you or I, infers to any old explanation rather than the best explanation available to them, then, I believe, we have a serious problem. All arguments in favour of “Just So” stories are fallacious; we need to be able to able to give reason as to why our particular inference to an explanation is warranted. We can understand this as being part of the parcel of issues that come with an understanding of the under-determination of theories (in this case explanatory hypotheses) suggested by the evidence. There will always be an infinite number of candidate explanatory hypotheses that will fit the evidence which can then be used to explain why some phenomenon or event occurred. Explanatory hypotheses provide a link between the particular facts and the specific event of phenomenon we want to explain the occurrence of and thus there will be a huge range of seemingly plausible explanatory hypotheses that an epistemic agent can choose from. It is clear, then, I think, that for something to be considered as a good explanation, then, it must be more than just a story that fits the facts.
Now, we can, I think, explicate the kind of problems that epistemic agents might face when inferring to an explanatory hypothesis. These are the standard kind of issues that people might easily be confused about. Epistemologists, like myself, can hopefully advise or adjudicate on such issues, as I am trying to do with respect to the vexatious issue of conspiracy theories in this thesis. Epistemic agents, when inferring to an explanation, might:
1. fail to consider some worthwhile contenders among the candidate explanatory hypotheses,
2. infer to an explanatory hypothesis that does not, in fact, entail or strongly suggest the phenomenon or event being explained,
3. infer to an explanatory hypothesis that is not the most plausible of the contenders, and
4. infer to an explanatory hypothesis that is not plausible enough to be regarded as likely.
These four problems cover the four most likely issues an epistemic agent will want to be able to provide some kind of answer to when she argues that her inference to an explanation is not merely an Inference to Any Old Explanation.
Here is an example.
George Leigh Mallory and Andrew Irvine made an attempt to reach the summit of SagarmÄthÄ (aka. Mt. Everest) on June the 8th, 1924ACE. Ever since, many historians and mountain-climbers have become obsessed with finding the remains of these two British mountain-climbers in the hope that amongst there personal effects will be found evidence that they successfully reached the summit before their demises, but this lack of evidence of a successful ascent has not stopped many of them from claiming that Irvine and Mallory did not die before making the summit but, instead, died after.
The only established fact of the matter is that Mallory and Irving were sighted at 1pm, several hundred metres from the summit. The various theories, which range from Mallory making the summit alone with Irvine’s last air-bottle to both Mallory and Irvine making the summit together, are, at best, intellectual fancies, given not only that there is no evidence they made the ascent but that the likelihood of their making the ascent, due to issues with the traverse they had chosen and the limited technology and resources they had available, was very low. Yet, some of the proponents of these theories present them not only as plausible candidate explanations for why it is that Irvine and Mallory never returned to their base camp, but as the actual explanation of their non-return.
The explanatory hypothesis that is used with respect to explain why Irvine and Mallory did not return to their base camp is something like the claim they successfully reached the summit and then died of oxygen deprivation and exposure on the descent. The arguments for the various explanatory hypotheses put forward for Irvine and Mallory’s demises are, at worst, Inferences to Any Old Explanation. Whilst we know that they did not return to the base camp we do not know that they reached the summit; we do not know what happened to them on the afternoon of June the 8th of 1924ACE. Given how little we know, there are a host of other plausible candidate explanatory hypotheses which account for their not returning to the base camp. Another, simpler and more probable explanatory hypothesis is that they died of oxygen deprivation and exposure during their attempt at the ascent to the summit. This is a much more worthwhile contender, as an explanatory hypothesis, than the more complex claim that they succeeded in their ascent and then died. Whilst the explanatory hypothesis does entail its conclusion it is not a particularly plausible explanatory hypothesis, given not only the technology of the time but also the fact that the mostly likely pathway Irvine and Mallory would have taken to the summit is now considered to be a much more dangerous traverse than the latterly discovered route subsequently taken by Norgay and Hillary; it is unlikely, given by what we now know, that Irvine and Mallory were successful in their ascent. These same considerations also make the claim that Irvine and Mallory were successful in their ascent implausible also means the explanatory hypothesis should be considered unlikely.
Now, aside from the exercise of asking “What might have happened that fateful day?” the other thing that makes these various theories interesting is that they are often used to refute the claim that Tenzing Norgay and Edmund Hillary were the first people to reach the summit of SagarmÄthÄ. Whilst this is no conspiracy theory ((Unless you think there is something to Popper’s divine agency analogy to belief in the conspiracy theory of society, in which case the claim would be something like “God and its agents conspired to prevent Britons from reaching the summit of SagarmÄthÄ.”)), the particular argument for the explanatory hypothesis of why Irvine and Mallory did not return to the base camp, which is that they died after a successful ascent of the summit of SagarmÄthÄ, has all the hallmarks of being an Inference to Any Old Explanation.
Now, such an Inference to Any Old Explanation is deeply problematic because if an epistemic agent claims that their particular inference to an explanation is warranted, then we should expect them to be able to give an argument which will put forward plausible and likely candidate explanatory hypotheses which, at the very least, strongly suggest the phenomenon or event to be explained. We want epistemic agents to put forward likely candidate explanatory hypotheses that are based upon the available evidence, which will range from the particular facts of the phenomenon or event (such as “Irvine and Mallory were last sighted approaching the summit at 1pm”) to claims or theories that are relevant to the phenomenon or event we want to explain (such as theoretical claims about the reliability of the technology used by Irvine and Mallory, along with what we now know to be safe and surmountable ascents to the summit of SagarmÄthÄ.
However, we should be aware that there is a tradeoff or tension between the probability of an explanatory hypothesis, based upon the evidence, and the extent to which said explanatory hypothesis suggests the evidence itself. Peter Lipton, in his 2004 book, ‘Inference to the Best Explanation,’ talks about this tension with respect to the likeliness and the loveliness of candidate ((Lipton actually uses the term “potential explanation” rather than “candidate explanation,” but, at least with respect to the following analysis, we can read “candidate” for “potential.”)). The likeliest explanation is the candidate explanation of the phenomenon or event that is best supported by the evidence whilst the loveliest explanation is the one that would promote [his terminology] the most understanding of the phenomenon or event. Lipton considers the loveliness of explanations to provide the deeper account of what makes some candidate explanation good because explanations must be more than merely suggested by the evidence. For some account to be a good explanation it must provide, by way of an answer to an explanation-seeking why-question, an understanding of why some phenomena occurred. Lipton uses the example of Newtonian mechanics to illustrate this; when Newton proposed his physical theories they were likely, because they were supported by the available evidence, and they were lovely, because they promoted an understanding of the physical world with reference to a set of simple yet powerful set of laws of nature. However, as more evidence became available the likeliness of the Newtonian theory diminished, even though they remained lovely.
Now, being overly concerned with loveliness might lead to the formulation of explanations which are specifically manufactured or designed to suggest the evidence that supports them.