I understand that this question might be difficult or even unresolved. But within a foundationalist view of knowledge, has anyone proposed a set of basic beliefs that seem to be the most rational for forming an accurate model of reality, or at least the most rational that we know of? If so, what are those beliefs?

To clarify what I mean by a 'rational basic belief', I think it's best to give an example of what I mean by an 'irrational basic belief'. An example of an irrational basic belief might be 'My favourite news channel is always reliable' or 'Whatever I want to be true is true'. Clearly, these two are irrational, but it's hard to pinpoint exactly what makes them different from a basic belief such as 'Deductive reasoning is reliable'. Nonetheless, it would seem this one is a basic belief that is rational or warranted.

So I'm not asking for a set of basic beliefs which by definition is the most rational. For example, I don't want a set of beliefs that includes a complete theory of physics, because right now such a theory is unknown, even though it technically would be the most rational thing to include in your set of basic beliefs.

Also, I'm not asking for a set of basic beliefs that most people seem to adopt, or that seems to explain the behaviour of most people, because most (if not all) people are not rational 100% of the time. I would rather just consider a set of basic beliefs on their own merits.

Another way of thinking about my question is: if I were to build an artificial intelligence/robot whose goal is to create an accurate model of reality, what assumptions should I program into it? Of course, the AI/robot is allowed to update its model of reality as it collects information, but which assumptions should underlie how it models reality?

For example, one assumption that an AI (or any rational agent) might have is "My memory is always reliable". Otherwise, the agent would not want to extrapolate from past experiences.


I found an article that provides an answer close to what I'm looking for, The Concept of Rational Belief (https://www.jstor.org/stable/pdf/27902897.pdf), but it's from 1985. So I'm looking for something similar to what's quoted below, but more recent.

From page 13:

So what we want is a policy that will bring our beliefs as nearly as may be into correspondence with the set of facts.

It could be that no policy can do this job. Indeed, it seems that the job can be done only if at least three conditions are met. The first is that there is enough lawful structure (it could be just statistical law)...

The second condition is that there be some input from the world for a person, in the sense that there be some fact, distinct logically from a judgement about it, with which judgement about it can be directly compared, although not necessarily one about which judgement is infallible.

[The third condition] ...there must be some sort of recording device, more or less, reliable, for information about earlier inputs.

Brandt, R. B. (1985). The concept of rational belief. The Monist, 68(1), 3-23.

  • 2
    Rationality is a mean to an end. No goal, no rational set of believes or actions to achieve it.
    – user14511
    Commented May 5, 2020 at 6:47
  • 1
    I think Chisholm and Alston are two of the most distinguished foundationalists and often cited when it comes to basic beliefs.
    – Philip Klöcking
    Commented May 5, 2020 at 11:22
  • 1
    Any model of reality is always possibly totally false on the basis that the brain in a bottle (Matrix film trilogy) thought experiment cannot possible be perfectly refuted. Because of this it is possible that at any moment every single cause-and-effect relationship could be totally changed by a programming change to the matrix. Thus no model of reality can ever be totally relied upon.
    – polcott
    Commented May 6, 2020 at 20:06
  • 1
    After we have addressed the brain in the bottle thought experiment the next best thing for validating one's model of reality would seem to be the scientific method. One aspect of this that almost no one has the discipline for is to apply the scientific method on the basis of first principles and not merely assume that any element of the body of knowledge derived by science is actually true.
    – polcott
    Commented May 6, 2020 at 20:10
  • 2
    I am saying that Modernism failed. There is no way to specify a view of the world in axioms and still know what those axioms mean. The meaning of the words themselves has to be established by experience, which will differ from being to being. There is a loop of usage and meaning that cannot be started. Foundationalism is a failed strategy. It is not about empiricism. it is about many, many people having tried this, until we found a really good reason to give up. And folks still not giving up. Commented May 6, 2020 at 23:48

3 Answers 3

  1. Always try to confirm or to refute your own conjectures.

  2. Be aware that also humans - alike all species - have their border of cognitive capabilities.


... it's hard to pinpoint exactly what makes them different from a basic belief such as 'Deductive reasoning is reliable'. Nonetheless, it would seem this one is a basic belief that is rational or warranted. blue-raven


Confidence in the truth or existence of something not immediately susceptible to rigorous proof: dictionary.com/browse/belief


The fact or state of knowing; the perception of fact or truth; clear and certain mental apprehension. dictionary.com/browse/knowledge

It seems that a key problem with most investigating foundationalism is that few carefully draw the distinction between knowledge and beliefs. Knowledge is necessarily true and thus impossibly false; whereas beliefs are always possibly false.

Foundationalists are united in their conviction that there must be a kind of justification that does not depend on the having of other justified beliefs They nevertheless disagree radically among themselves as to how to understand noninferential justification.


It is understood rather than merely believed that sound deductive logical inference derives conclusions that are necessarily true thus impossibly false.

The strongest possible justification for a proposition is the case where a proposition can be determined to be certainly true entirely on the basis of the meaning of its words.

The next best thing after this (does not withstand radical skepticism2) would seem to be information obtained through the scientific method, yet the reliability of this is limited by the problem of induction.

The original problem of induction can be simply put. It concerns the support or justification of inductive methods; methods that predict or infer, in Hume's words, that “instances of which we have had no experience resemble those of which we have had experience” (THN, 89). Such methods are clearly essential in scientific reasoning as well as in the conduct of our everyday affairs. https://stanford.library.sydney.edu.au/archives/sum2016/entries/induction-problem/

2 Radical Skepticism

Brain in a vat

Five-minute ago hypothesis


The set of propositions that can be known to be true with 100% logically justified complete certainty are those analytic propositions that are verified as true entirely on the basis of their semantic meaning. These are all anchored in sound deductive inference and semantic tautology.

All propositions that have any empirical aspects are limited by the problem of induction in that they depend upon possibly false assumptions.

Problem of induction
...all such inferences rely, directly or indirectly, on the rationally unfounded premise that the future will resemble the past. https://www.britannica.com/topic/problem-of-induction

i. Strong Foundationalism
Strong foundationalists hold that the properly basic beliefs are epistemically exalted in some interesting sense. In addition to basic beliefs possessing the kind of justification necessary for knowledge (let us refer to this as “knowledge level justification”) strong foundationalists claim the properly basic beliefs are infallible, indubitable, or incorrigible. Infallible beliefs are not possibly false. https://www.iep.utm.edu/found-ep/#SSH4ai

The Analytic/Synthetic Distinction
An “analytic” sentence, ... has historically been characterized as one whose truth depends upon the meanings of its constituent terms (and how they’re combined) alone... https://plato.stanford.edu/entries/analytic-synthetic/

Expressions of language verified as totally true entirely on the basis of their semantic meaning encoded as relations between the expressions of this language. PL_OLCOTT

To make analytical knowledge subject to a formal proof the expressions of language and the relations between them are encoded as finite strings. A formal proof would merely verify that an expression of language encoded as finite string has the required semantics meanings. PL_OLCOTT

I uniquely named this foundationalist basis mutually self-defining semantic tautologies The following is an example of this analytical knowledge:

Successor(Successor(Successor(0))) = 3

We know that Successor(Successor(Successor(0))) equals 3 is true on the basis that the symbol: "3" on the RHS is stipulated to correspond to the (Peano axiom) algorithm on the LHS.

  • 1
    Just to be clear, I don't mind at all if you try to point out a flaw in the premises of my question by saying: 'Some of the things "known" to be true depend upon a basic assumptions that are possibly false, so these things would not be most rational.' However, I'm not convinced that any assumption fails to be rational just because it could be false. Could you explain that statement further?
    – blue-raven
    Commented May 7, 2020 at 13:06
  • @blue-raven if your form complete certainty in your mind when complete certainty is not fully justified you are being less rational than when you are forming complete certainty in your mind when complete certainty is fully justified. When the mailman delivers your mail you don't say to yourself "either the mailman just delivered my mail or the simulation of what appears to be the mailman appeared to have delivered what seems to be mail". Because you form complete certainty when certainly is not perfectly justified you are being less than perfectly rational.
    – polcott
    Commented May 7, 2020 at 15:06
  • Okay I agree with you when you say "if your form complete certainty in your mind when complete certainty is not fully justified you are being less rational..." However, I think certainty is irrelevant in cases where somebody is forced to make a decision. For example, if someone were to pull a gun on me and demand I give them money, I would, because it's rational to believe their gun is loaded, even though I'm not certain that it is loaded. So my original post is asking for a set of (possibly uncertain) assumptions that explains how one would come to these apparently rational beliefs.
    – blue-raven
    Commented May 8, 2020 at 12:35
  • By the way, I edited my original post and linked an article since you last answered, so you might find that article helpful.
    – blue-raven
    Commented May 8, 2020 at 12:37
  • @blue-raven There are some situations where the most rational choice in the situation cannot possibly be determined. In these cases choosing the most rational one of a set of alternatives depends the actual nature of reality and the actual nature of reality may be impossible to know.
    – polcott
    Commented May 17, 2020 at 13:32

You must log in to answer this question.

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