Enjoy!

Please note that this is part three of a three-part discussion

Significantly-built into the notion discussed in part two- is the concept that the inferential transmission of our justification does not need to be deductive, but rather can be probabilistic or inductive. Epistemologist Robert Audi, in The Foundationalism-Coherentism Controversy, states, “Superstructure beliefs may be only inductively, hence fallibly, justified by foundational ones and thus (unless they are necessary truths) can be false even when the latter are true.” (Knowledge and Inquiry, pg. 97) Thus, although we may have very good inductive evidence for holding a belief, we ultimately might have to reject it because of the introduction of new refuting evidence. For, no judgement aimed at finding truth can ever attain apodictic certainty. So, in the absence of infallible knowledge, as reasonable beings, we should be satisfied with less than that. As such, within this framework, fallibly justified beliefs are likely, but not guaranteed to be true, and for this reason, mutual coherence may also play a key role in our inferential justification.
Accordingly, a large part of mutual coherence, and its provision of justification for our basic beliefs, will depend on employing the principle of independence. The independence principle simply states that the larger the number of independent and mutually coherent factors one believes that support a truth, the stronger one’s justification will be for believing it. So, it is not enough to determine that one is having the internal perception of red; one must also have an external and independent factor corresponding to the perceptual state. For example, let’s say that a knower sees a red apple on the table:
(P1) I have a perceptual belief that I see a red apple (justified via introspection)
(P2) There is a red apple (justified via independent external factor)
(C1) Therefore, I have a true perceptual belief that there is a red apple
In this example, the knower infers deductively that their perceptual belief is reliable. And, since both P1 and P2 are justified, they jointly entail C1.
Conclusively, the individual appeals of foundationalism and and coheretism may lie in the fact that they provide answers to different questions. Or, it may simply be that we should let go of our traditional epistemological labels. Either way, the notion of fallibilist foundationalism seems superior to either ideology taken alone in isolation because it can give us an attainable theory of knowledge through an account of coherence as justification. Additionally, by making our theory of knowledge plausible, it avoids the impossibility of the skeptic because it grounds our beliefs in observation and past experience. So, again, although our perception may decieve us, it does not follow from this that we should abandon the realities of our senses altogether. Audi states, “In working from the experential and rational sources it takes as epistemically basic, fallibilist foundationalism (in its most plausible versions) accords with refective common sense: the sorts of beliefs it takes as non-inferentially justified…are pretty much those, that, on reflection, we think people are justified in holding.” (Ibid, pg. 123) As such, can we allow into our theory of knowledge, as privileged basic beliefs, those beliefs that are inferentially justified? Or, moreover, is it more accurate to question whether or not our immediately justified basic beliefs are infallible. Accordingly, if the answer to either of these questions is yes, should we further question whether or not our justification is strong enough?
I would argue that, as humans, the notion of achieving infallibilty is both impossible and unrealistic. The Cartesian, who proposes the goal of indubitability, ultimately ends up rejecting all inductions we make of the external world, which then leads us to a state in which we possess virtually no knowledge. And, clearly, the practice of making inductive inferences of the external world is reasonable and does constitute knowledge. Chisholm states, “It is difficult to think of any claim to empirical knowledge, other than self-justifying statements we have considered, that does not to some extent rest on an appeal to memory.” (Ibid, pg 24) Yet, although we must accept our fallibility as unavoidable, we should, perhaps retain the Cartesian concept as an unachievable ideal. Or, maybe the role of coherence in any theory of knowledge, simply becomes the negative role of incoherence. And, finally, perhaps it is enough that our beliefs, which arise from perception, memory, introspection and human reason, generate outcomes that are least likely to be true.

Please note that this is part two of a three part discussion

Empirical foundationalist Roderick Chisholm posits in The Myth of the Given that memory and perception are fallible and are some what a matter of coherence. Yet, he also thinks they belong to the class of justified basic beliefs. Why? Because we are human, our perception, our memory, and our reason are all imperfect. However, when we fall into error regarding the objects of sense, we can generally correct our error by acquiring more accurate information, or by making an appeal to our prior experience. So, subsequently, Chisholm believes that memory and perception seek to provide bridge principles from our first-person epistemic or internalistic beliefs, to our justified external beliefs of the world. As such, if an external world proposition is accepted by one, and the rest of one’s beliefs do not go against it, then the proposition in question is made probable. And, consequently, accepting any external world proposition under these circumstances gives it pirima facie probability. So, what seems like an initially plausible case could eventually prove to be implausable given more evidence. For this reason, prima facie justification is always defeasible. However, it is also for this reason that Chisholm states, “We cannot say precisely what is meant by ‘fitting in’, ‘coherence’, ot ‘congruence’ until certain controversial questions of confirmation theory and the logic of probability have been answered.” (Pg. 24) And that, moreover, a foundationalist theory that includes memory and perception as regress stoppers can never merely be “a system of blind posits.” (Pg. 11) Our probable perceptions must, at some point, be grounded in some type of certainty and not guesswork. So, although Chisholm’s theory may be a step in the right direction, his version of fallibilist foundationalism does not entirely commit to the viability of induction, or probabilities based on probabilities.

So, the notion of fallibilist foundationalism that I wish to defend is based on the understanding that our basic beliefs can be fallible, and that, furthermore, the links in our epistemic chain can be connect to a final link that is false. Since humans are prone to error, it is always possible to discover a lack of justification somewhere in our hierarchical framework, and the foundation is no privelaged exception. Importantly, this requires an undeniable commitment to induction. So, accordingly, if we can allow our basic beliefs to have both immediate and inferential justification, based on the idea that our meta structure is foundational, then it follows that its content will be inferentail. It simply makes sense for people to believe what the perceive and experience. Realistically, should we not trust our eyes and sense of sight; or ears and sense of hearing? Indeed, our senses can sometimes be deceiving. However, does it necessarily follow from this that appearances and reality mostly diverge? This skeptical argument seems like a theoretical artifact, since we can generally make the distinction between what ‘appears to oneself’ and what actually ‘is’. So, in this regard, it is possible for our true premises to lead to a false conclusion, and for our knowledge to be defeated by the discovery of some new counter evidence. Yet, this does not entail that we lack either justification for our beliefs or, more importantly, a theory of knowledge.

Still, problems can arise when we rely on this type of psychological sense datum. Can a proposition in the form “I am being appeared to” be considered foundational if understanding the proposition requires antecedent knowledge? And, if there is a lack of antecedent knowledge, can we really trust what we know from our senses? Although it may be the case that no belief is ever more than conditionally justified, this does not mean we must forgo antecedent security. On the contrary, the sensible content of our past experiences will provide what is necessary and sufficient. And, in cases where a perceptual error does occur, we can generally correct that error by making an appeal to our existing stock of knowledge. For, all truths of the natural world are discovered through ampliative induction- it is simply unavoidable. So, as an example, we may not be able to see that the inside of a banana is soft and mushy when we examine it from the outside, but we know that from peeling back the skin on previous bananas, that the fact that this banana will be soft and mushy on the inside, is more than likely true. It would be ridiculous to anticipate otherwise.

Please note that this is part one of a three part discussion

Foundationalism dictates that knowledge, in the form of justified true belief, must be grounded in immediately evident, or non-inferentially justified basic beliefs. It therefore condones a form of knowledge that is hierarchichal, with epistemic chains arising from these basic beliefs. Coherentism, on the other hand, holds that all justified belief is a matter of structural inference, and that increasing coherence de facto increase justification. As such, this framework condones a web of beliefs that is ultimately justified through inferential loops. Yet, although these two theories may seem incompatible on the surface, is that really the case? If we can formulate a theory of knowledge that combines a foundationalist structure, with a form of coherentist justification, we can formulate a theory of psuedo-fallibilist foundationalism. For, if sense data, perception, and past experience are significant sources of our basic beliefs, and if we have immediate and not inferential justification for them, than we must have a theory of knowledge that bottoms out with fallibilism. So, perhaps, for this reason, we can view the traditional conception of knowledge as an ideal, and coherence theory as a pragmatic and useful tool. And, together it may be that they form a more commonsensical view of epistemology. But before we explain how this conflation may work, we must first understand the origins of these two apparently divergent views, and why they are plausable solutions to the epistemic regress argument.

The coherence theory of justification, from the holistic point of view, reduces the source of all justification to logical consistency, symmetry, and a lack of anomalous beliefs. As such, all claims are mutually explanatory, and a proposition is justified IFF it is a member of a coherent set. So, for this reason, coherence is always the logical property of a set and never a single proposition in isolation. As such, if we assume an empirical property E, we can discover if E is justified, by comparing E to the entire belief set. If the entire set is more coherent when E is added, then E is justified. Yet, although the clearest example of incoherence may be inconsistent belief, mere inconsistency in and of itself does not necessarily provide justification enough. So, as a consequence, the comprehensiveness requirement is deeply important. The comprehensiveness requirement is a form of completeness that incorporates empirical data. As Such, the main attraction of coherence theory lies in the core idea that, if a system of beliefs coheres, and if it is also comprehensive, then it is at least likely to be correct. Subsequently, the descriptive is the prescriptive.

So, how does a holistic coherentist address the epistemic regress problem? From this perspective, the argument never gets off the ground, since a coherentist outrightly rejects the notion that all justification is linear or hierarchichal. Anf, for this same reason, the coherentist can avoid endorsing a complex form of circular reasoning, since a web of belief that is non-linear, simply circumnavigates the problem’s existence. Since circular justification requires that we move in an epistemic chain from a conclusion through premises, and then back to the conclusion as a premise, if we deny that there is a chain, we deny that there is either a regress or circularity. Consequently, coherence employs a concept of symmetry- if p justifies q; it is okay that q justifies p, as long as there are no anomalies. Each will be a small part of the justification for the other. But if this is the case, what grounds the web? Or, for that matter, does our web even need to be grounded? If we posit that our web of coherence must utilize some means of support, what will ensue is a form of fallibilist foundationalism.

Some philosophers that endorse coherence theory, such as Dancy, additionally posit that when we are maximizing coherence we are de facto maximizing truth. But, clearly, this is where some trouble can occur. For, if we define objective truth as coherence, we risk buying into a worrisome form of epistemic relativity. Clearly, a system of coherent beliefs, no matter how coherent, does not necessarily equate to objective truth. If we think that this is the case, we are missing the boat on what truth objectively is. Dancy also believes that there is nothing sacrosanct about any basic belief, and that all of our justification is a matter of structural inference. The problem here is that our web is left unanchored and floating. So, although this view may allow for the constant revision of a belief set, it simply goes too far. Consequently, although Dancy’s version of coherence theory may be a pure and pragmatic approach to the theory of knowledge, the idea that there are no priveleged basic beliefs, or that our web is not in need of grounding, renders it unworkable.

In order to motivate this argument further, however, we must first explicate the structural claim that foundationalism makes in regards to our priveleged basic beliefs. Traditional foundationalism views knowledge as an epistemic chain of justification, that concludes in a final non-inferential, or immediately justified belief. As such, these immediately justified basic beliefs can include a priori statements, self-evident beliefs, or first-person cognitive experience. And, importantly, in order to stop an infinite regress, they must be self-justified or justified by something other than a proposition. So, in this respect, we can view knowledge as an ideal cognitive state that contains both certainty and justification at its foundation. with a solid structure for our epistemic inquiry built on top. Consequently, reason, as a reference to the Cartesian viewpoint of clear and distinct perception, becomes the perception that allows us to determine what knowledge is. Yet, a large problem with this perspective lies in its unattainability. It is obvious that the struggle for indubitability can ultimately lead to the possession of virtually no knowledge whatsoever. Nonetheless, the simplistic fact that ideals may very well be unattainable does not mean that we should abandon them either. After all, is it not the case that most of our human ideals are rarely acheived?

An additional argument in favor of foundationalism is that it more clearly responds to the problem of an epistemic regress, since instead of circumnavigating the issue, it tackles it head on. The argument basically assumes that since we have justified basic beliefs, the ultimate justification for those beliefs cannot be produced by circular reasoning, and therefore the chain of reasons must terminate in other non-inferentially justified beliefs. So, following, no proposition ever plays a role in justifying itself because the structure is asymmetrical. In other words, if P justifies Q, Q cannot justify P. Thus, by denying the claim that justification is always inferential, the traditional foundationalist requires that our basic beliefs be a priori or self-evident facts of the matter. Yet again, as with the aforementioned indubitability, are we aiming too high? And, more importantly, are we stopping the regress at the risk of losing what it is to engage in humanistic inquiry and reasoning?