Are two persons equally rational in choosing different dogmatic stopping points in their chains of justification as per the Münchhausen trilemma?

Question

In epistemology, the Münchhausen trilemma is a thought experiment intended to demonstrate the theoretical impossibility of proving any truth, even in the fields of logic and mathematics, without appealing to accepted assumptions. If it is asked how any given proposition is known to be true, proof in support of that proposition may be provided. Yet that same question can be asked of that supporting proof, and any subsequent supporting proof. The Münchhausen trilemma is that there are only three ways of completing a proof:

• The circular argument, in which the proof of some proposition presupposes the truth of that very proposition
• The regressive argument, in which each proof requires a further proof, ad infinitum
• The dogmatic argument, which rests on accepted precepts which are merely asserted rather than defended

The trilemma, then, is the decision among the three equally unsatisfying options. Karl Popper's suggestion was to accept the trilemma as unsolvable and work with knowledge by way of conjecture and criticism.

Source: Münchhausen trilemma - Wikipedia

Suppose that two persons, X and Y, are asked the same question "why do you believe what you believe?", and for each justification they offer, they are subsequently asked the question "why?".

Suppose also that both X and Y avoid circular arguments and infinite regresses. Thus, their chains of justification (or DAGs, to allow for potential branching in the justification process while avoiding cycles) end up terminating, in one way or another, in dogmatic assumptions.

Considering this, are X and Y equally rational, no matter what dogmatic assumptions they adopt as their stopping points? What if X, for example, decides to adopt theistic dogmatic assumptions, whereas Y decides to adopt naturalistic ones?

Furthermore, suppose that we attempt to break the symmetry between X and Y by adopting a meta-criterion to assess the rationality of their dogmatic assumptions. Wouldn't this meta-criterion also suffer from the same justification issues highlighted by the Münchhausen trilemma? Would we need to come up with meta-dogmas in order to justify our meta-criterion, and meta-meta-criteria to justify those meta-dogmas, and meta-meta-dogmas to justify those meta-meta-criteria, and meta-meta-meta-criteria to justify those meta-meta-dogmas, and so on and so forth ad infinitum?

Answer 1

As a preliminary, let us note that building up a structure of reason from the ground need not be dogmatic. Dogmatic carries somewhat pejorative tones as resistant to reason. In a debate between an atheist and a priest, it may very well be that either the atheist or the priest or both may be dogmatic. I think the term you are searching for is foundational, as in foundationalism:

Foundationalism concerns philosophical theories of knowledge resting upon non-inferential justified belief, or some secure foundation of certainty such as a conclusion inferred from a basis of sound premises.

Second, you're in very nebulous semantics with the phrase "are equally rational" because devising a metric for rationality is no small feat. What does it mean to be reasonable or rational to begin with? Robert Audi has written The Architecture of Reason (GB) and provides a theory on the matter. And Robert Audi is but one epistemologist.

Even if two persons both have a foundational argument, there is an exhausting list on how the use of reason might differ between two parties: differences in knowledge of facts, differences in use warrant and rebuttal, linguistic sophistication, etc. What constitutes an effective rubric of rationality? That's hard to say. Whether or not an argument is easy to digest as a DAG with a root and leaf nodes is somewhat irrelevant, TBH.

One could consider the types of logic that each party uses. If one arguer is a high school student and relies on informal logic in natural language, and the other arguer is a professional logician who can rely on formal and non-classical logics and has a detailed knowledge of fallacy and theory, then the argument could be made that whether the argument relies on regress, foundations, or circularities is somewhat orthogonal to rationality all together. The primary characteristic, then is not argument structure, but rather then use and application of inference moving from premises to conclusion.

Lastly, consider that the topic in question may have a role in regards to rationality. Is the argument over the philosophy of quantum physics in which new discoveries play an important role? Perhaps the logician in our scenario refuses to admit new evidence because of confirmation bias. Then, indeed, this dogmatism might impact an assay of rationality. Thus, a younger student might not have the same mastery of inference, but at the same time accept some important defeaters that an more conservative thinker refuses to recognize.

Ultimately, rationality is multidimensional. So without more details, it's hard to commit to an answer to this question. It might be helpful to familiarize yourself with some contemporaneous thinking on the matter. Consider The Oxford Handbook of Rationality (GB). There are some SEP articles that might be of interest:

• Structural Rationality
• Instrumental Rationality
• Bounded Rationality

Remember, being rational is a subjective claim, because logic itself is goal driven and very subjective in its application.

Answer 2

Your criteria of rationalism to evaluate two responses to the Trilemma is, itself suspect. Rationalism is -- not rationally justified. See this question and answer: Logic and its beginnings and why it is

My belief is that between logical pluralism, and the Trilemma, that rationalism (and hence Analytic philosophy) is fatally compromised, and ultimately all valid responses to the Trilemma have to adopt pragmatism of one kind or another. See this answer: https://philosophy.stackexchange.com/a/64646/29339

One pragmatic response is to accept that no justifications are ever complete, hence there is not chain of justifications with a termination, there is only a network of better or less justified assumptions, that eventually trickles out into assumptions that are generally not even recognized. This pragmatic network of justifications are generally based on methodological naturalism/empiricism, at least in the West, but the justifications of empiricism are circular, so there is another leg that is also usually being resorted to. The Popperian solution is to treat ALL justifications as incomplete, always -- and subject to revisiting in the future.

There is no consistently reliable way to measure or otherwise apply a metric between two such networks of decreasingly supported assumptions. This is illustrated with Godel's Incompleteness Theorem, the way the Verification Principle refutes itself, the way Fasifiability fails Fasification, Popper's own failure to develop a valid criteria of Versimilitude, and Lakatos' failure to develop a valid criteria to measure progressivity or regressivity of a Research Programme.

What we have, as the best alternative I know of, is that searchers of good will and honest intentions can dialog, questioning each others assumptions, and apply their best judgement to the validity of the answers. This process requires intellectual honesty, and openmindedness, to be successful. Both philosophy and religion tend to attract absolutists and ideologues, which means an unfortunately large number of dialog partners may not meet the criteria for this method to be successful.

Answer 3

Your question has no objective answer as per the trilemma itself.

In order to say if something is rational, people need to agree on a standard of rationality, but this can’t be done according to the trilemma. Thus, this also applies to the question of whether two people are equally rational in choosing different “dogmatic stopping points”.

Answer 4

The Manchhausen "trilemma" states that circular arguments are unsatisfying:

The circular argument, in which the proof of some proposition presupposes the truth of that very proposition...

... is among the three equally unsatisfying options.

The Manchhausen "trilemma" presupposes:

• the three proposed "argument types" are exact, exhaustive, exclusive, valid
• all proofs are one of the three proposed types
• all three types are equally unsatisfying

as proof that all proofs are unsatisfying.

Maybe Karl Popper was pranking philosophy students?

It makes sense that he concludes:

"work with knowledge by way of conjecture and criticism." - Karl Popper

---

Per "Suppose that two persons, X and Y, are asked the same question "why do you believe what you believe?", and for each justification they offer, they are subsequently asked the question "why?"."

Where I am "Y", my response is...

"I have no "beliefs". I have in my head, no thoughts or notions or concepts or conclusions that I in any way characterize as a "belief".

I have conclusions.

I have suspicions.

I have theories.

I have guesses.

I have suggestions.

I have knowledge.

I have suspicions.

I have predictions.

I have estimates.

I have awarenesses. Recognitions. Understandings.

But nothing I need to apply the word "belief" to.

SOME would want to call my knowledge "justified true belief".

I would call it informed conclusion... knowledge.

Therefore, when asked "Why do you believe what you believe?"... my answer is "I have no beliefs".

If the question is "How do I know what I know?" or "Why do I think I know what I know?" that is an answerable question.

But really needs more context. More specifics.

---

I can give some examples, to elucidate.

FUNDAMENTAL FORCES... I do not believe in them. I expect the do not exist. I may be wrong. I expect they are emergent... a result of interactions at a sub-atomic scale that are not "forces". I may be wrong.

THE BIG BANG... I do not believe in such an event. I expect the universe is eternally old, infinitely vast, populated by an infinite number of planets and stars and galaxies, and galactic groups and galactic clusters and galactic superclusters... in a steady state situation. But again, I may be wrong.

THE AGE OF THE EARTH... I trust that the academics that have studied the evidence have a good enough set od data that reconciles enough to allow for just conclusion that the Earth is at least some billions of years old. I consider it fair to expect that they might be off by some margin of error due to some gaps in knowledge. Maybe a few percent. Not more, because the "gaps" in knowledge are not significant. My choice to trust the academic position to be pretty much correct may be misplaced. But I expect not and have zero reason to expect that is the case.

THE SHAPE OF THE EARTH... doesn't need discussing. dissidents aren't worth addressing.

One really truly does not have to hold a place in one's head for "beliefs", if one doesn't want to and has no time for them.

For "justification"... go with things such as:

• reproducability
• confirmability
• reviewability
• reconciliation
• coherence

Answer 5

I find it slightly ironic that your question presupposes that an infinite regress is associated with any attempt to determine whether two positions are equally rational, yet you still expect a definitive answer!

Yes, whatever criteria you use to determine whether a particular dogmatic view is more or less rational than another, someone could challenge you to justify your criteria and you will eventually have to resort to your own dogma for the justification.

If you reflect upon the process by which a person decides whether a thing is reasonable, you ought to conclude that many factors are involved, and some of them inevitably are the result of the person's upbringing, social milieu, education and so on. People develop their beliefs from birth onwards, based on countless experiences and inputs. Many of our beliefs are held unselfconsciously- I suspect you believe, for example, that when you walk down a street you are unlikely to find the ground vanish beneath you. That will not be a belief you will have consciously arrived at through analysis of what is more or less reasonable to believe- you will have formed the belief on auto-pilot, as it were, from having repeatedly and consistently experienced the ground as something solid. You will have thousands of other beliefs that you have acquired in a similarly uncritical way. I imagine, for example, that you belief the Earth is not flat, that the arctic is colder than the Sahara, that Paris is the capital of France, because you have absorbed that sort of information consistently through your life. You might first have been taught at school that Paris was the capital of France, having earlier been conditioned to assume that what teachers tell you is correct, and thereafter your belief in the fact will have been buttressed again and again in other ways, for example, by seeing Paris mentioned on TV, or reading about it in books etc.

In short, you acquire views of the world from an interlocking set of inputs and experiences, each of which you have been found to be consistent with many others. It is against that set of beliefs that you will form conclusions about whether other proposals are reasonable. If I am asked to say whether a naturalistic or theistic view is more reasonable, I will tend to favour the naturalistic, because I find theistic beliefs to be implausible in the context of my personal belief system. However, someone raised to be religious might take the opposite view. There is no absolute way to say which is more rational, and that is why the same old arguments about the existence of god continue century after century.

Answer 6

The trilemma points out a problem, but that problem is particular to a specific approach, the approach that demands that all beliefs must be justified with some source which itself is subject to justification. This seems to be a rational demand, and at first it seems universal, but it is not. Therefore, we can find a few off-ramps which are available to rational people. Some of these may be more rational than others.

The first is, as Professor Sushing's answer mentioned, that we come into the question with certain forces compelling us toward beliefs. This includes things like habit and cultural influences or even intuition. In a sense, our impulse to seek rational answers is simply to satisfy a habit or intuition to suppose that a certain belief may be suspect, and we might pursue that question only so far as necessary to rule out any suspicions.

Another, which may have been hinted towards here, is that in the absence of conclusive justification for one belief over another, we could still compare in some sense the various implications of the core beliefs and discard as irrational any which result in beliefs which cannot be coherent, such as those which are self-defeating.

However, my preferred approach is bit of a mix of the other two. I propose that when we invoke the trilemma, we are invoking at least one and perhaps two necessary assumptions. The most easily observed is the fact that the trilemma presumes that we have an ability to reason (which itself seems to require certain concepts, such as the three laws of reason and some sort of a demand for justification). The second is that there exists an "ought" of some sort such that we ought to believe that which is rational. If we excluded either, then the trilemma itself would be incomprehensible, and thus, without proving that they exist and without demanding further justification, we have already presumed them. In my experience, we end up making the same presumptions any time we ask these questions, and so they hold even if we approached from an angle other than the trilemma.

If we allow those two assumptions, we have actually created a framework which can escape the trilemma because we are no longer forced to continue ad infinitum. Yes, we do presume something, but something such that if we didn't presume it, then there would be no force to compel us to not presume it, and we can fall beck to "lesser" methods of inquiry, such as our underlying intuition or habit (in believing that these two things exist). With these two presumptions, implied by the trilemma, we have a foundation which can be the stopping point to prevent infinite regress, or non-circular DAGs as you put it.

You mentioned deciding between, for instance, a theistic set of assumptions versus a naturalistic set of assumptions. This is apt because such an approach would involve questioning the most core assumptions regarding our beliefs, and as far as I can tell, this is the only rational method of judging between such systems. With this method, we would simply be asking if X or Y better satisfies (and accounts for) both reason and impetus (ought). This happens to be a popular form of argument used by Christian Apologists. For instance, the Transcendental Argument for God asserts that our very inquiry demands certain "transcendentals" (such as perhaps reason and impetus), and proceeds to show how such can only be supplied in the case that a god exists. Also, some philosophers do as Plantinga and critique the Naturalist model by suggesting that it contradicts (or at least makes dubious) our ability to trust in reason. I am not personally aware of any such argument for Naturalism, leading me to suspect that it may be less rational, tending to rely on additional assumptions, but it is possible that such might exist.

Answer 7

If, "There is an epistemological trilemma, with its horns being these inferential sequences of these relatively specific kinds, such that..." is used as a premise in an argument, then is not the existence of this trilemma being used as a foundationalistic starting point in some chain of reasoning? If we argue for foundationalism from objections to two horns of the trilemma, are we inferring foundationalism, after all? Is, "There are axioms," itself an axiom or a theorem, then?

Now, there is something of a turnabout-is-fair-play attitude to the way Reformed epistemology plays the game of foundationalism, except that often enough the Reformed epistemologist will go on to appeal to a sensus deitas or suchlike to say that a believer can (to put it in a sort of first-order/second-order manner; c.f. the SEP article on higher-order evidence) Level₂-rationally claim to be Level₁-more rational than nonbelievers. Yet otherwise, even nonbelievers might be just as Level₂-rational as believers, here, it seems, sensus deitas or no (unless we were illuminationists, perchance...).

Pluralisms can be construed as allowing for mutually rational (or reasonable, or justifiable, or whateverable) but distinctive systems of starting points, though. There is logical pluralism, for example:

(C.f. the distinction between religious pluralism and religious inclusivism: "Religious exclusivism and religious pluralism appear in most categorizations, but not always with the same meanings. Religious inclusivism also sometime appears, but primarily in discussions about sufficient conditions for spending eternity with God, as it does in the discussion below (section 8) on the Eternal Destiny of Humankind.")

Answer 8

Munchausen's trilemma explains that if you are trying to justify your ideas, show they are or good or whatever, then you have a choice between (1) infinite regress of justifications (2) dogmatism about some assumptions (3) circularity

One solution to this problem, suggested by Karl Popper in On the Sources of Knowledge and of Ignorance, which is in his book "Conjectures and Refutations", and in "Realism and the Aim of Science" Part I, Chapter I, is to abandon the standard that ideas can and should be justified.

Instead we should look for problems with existing ideas and try to solve them by proposing alternatives that solve those problems. The relevant standard of rationality is that you should solicit criticisms of your ideas from other people as well as looking for them yourself and creating institutions that make it easier to criticise and replace ideas.

This eliminates the problem of Munchausen's trilemma by pointing out that if justification is impossible we should stop trying to do it and get on with fixing problems.

For some philosophy that tries to build on and improve Popper's ideas see

https://criticalfallibilism.com/introduction-to-critical-fallibilism/

Answer 9

Are X and Y equally rational, no matter what dogmatic assumptions they adopt as their stopping points? What if X, for example, decides to adopt theistic dogmatic assumptions, whereas Y decides to adopt naturalistic ones?

Taking a colloquial understanding of "rationality", I do not believe so. Certain sets of dogmatic assumptions will inevitably result in the invalidation of some beliefs held by one who upholds a sufficiently inconsistent set of dogmatic assumptions. Of course, one could abandon consistent logic, though that would seem to entail a lack of rationality (given that paraconsistent logics are not considered rational), and rationality is the metric of comparison here.

More interestingly though is the case in which both individuals have differing sets of dogmatic assumptions that are internally consistent. Such dogmas seem to be equally "rational" if we are considering only this horn of the trilemma.

Furthermore, suppose that we attempt to break the symmetry between X and Y by adopting a meta-criterion to assess the rationality of their dogmatic assumptions. Wouldn't this meta-criterion also suffer from the same justification issues highlighted by the Münchhausen trilemma? Would we need to come up with meta-dogmas in order to justify our meta-criterion, and meta-meta-criteria to justify those meta-dogmas, and meta-meta-dogmas to justify those meta-meta-criteria, and meta-meta-meta-criteria to justify those meta-meta-dogmas, and so on and so forth ad infinitum?

You seem to be implicitly assuming the interchangeability of "rationality" and "justification". If we are to consider them interchangeable, assume that consistent dogmas are chosen, and allow for the adoption of "meta-criteria", I would be inclined to agree with you. However, it seems that implementing such criteria would defeat the notion of "dogma", as "dogma" is meant to be axiomatic and rigid; it is that with which you scrutinize your beliefs, not that which is scrutinized.

Answer 10

Münchhausen trilemma only relates to the impossibility of absolute deductive proof.

But we were never using absolute deductive proofs to gain knowledge to begin with. This doesn't preclude other means of justification, such as coherence, explanatory power and considering the reliability of a method.

This would avoid the trilemma altogether, because it wouldn't be a foundation of deductive proofs, which the trilemma is specifically about.

---

As an example, consider the question of determining whether someone is a murderer. Courtrooms don't consist of people trying to write down absolute deductive proofs that someone is a murderer. Instead, we take all the evidence and build a coherent model of the events that transpired, which has high explanatory power in terms of having fewer assumptions, having fewer "surprising facts", better accounting for more of the evidence, etc. And we make use of reliable methods for evaluating evidence, like robust science, instead of just some guy speculating about what happened.

Also consider science. It's commonly accepted (especially by scientists) that scientific models aren't an exact description of the foundational nature of what actually exists. It merely models the behaviour of what exists (or of what is perceived to exist, if you want to go down the solipsism route). There is some room for uncertainty and for it being incorrect. But such models are nonetheless still extremely helpful in navigating the world, and it was foundational in the development of most of what you see around you and what you take for granted in the modern world.

Yes, it is the case that courtrooms and science do build on top of some already-accepted claims. But one can use similar means to what's been described above, to conclude, for example, the truth of induction and the existence of the world outside of your own mind, which are two fairly foundational beliefs.

Münchhausen trilemma is more a useful note about the limits of knowledge, how we should think of what knowledge is, and that there should always be at least a hint of uncertainty. It doesn't eliminate all forms of justifying beliefs, and it certainly doesn't mean "anything goes".

---

Side note: Despite the perceived robustness and general-objectiveness of deductive proofs, they often end up being a disappointingly unreliable tool for gaining knowledge about reality. You need to start from some premises and make relatively small (even if far-reaching) inferences, which often results in some people accepting the proof and some rejecting it, based on whether they accept or reject the premises. This typically merely validates the conclusion they already hold, rather than helping us figure out what's true or justified. Don't get me wrong: it's extremely useful to learn logic and learn about fallacies and such. But like any tool, it's more useful for some things than it is for others, and it's not perfect.

Münchhausen trilemma is a problem for anyone who wants to say deductive proofs are perfectly objective or perfectly reliable (if the fact that so many people disagree about them isn't enough evidence of that already). Although people also often employ plenty of questionable logical leaps, even if they start from a set of facts that others may agree with.