Claiming that the fact that one used to hold the idea of the opponent to bolster one's own argument.

Example: I used to believe in your idea of X and would have argued your side supporting X until I had this experience which led me away from X to Y...

In other words, identifying with the opponent's position and claiming that now he sees the "truth" to suggest that the opponent is just missing something, without providing a concrete argument.

It seems that this is a type of non sequitur (or maybe an argument from one's own authority, if such a thing exists) but is there a more specific fallacy describing this? This type of argumentation is all too common.

  • 2
    Why do you think it's a fallacy? Mentioning earlier agreement is a rhetorical trick to gain the hearer's trust. It isn't used in the argument, which is solely based on his experience that led him away from this point of view.
    – user2953
    Oct 20, 2015 at 6:38
  • Great point! That makes a lot of sense.
    – Jeff111
    Oct 20, 2015 at 6:55
  • I use a construction like this as a header, so that the non-ovbiousness of the counterpoint is made clear.
    – Joshua
    Oct 9, 2016 at 16:22

4 Answers 4


You are correct, structurally speaking this is the fallacy of appeal to authority. The person isn't providing any actual argument against your position, he is merely claiming to be in a position of authority to judge between the two positions.

As with all informal fallacies, this argument gains its force from superficial similarity to a good argument. For instance, if I tell my child "I was once your age, and I ate a whole bag of candy, and I was sick all night, so you should not eat that whole bag," that is a good argument that is very similar to the bad one, except in this case I've actually provided support for the relevance of my experience.

  • But if the rhetoric drops out and it is strictly formalized, can't "new information" increase the probability that initial belief is wrong, even if the "reason" is black-boxed? As argued below. Oct 20, 2015 at 16:45
  • @NelsonAlexander It's a mistake to believe any informal argument can be "strictly formalized." At most you can create a formal argument with structural similarities to the informal one. In the case of your argument, it's not clear to me why you formalized it the way you did, and what formal system operates under the rules you describe --which precludes me being able to judge it as formally valid or invalid. It looks like you're assuming Bayesian probability or some other system of probabilistic logic... but which one, and why? Oct 20, 2015 at 17:47
  • Yes, Bayesian probability, somewhat, and foggy memory of the Monty Hall paradox. But I have no formal training in logic or probability, though I do not thereby toss myself out of court. My "formalization" is simply stripping the personal unknowns that everyone else was assuming and treating this as a rational interaction with some form of "new information." I think it is defensible. It is also not arbitrary. It indicates why such arguments may (and sometimes do) legitimately carry force in some situations. I believe it captures what actually happens better than just "appeal to trust." Oct 20, 2015 at 18:16
  • 1
    @NelsonAlexander There are a host of technical challenges involved in translating from natural language into a formal language. You can't make arbitrary decisions and call it "formalization." Your current argument isn't coherent. If you are really interested in formal logic, you should study what has already been done and use that as a starting point, rather than reinventing the wheel. Oct 21, 2015 at 2:19
  • I appreciate your comments. I understand that there are formal systems of logic and that there are problems translating from natural language. But I do not accept your own "argument from authority." My so-called "formalization" may be crudely handmade. But I do not see where my very restricted case is invalid or "arbitrary." "It explains, using the Monty Hall analogy, why such arguments ARE accepted in certain cases, which everyone else dismissed merely because the "reason" remained blackboxed. I am treating the "reason" as "new information."Not at all arbitrary or unrealistic. Oct 21, 2015 at 13:17

I share your dislike of this kind of argument. If someone wishes to say that they used to believe X and now believe Y and proceed to offer the reasons why they made the change, then that's fine - the reasons are what matter. But it is all too common for people merely to say, well I used to think X but now I believe Y as if that magically makes Y more plausible than X just because they changed their minds. Tell me the reasons or go home.

As to whether it is an instance of a fallacy... I'm not a big fan of hunting for fallacies - to me that is just kindergarten logic and if you progress in your understanding of logic then you should grow out of it. But as it happens, it is an example of non-sequitur. The fact that your opponent used to believe something else does not of itself make Y more plausible.


If your debating opponent says he once shared your position X, then this can be to gain your trust and establish a shared foundation for debate and engagement and because he may move onto a position that extends this in an organic way or in an interesting way.

Or it can be, and sometimes is used to already move on to a different position that is diametrically opposite; and here the tone matters - for it might be a pitying tone, or a patronising tone supposing himself to have moved on to somewhere you have yet to follow - which would then make you a follower.

And this is exacerbated if no reasons are given, or promised, either in full or in part.


Allow me to attempt a dissenting opinion.

While I agree generally with all of the above observations, and find such arguments generally annoying, they can have a certain rational force in certain circumstances, and I am wondering why.

First, we have two "very dumb" (pardon the technical term) interlocutors, In1 and In2 and one belief (B1). The first interlocutor proposes only B1 for both In1 and In2. The second interlocutor proposes, let's assume, a second belief B2. We now have two interlocutors and two beliefs B1 + B2 with, mutatis mutandis, equal likelihood.

Now In2 proposes B1 + B2 - B1 = B2. She has effectively doubled the number of possible beliefs and then negated one. Even if she gives no good reason, In1 now has 50% less reason to believe only B1. In information terms, the number of possibilities to be rationally reduced to "certainty" has doubled.

Moreover, In2 has implied that there is some blackbox "reason" for B2, assuming they are two agents "reasoning" with one another. She implies that new evidence arose as time passed. By the principle of sufficient reason, there should indeed be some "reason" for the appearance of the new belief B2. Since In1 and In2 are already "reasoning," there is a non-negligible chance that she may have some "valid reason" for the new B2.

Nor is the matter of "trust" negligible or necessarily a rhetorical device. That "reason" too is unknown. But since both In1 and In2 shared B1 prior to B2, there is evidently some overlap of what they accept as "valid reasons." Hence a greater likelihood that the blackboxed "reason" may be valid.

In reality, this type of underwhelming argument does work best the more information and "possible beliefs" are limited, with "young" people, say, or people with "less expertise." Hence my qualification of "very dumb" and my limitation to just two possible beliefs between equally "reasonable" people.

In practice, the argument will carry scant weight in a world of infinitely many possible beliefs. We indeed quickly "grow out" of such arguments. But even where the "reason" is not given, good or bad, there is some probabilistic weight to the argument given very restricted beliefs. It is a bit like the Monty Hall paradox. All else equal (as it never is) In1 will technically improve odds in the new domain by switching to B2.

  • I've tried my best but I can't understand this at all. 1) What does this have to do with the original question? 2) What does it mean to propose B1 "for" In1 and In2? 3) What does it mean (to you) to double the number of possible beliefs? 4) What does it mean (to you) to negate a belief? 5) Do you have any methodology for reducing indistinguishable beliefs? 6) How are you calculating these probabilities and why? 7) Why does In2's proposal change the probabilities for In1? 8) What is the "principle of sufficient reason" and how are you applying it? Oct 21, 2015 at 14:47
  • Yes, my terminology is pretty sloppy. Here's how I'm framing and restricting the question. We have two "rational" interlocutors. (They aren't liars, insane, enemies, etc.) One proposes a "belief" for the other to accept, "B1" for simplicity. Two interlocutors, one proposed belief. (Note that I am stripping away all sorts of "possible beliefs," hence "very dumb.") The second person says, " I used to believe B1. Now I don't. I believe "B2," having "seen the light," or whatever. I am assuming the scenario does imply some sort of B2 or alternative to B1. Now an admittedly debatable assumption... Oct 21, 2015 at 15:29
  • ...I assume that the first interlocutor is "very dumb," meaning has very limited number of possible beliefs, which did not include B2. (This is not unreasonable in the cases of children or "non experts" or "foreigners." Not utterly arbitrary.) So she is confronted now with B1 or B2. More possible beliefs to choose from. Why not stick with B1? The other person has still offered no "reason." Here, other commentators were assuming until the "reason" is given the "logical situation" has not changed at all. I am arguing that it has. Here's why. The first interlocutor has new information.... Oct 21, 2015 at 15:41
  • ...She now has the expanded domain of B1 and/or B2. But she has other information as well. The blackboxed "reason." She knows there must be "some reason" for B2, good or bad. (That's where I mentioned, Leibniz's "law of sufficient reason," a distraction.) She also knows the proponent of B2 "used to" share her reasons for B1. So they can "share" reasons. So even without opening the box and seeing the "reason," she may have good cause to switch. (Here I also refer to the "Monty Hall" problem in probability, which I find hard to grasp, but it is perhaps not really necessary in this case.) Oct 21, 2015 at 15:56
  • 1
    So you're basically saying that just knowing someone else holds a incompatible different belief weakens, at least to some extent, your warrant for your own belief, even if you don't know why that person holds that belief. That makes sense. Your technological terminology is just obscuring your argument, not illuminating it --terminology never helps unless it is shared (or well defined and explained). Oct 21, 2015 at 16:55

You must log in to answer this question.

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