4

Foundationalism, once considered a valid and popular philosophy, now receives nearly universal contempt. There seems to be a consensus, in both analytic and continental camps, it is dead.

Are there any modern attempts at a resuscitation – justifying it and addressing the seemingly fatal arguments against it?

9
  • Well, you have a reasoning agent who investigates the world. Could be a human, could be a machine learning program. This agent must necessarily begin with some structure to its mind, or it obviously would not be able to investigate or have a mind at all. This initial structure may encode initial assumptions, and tells the agent what evidence it will find persuasive of what propositions. For a machine learning agent, the initial structure would be however it was programmed. For a human, the initial structure is given to us by our genes.
    – causative
    Mar 28, 2021 at 23:09
  • So that's where the foundation of knowledge rests for that agent - in whatever structure the agent's mind had initially. All other beliefs of the agent rest upon and are derived from that initial structure, which causes the agent's mind to change over time in combination with observations of the world.
    – causative
    Mar 28, 2021 at 23:10
  • 3
    Foundationalism is down but not out, especially in religious circles. Currently the most popular form of it is reformed epistemology, see IEP, Modest Foundationalism.
    – Conifold
    Mar 29, 2021 at 5:11
  • @Conifold so what is en vogue instead? Formally there remain three answers to the infinite regress problem: (1) coherentism, (2) infinitism and (3) skepticism. It can’t be (2) and (3) … but is it really (1) that is popular? It seems we are still stuck with foundationalism. Just of a more moderate sort. It’s not even called “foundationalism” anymore…
    – viuser
    Nov 25, 2021 at 23:47
  • @viuser None of the above. Agrippa's trilemma presupposes that claims must be justified individually and from other claims. Both premises are rejected, and so it goes by the wayside. What we get is a system of claims that is only justified as a whole, but not by coherence with some pre-existing assumptions (as in coherentism) but rather by "pragmatic success", or something else non-propositional. And since the system is open to perpetual correction, the standard of justification is much lower than in foundationalism, being best available option at this time is enough.
    – Conifold
    Nov 28, 2021 at 6:21

5 Answers 5

5

To expand slightly on what Conifold mentioned, according to IEP the "modest foundationalism" has Alvin Platinga as a prime exponent; Wikipedia mostly covers that under "reformed epistemology" although it does say under "modest foundationalism" that:

Reformed epistemology is a form of modest foundationalism which takes religious beliefs as basic because they are non-inferentially justified: their justification arises from religious experience, rather than prior beliefs.

And in the more focused wiki article:

In Plantinga's view, warrant is defined as the property of beliefs that makes them knowledge. Plantinga argues that a properly basic belief in God is warranted when produced by a sound mind, in an environment supportive of proper thought in accord with a design plan successfully aimed at truth.[10] Because there is an epistemically possible model according to which theistic belief is properly basic and designed to form true belief in God, belief in God is probably warranted if theism is true. Plantinga does not argue that this model is true, but only that if it is true, theistic belief is also likely true, because then theistic belief would result from our belief-forming faculties functioning as they were designed. [...]

Other prominent defenders of Reformed epistemology include William Lane Craig, William Alston, and Michael C. Rea.

(I think Nicholas Wolterstorff can be safely added to that list.) From Plantinga Wikipedia bio:

Ultimately, Plantinga argues that epistemological naturalism- i.e. epistemology that holds that warrant is dependent on natural faculties—is best supported by supernaturalist metaphysics—in this case, the belief in a creator God or designer who has laid out a design plan that includes cognitive faculties conducive to attaining knowledge. [...]

William Lane Craig wrote in his work Reasonable Faith that he considers Plantinga to be the greatest Christian philosopher alive.

According to Wikipedia's bio of Platinga he "was the 30th most-cited contemporary author in the Stanford Encyclopedia of Philosophy" according to one analysis. (N.B.: William Alston was #48 in that list. Robert Audi (see below) #114; Nicholas Wolterstorff #152.)

Although Platinga may be most famous now, according to IEP:

William Alston probably did the most to rehabilitate foundationalism.

You can probably find more related figures in the list of presidents of the Society of Christian Philosophers. Robert Audi who is on that list also subscribes to some weakened version:

Audi has defended a position he calls "fallibilistic foundationalism." He thinks that the foundationalist response is the only tenable option of the epistemic regress argument. This states that if every belief has to be justified by some other, then the only options are four: infinite regress, circularity, stopping at a belief that isn't knowledge, and stopping at a basic belief that is itself justified. If the only alternative is the fourth, then if one has knowledge, one has foundational knowledge. Audi considers that foundationalism is usually taken to be infallible. That is, it is normally associated with the view that knowledge is founded on basic beliefs that are axiomatic and necessarily true, and that the rest of knowledge is deduced from this set of beliefs. Audi thinks that foundationalism may be fallible, in the sense that the suprastructure of beliefs may be derived inductively from the basic beliefs, and hence may be fallible. He also thinks that basic beliefs need not be necessary truths, but merely have some structure which makes epistemic transition possible. For example, the belief that one is in the presence of an object arises causally from visual perception.

I'm not sure where Robert Merrihew Adams stood on that as his wiki bio is a bit undeveloped (he's #37 in the SEP mention ranks though.)

The "weak foundationalism" of BonJour gets a honorable mention of sorts on IEP but mostly a subject of criticism from another modest foundationalist--Van Cleve. (Amusingly enough BonJour and Van Cleve "tied" on SEP at #179.)

IEP also mentions that reliabilism might fall under the broad tent of foundationalism:

reliabilism makes plausible a form of structural foundationalism which stops the regress of justification

Wikipedia [also] classifies reliabilism an externalist foundationalist account:

Reliabilism is an externalist foundationalist theory, initially proposed by Alvin Goldman, which argues that a belief is justified if it is reliably produced, meaning that it will be probably true. Goldman distinguished between two kinds of justification for beliefs: belief-dependent and belief-independent. A belief-dependent process uses prior beliefs to produce new beliefs; a belief-independent process does not, using other stimuli instead. Beliefs produced this way are justified because the processes that cause them are reliable; this might be because we have evolved to reach good conclusions when presented with sense-data, meaning the conclusions we draw from our senses are usually true.

And a number of SCP notables like Linda Zagzebski although declaratively responsibilists have argued closer to reliabilists:

Some virtue responsibilists have adopted an approach similar to that of virtue reliabilists by giving virtue concepts a crucial role in an analysis of knowledge or justification. Linda Zagzebski, for instance, claims that knowledge is belief arising from what she calls “acts of intellectual virtue” (1996).

(I don't understand this bit too well, but it seems to be some kind of argument that virtue is foundational or some such.)


I'll also note that SEP (unlike IEP and Wikipedia) uses the term "modest foundationalism" with a fairly different meaning, denoting some other efforts, not seemingly related to Christian philosophers:

Some contemporary epistemologists seek a more modest foundationalism that will make it much easier to respond to the skeptic’s arguments. Michael Huemer’s (2001) phenomenal conservatism and Jim Pryor’s (2000) dogmatism are both views that are far more “permissive” in allowing a far more extensive range of beliefs to have foundational justification. And their views are not unrelated to Chisholm’s longstanding efforts (e.g., 1989) to locate noninferential justification for believing various propositions about one’s past and physical environment in the character of one’s experiential states.

But otherwise (agreeing with Wikipedia on this) it also classifies reliabilism as an externalist foundationalist viewpoint...

While the externalist defends radically different views than those of classical foundationalists, the structure of knowledge and justification that emerges from such theories is still often a foundationalist structure. We might first illustrate the point by examining the view defended by the most prominent of the externalists, Alvin Goldman’s reliabilism. [...]

[However...] On most externalist accounts of noninferentially justified belief there are literally no a priori constraints on what might end up being noninferentially justified.

[And among the criticisms listed] Reliabilism seems to yield the wrong result, justifying too many beliefs.

Although Plantinga is very cited on SEP, (amusingly) he isn't cited on that page, but some other works quickly point out that Plantinga also resorts to a form of reliabilism but that he bases his on a warrant (from God basically).

Comparing reliabilism with the theory of warrant, Nicla Vassallo claims that “the difference between proper functionalism and reliabilism is slight, since the cognitive faculties that function properly are usually reliable and since, even from a reliabilist point of view, the cognitive processes considered are those that bring to truth” (Vassallo 1999, p. 131). [...] She sustains that with appropriate modification – essentially renouncing to the concept of proper function – Plantinga’s theory of warrant would be “compressed in Goldman’s sophisticated reliabilism”. [...]

Plantinga chooses Goldman’s reliabilism as the paradigmatic theory that is, in his words, “clearly compatible with the existence of God” (Plantinga 1996, p. 356). [...But] In order to accept Goldmanian reliabilism, Plantinga needs to take it to a higher level of reliability, for once in a Cartesian fashion, as only the Christian God can offer such level of warrant.

1
  • 4
    "Slightly" :) This is very comprehensive. It is worth noting that reliabilism shows that Audi's tetralemma is incomplete, a belief can be justified by a non-belief, a reliable process in that case. To add to your list, relativized Kantian apriorism that Friedman defends in Dynamics of Reason seems to me akin to the "fallible foundationalism", although he does not use the label. Fallible relativized a priori are not that far from properly basic beliefs, but have the advantage of being more traceable to accumulated empirical experience.
    – Conifold
    Mar 29, 2021 at 20:09
0

I'm not a published author or what, but I do have an argument or line of reasoning to offer on behalf of foundationalism. The reasoning is a response to the no-inference-from-nonbeliefs objection to foundationalism from experiences. The objection is that one cannot infer things from nonpropositional structures like bare experiences. So far, oh well. But don't some questions have propositional structure? So might there be things besides beliefs from which beliefs could be inferred? (Not that this supports foundationalism from experiences itself, as such, then, unless inquiry is a form of experience?)

Granted, not all questions have propositional structure, e.g., "Go to the store?" doesn't. But otherwise, if there were questions from which we could infer their answers... Self-answering questions, then, rather than self-evident answers (or: this is self-evidence!).

This reasoning depends on an underdeveloped erotetic logic, for starters, but it seems promising to me. Cases where this erotetic perspective seem helpful include thinking about the liar paradox, demarcating analytic from synthetic knowledge, playing the skeptic's game, and bridging the prescriptivism-cognitivism gap in metaethics. With respect to the liar paradox, for instance, the erotetic perspective has us convert the liar sentences into questions, like, "Is this sentence false?" which we then see is rather peculiar, as questions are neither true nor false, so the "this" at play doesn't seem to be intelligible after all.

This leads us to note that the liar sentence in question is seemingly unquestionable, which is surprising or absurd: certainly this liar sentence is not axiomatic! Wherefore this liar sentence is testified against such that it seems not to be a "good" counterexample to the law of noncontradiction. What I mean is that the erotetic perspective supports, "The liar paradox can be resolved," not, "The liar paradox is unsolvable."

Or consider questions like, "Why go to the store?" If something is why, then questions like this can take us from prescriptive to assertoric ("cognitive") functions. For, "X is why," is assertoric, wherefore... Or we can perhaps differentiate analytic and synthetic knowledge in erotetic terms.

Ultimately, what if even libertarian free will is grounded erotetically? What else but our pure ability to question everything, allows us to question not only every individual claim of cause and effect, but even causality in general itself? It would be axiomatic, then, that we have libertarian free will, as this will would be a presupposition of our faculty of inquiry. (Or so it seems to me.)

EDIT: to quote Wisniewski:

... we often pass from some "initial" question to another question.

The regress of reasons, if inferential, might be structured like so: inference sequences among answers retroterminate in fundamental answers, but the questions from which these answers are inferred form an infinitistic sequence in their own right.

4
  • The ability to understand and utter propositions (=language acquisition) has been shown to work holistically, so I do not really see the point of your wordplay with questions...
    – Philip Klöcking
    Mar 29, 2021 at 9:04
  • I suspect my position is closer to foundherentism with infinitism mixed in somewhere, than to strict foundationalism. At least, I suspect my position should be closer to such a mixture. Mar 29, 2021 at 12:26
  • 1
    But how does this help foundationalism? Questions are, arguably, propositions with a question mark, and games with the Liar can be played with or without putting it into the interrogative. Moreover, Liar, Truth Teller and the like are completely sterile as far as producing substantive knowledge that people in epistemology want to justify. If the strategy is to have propositions justified by non-propositions it would have to involve something much more non-propositional than questions, and much more substantive than semantic puzzles.
    – Conifold
    Mar 29, 2021 at 23:29
  • I need to reword what I said in the beginning... Mar 29, 2021 at 23:41
0

According to Wikipedia referenced here for foundationalism.

In the 1930s, debate over foundationalism revived.Whereas Moritz Schlick viewed scientific knowledge like a pyramid where a special class of statements does not require verification through other beliefs and serves as a foundation. In the 1950s, foundationalism fell into decline – largely due to the influence of Willard Van Orman Quine, whose ontological relativity found any belief networked to one's beliefs on all of reality, while auxiliary beliefs somewhere in the vast network are readily modified to protect desired beliefs.

Around 1975, weak foundationalism emerged. Thus recent foundationalists have variously allowed fallible basic beliefs, and inductive reasoning between them, either by enumerative induction or by inference to the best explanation. And whereas internalists require cognitive access to justificatory means, externalists find justification without such access.

Seems to me the key to be safe for modern foundationlists is to either have cognitive access to some kind of internal justificatory means of their foundations, or have some external reliable process to justify them. As long as your foundations can be internally or externally justified, you can still be called a (weak) foundationalist in modern times, even some may be turned out to be fallible.

Besides above non-ideal limitations, however, you may have other difficulties, like in physics (unlike mathematics), you can hardly call any law is more foundational than another. Is Newton's law more foundational than Lagrangian principle of least action in the realm of classic mechanics, though one law can be deductively transformed to another? If you stubbornly hold Newton's law as your epistemic foundation and try to apply it to QM, it'll be much less effective than Lagrangian's which later is successfully formulated as path integral approach in QM. So it may be controversial to rigidly and subjectively put a certain foundation simply per one's preference or chronology. That's why coherentism is on the rise these days without worrying about a hierarchy unnecessarily, metaphorically in line with holographic principle...

5
  • I guess you could call weak foundationalism "nonmonotonic" or "defeasible" in logic terms.
    – Fizz
    Mar 29, 2021 at 7:23
  • The problem here is that internal justification could work just as well with holism, ie. completely without foundational knowledge. Whether they are fallible or justifiable does not decide the question of foundationalism.
    – Philip Klöcking
    Mar 29, 2021 at 9:02
  • @PhilipKlöcking: yeah, "reliabilism" and foundationalism seem to be distinguished on this aspect, although some of the writeups (e.g. IEP) assert that reliabilism is a kind of "structural foundationalism" FWTW.
    – Fizz
    Mar 29, 2021 at 9:58
  • @Fizz In my book, reliabilism and holism still have to be distinguished in the sense of how we get to the reliable structures: It makes a huge difference whether the structures are just "apodictically" reliable or "justifiably" reliable, ie. whether they are just taken to generally produce knowledge or embedded into a multi-directional framework of beliefs justifying each other as knowledge/reliable process.
    – Philip Klöcking
    Mar 29, 2021 at 11:08
  • 1
    @Fizz more accurately speaking, we may say "nonmonotonic" or "defeasible" logic is a subset of weak foundationalism since I'm not sure we can correctly transfer all knowledge (like the intuitive knowledge to process image recognition) into a linguistic symbolic formal logic system. Mar 30, 2021 at 0:22
0

The Münchhausen trilemma is a key counter-argument against foundationism.
In epistemology, the Münchhausen trilemma, also commonly known as the Agrippan 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.

Analytic truth 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.

We don't merely assume that a "dog" is an "animal" and not an "office building", this is completely proven entirely on the basis of the meaning of the words: {"dog", "animal", "office building"}.

The dogmatic argument of the Münchhausen trilemma is refuted:
Stipulating relations between finite strings(words) is the only way that arbitrary finite strings(words) acquire any semantic meaning thus is not merely dogmatic. Language only acquires meaning on the basis of stipulating relations between otherwise arbitrary finite strings(words).

Although "a rose by any other name would smell as sweet" if we fail to assign any name to {a rose} then we cannot even refer to it.

The relationship between "dog" and "animal" is stipulated to be "is a type of". The only way to prove that a dog is an animal is through other stipulated relationships.

The foundation of analytical knowledge is based on stipulated relations between otherwise meaningless finite strings. The finite string(words) of "dog" and "animal" are utterly meaningless until meaning is assigned to them. This same process is how every element of language acquires any semantic meaning.

Meaning Postulates Rudolf Carnap 1952 -------> ∀(x) (Bachelor(x) → ¬Married(x))
The above shows how one element of semantic meaning of the otherwise totally meaningless terms of: "Bachelor" and "Married" is specified by their relation to each other. To prove that Bachelor(x) is ¬Married(x) we simply refer to this established axiom.

0

There exists a subset of language expressions that are necessarily true and impossibly false. This subset has a completely different character than all other expressions of language. Analytic truth is simply the set expressions of language known to be true that are connected together on the basis of their semantic meanings.

If we define knowledge as a fully justified true belief such the the justification necessitates its truth then a foundation of analytic knowledge might be derived.

Self-evidence In epistemology (theory of knowledge), a self-evident proposition is a proposition that is known to be true by understanding its meaning without proof...

"This sentence is comprised of words." is proved to be true entirely on the basis of the meaning of the terms: {sentence}, {comprised}, and {words} combined together to form the compositional meaning of the whole sentence.

The foundation of analytic knowledge can be understood as the semantic interconnections between the body of true sentences as demonstrated in the above example. These sentences are expressed using language and verified as completely true entirely on the basis of the meaning of these language expressions.

You must log in to answer this question.

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