This came up from a discussion about the design argument.

Suppose you roll a dice and it lands on 6 10 times. Now, this could be explained by cheating. If we live in a deterministic universe, it could also be explained by the laws just being set in a way that results in that without humans being involved.

Of course, the latter theory has no evidence. The first does. But it seems that this evidence is based on past experience: we know humans exist and are capable of cheating.

Technically though, we don’t have any direct observational evidence of humans cheating on this particular dice roll (assuming that we didn’t observe any rigging process). We make an inference based on the past and consider our experience of humans cheating to be reliable enough to show that a human probably rigged the dice.

But clearly, this analogy would break down if we considered extraterrestrial designers rigging the dice or god rigging the dice or a fairy rigging the dice etc.

This is presumably because we have no direct observational evidence ever of those kinds of entities designing things. However, what if one argues that we have evidence of designers doing things, and god is a designer; hence, the entities are similar enough to postulate god to explain an event that looks designed. This sounds weak but why exactly? How does one identify how strong or how weak an analogy has to be before making an inference?

  • 1
    All reasoning is ultimately from analogy. When from A, B, C we conclude D, then we are judging that the inference ABC->D is similar enough to inferences that worked for us in the past. Or specifically that ABC are similar enough preconditions to those preconditions from which we in the past correctly concluded the similar equivalent to D. This even applies to deductive reasoning. Analogy = pattern matching = how the whole brain works.
    – causative
    Feb 4 at 9:01
  • @causative That might be how deduction happens in the brain but it has nothing to do with deductive reasoning generally.
    – user71009
    Feb 4 at 9:17
  • 1
    @abcga What distinction are you drawing between deduction in the brain and deduction "generally"? Are you thinking of deduction as done by a computer? All deduction is pattern matching, whether in the brain or in a computer, and pattern matching is essentially analogy. If pattern A matches pattern B, then pattern A is analogous to pattern B.
    – causative
    Feb 4 at 9:27
  • Your question turns out to be broader than you've described it; I've given a broad answer in response. If you want to investigate a specific analogy, then please open a fresh question which doesn't admit a generalized response.
    – Corbin
    Feb 4 at 22:00
  • This is equivalent to the famous Hume's problem of induction, once it's deemed to be justified empirically, then you can invoke argument from analogy to arrive at similar effects from similar causes... Feb 5 at 7:13

3 Answers 3


Yes, you cannot ever conclusively determine that a random process was hijacked, but you can adjust your beliefs accordingly (What is the probability that the chance of getting heads/tails is 50/50 if you try 1000000 times and get only tails?). I cannot perceive how it's about reasoning from analogy.

  • It’s reasoning by analogy or atleast past experience because in order to think there is cheating, you would need to first know if it’s possible to rig a coin in the first place, and whether that has happened before. The current sequence would then map to your previous experience and knowledge which will then allow you to infer that there was cheating. Without this knowledge and experience, you have no reason to prefer cheating over the hypothesis that the universe was simply determined from the start to result in those tosses. Feb 4 at 8:57
  • And presumably, without that knowledge, you might even have reason to prefer the non rigging hypothesis over the rigging hypothesis since for any other sequence, you prefer that anyways. Feb 4 at 8:59
  • @Baby_philosopher That's correct (although I don't see why the universe couldn't be "determined from the start" to result in someone cheating). But I still don't see how it's about analogy at all. Yes, there might be other equally good explanations than cheating. Why would we suspect someone to cheat? I assume because humans are discursive beings and we might attempt to question them to verify that etc.
    – user71009
    Feb 4 at 9:07
  • So if by "analogy" you mean "past experience" then there is nothing puzzling about the fact that there might be many equally good explanations. If you mean something more specific, I don't see how it's relevant.
    – user71009
    Feb 4 at 9:08
  • You are correct that the universe could be determined from the start to have cheating. But you’re still then comparing cheating with a determined result without cheating. And the former has more evidence. One should still prefer the former in the case of dice since we have reasons to think that humans can cheat and do so and that it’s possible for dices to be rigged Feb 4 at 9:09
  1. So called “arguments from analogy” are no arguments at all.

    The two situations, which are compared in such cases, differ in at least one property. There is no certainty, that it is not exactly this difference which violates the whole reasoning.

  2. Your second proposal “the laws just being set in a way that results in that” is no theory at all. It is just the claim that the observation underlies a certain law.

    To explain the observation by a law requires much more generalization. The law has to be stated explicitly, covering much more cases than the one under observation, and proposing some mechanism behind.

  • I think you misinterpreted my post. I am referring to existing laws, not any new ones. So in the case of a coin landing on tails 100,000 times it could be the case that “by chance” natural laws resulted in a fair coin landing in that sequence. Feb 4 at 9:38
  • 1
    Are you suggesting that an argument without certainty is not an argument?
    – causative
    Feb 4 at 13:44
  • @causative I do not suggest this. Why do you think?
    – Jo Wehler
    Feb 4 at 14:26
  • You said, "There is no certainty, that it is not exactly this difference which violates the whole reasoning." Is this lack of certainty relevant to your preceding claim that argument from analogy is not an argument? Seems you were claiming that argument from analogy is not certain and therefore not an argument.
    – causative
    Feb 4 at 14:35
  • 1
    Yes, so it is exactly as I described - you are suggesting it is not an argument because it is not certain. Of course, many times argument from analogy can be very strong evidence.
    – causative
    Feb 4 at 15:50

I'm answering the question as asked and ignoring the given example, which is already discussed in existing answers.

Analogies can be formalized with ontology logs, or "ologs", which are structured ontological analyses. An analogy is a partial bijection between ologs. An argument from such an analogy is as strong as its bijection, because the analogy can be used to map components of the argument from one domain to another.

As an example, consider Giesa, Spivak, & Buehler 2011, which draws an analogy between spider silk and music. On p10, there is a table which enumerates portions of the bijection informally. On p12 and p13 each, a richer diagram is drawn which illustrates each olog and lays out its internal hierarchies in a way which eases verification of the bijection.

More generally, take any formal proof within any deductive system. By basic abstract nonsense, this proof has an underlying poset; intuitively, two steps of the proof are comparable when one step requires the other. Then, any bijection between two posets of two proofs yields an analogy, regardless of which deductive systems were used to obtain them.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .