Here's a stripped-down view of how beliefs are justified. (See part one here.)
One view is that there is a single basis for all beliefs. Anything you believe can be justified by reference to a foundational belief, such as belief in a deity or even belief in one's own existence. This view, relying as it does on a single foundational belief, is called foundationalism, and the test of justification is how well a belief accords with the foundational belief, which cannot be called into question. (Religious belief is foundational, and foundationalism is basically a dogmatic position, wherein the foundational belief cannot be called into question without all other beliefs collapsing like a house of cards.)
Foundationalism clearly doesn't work, because any foundational premise can be disputed and turn out to be wrong. For instance, there don't seem to be any deities, and basing the justification of your other beliefs on the premise that a deity exists is surely not a good idea if their existence can be called into question. This implies that people don't base their beliefs on a single foundation, and, instead, have a different and more basic epistemology. Moreover, it isn't often clear how these foundational beliefs justify
other beliefs. How does belief in a deity justify the belief that water
consists of hydrogen and oxygen atoms?
So foundationalism is out, although some people refuse to accept this. The theologian William Lane Craig frequently tries to justify belief in god on the basis that without that belief, all other beliefs have to be called into question. Lane Craig seems to believe that without belief in god, you can't have a conversation or like music. This is self-evidently wrong, and, of course, Lane Craig never justifies why appreciation of music depends on god.
It's also a common misconception that science relies on a foundationalist epistemology, but it doesn't. We'll come to how it works a bit later. But what's important is that attacks on the scientific enterprise by various figures are unfounded if they attack the supposedly 'foundationalist' basis of it (as does, say, Stanley Fish), because science doesn't depend on any single belief.
Another view is called coherentism. It takes the idea that no belief is foundational and runs with it. Instead of all beliefs relying on a single basis, coherentism takes it as read that beliefs support one another, and that epistemology has to consist of understanding how beliefs can coherently support one another within a closed system of beliefs. If there's no foundation, then the belief that rain is wet is supported by other beliefs that you have about rain, related not in a hierarchy of beliefs (a la foundationalism) but in a circle. The test of justification is how well a belief fits in with other beliefs that you already hold.
This doesn't cut it either, for another obvious reason: if beliefs are sorted into closed systems of self-supporting ideas, then how does new information intrude? On what basis can new information override previously-held beliefs, if justification of belief is held to reside in the belief's ability to fit in with other beliefs? That would imply that experiment and evidence aren't important, except to people who already believe that they are the criteria of justification.
Coherentism is clearly not a good position, and it can lead to extreme epistemological relativism - the view that any belief is true if it is justified by the person's belief system.
Before I get onto the next logical step in advancing epistemology, I want to point something out here: neither of these epistemological views can be overridden. If you genuinely believe that there is a single foundational belief that is necessary to justify why rain is wet and why oats are good for you, then you can disregard all the evidence and logical connections between beliefs, because you already believe that you know the truth, and that it boils down to one essential feature. And if you genuinely believe that there are no 'true' beliefs and no independent justification of belief (which also relies on a certain ontological view - I'll get onto that in another post), then no amount of evidence and reason can convince you otherwise, because you can simply wave your hand and say that it comes from a particular school and view of justification, and, like any other belief, has no independent justification. So they're both annoyingly circular positions.
The latter, of course, is found most commonly in the academy - no normal person could hold such a silly view - and it is a favourite of critical theorists who, in their daily lives, rely on a more realistic view of the world than the one they profess to hold academically. The former is most commonly found outside of the academy, and is a particular favourite of religious people.
So what's the resolution? Well, the British philosopher Susan Haack, now at the University of Miami, has proposed a third distinct approach to the problem of the formal justification of beliefs. Instead of a circular coherentism or a baseless foundationalism, Haack has proposed that beliefs may be justified like the choice of letters and words in a crossword. No individual answer on a crossword can be taken as guaranteed. It may appear to fit the spaces on the puzzle and it may appear to answer the related question satisfactorily, but new information may come to light that overrides the previous choice. In trying to answer another question, you realise that another word would fit the question and the spaces even better than your former answer, and so you replace it.
All the answers are provisional - they could be shown to be wrong by new evidence at any point. But as the answers pile up and as they continue to support one another, they become increasingly better supported, even though a new answer could potentially cause a rethink of any or all of them.
The difference between the attempt to understand the universe (from a human position) and a crossword is that crosswords are finite and you can get all the right answers. In science, we can keep going and going, and we'll never know if our answers are more than best guesses justified by the evidence we've got (instead of, I guess, metaphysical truths). But this approach - which Haack labels foundherentism, a neologism composed of both coherentism and foundationalism - allows us to believe that beliefs can be justified independently through investigating the world in some way, and that research can disprove or corroborate existing ideas, without holding to the view that all beliefs have to be justified by an unquestioned foundation.
That's how science works, and it is also how we justify ourselves in ordinary life. A lot of people seem to believe that science depends on an a priori belief in the power of reason and in the truth of naturalism, but it actually doesn't. It's just that those beliefs come about through the progressive attempt to solver lower-level problems, correlating the knowledge attained through this process in a similar manner to a crossword puzzle, seeing where certain new answers don't fit and overturning older answers on the basis of better-justified, newer information.
The belief that water consists of atoms of hydrogen and oxygen isn't justified by recourse to a foundationalist claim about the nature of the universe. It's justified by experiment after experiment showing that no other answer fits with either new experiments or old data, and by the fact that no other model permits a better understanding of water and its properties. And the belief that experiment is a criterion of truth is justified by the fact that it tends to work to resolve problems, and that we have no real reason to doubt it. Either of those beliefs could turn out to be untrue, but what we need is a reason to believe it - some piece of evidence that shows why those specific beliefs are wrong. It's not enough to say that scientists are dogmatic in proposing the material nature of water if all of the actual relevant evidence says that water is material.
Let's say I have an argument with my girlfriend about whether I cleaned the dishes yesterday. She might say that I didn't, because they weren't clean. I distinctly remember cleaning them, but maybe I'm mistaken. If I take my memory - possibly flawed - as foundational, then going to check the dishes isn't going to help, because I already know that I cleaned them. If I take a coherentist position, then my beliefs are justified by my own worldview (dependent on the memory of cleaning the dishes) and her beliefs are justified by hers (dependent primarily on the fact that the dishes aren't clean), and neither of us can be 'wrong', because our systems of belief justify our beliefs about the situation. (This is why coherentists of various stripes place such an emphasis on ideas like the 'paradigm', because a paradigm shift would mean a radical rethink of the very basis of the universe.) But if I take a sensible position and asssess the actual evidence, then maybe I have to rethink my belief in the power of my memory.
Then again, if I go and check the dishes and they turn out to be clean, then maybe I don't have to do that. Or, they could be dirty, but if I had reason to believe it, I might suspect my girlfriend of making them dirty in order to find a reason to be angry at me (which I suppose is possible). The important thing is that we look at the different reasons we have for believing things about the world, taking no belief as foundational.
Foundherentism also means that 'scientific' facts are not different to other kinds of facts. A scientific fact is just a fact. Science is just an extension of how people already reason, not a different sort of activity. No one is actually a foundationalist; if William Lane Craig lost his belief in god, that wouldn't necessarily mean that he would suddenly hate Wodehouse and curl up in a nihilistic ball, because he is a foundherentist in real life, just like everyone else, and his belief in god is fundamentally unconnected to the enjoyment he gets from reading and listening to music. If you accept that the universe exists, is reasonably well-ordered, and that logic dictates certain conclusions - as you must do if you are alive, reading this, and not insane - then you have to accept the conclusions of science, as they are justified in the same way you justify every belief you've ever held, and follow from the same basic ideas.
In popular archaeology, you come across a lot of people who claim that the academy is dogmatic in its rejection of, say, the 'ancient alien' theory. They say that archaeologists are dogmatists who rigidly pursue a naturalistic, materialistic, scientific agenda and aren't open-minded enough to accept the truth: that ancient astronauts came down to earth and inspired human civilisation globally.
But that's not how it is at all. How it works is this: not only does ancient alien theory not fit with any of the existing evidence - and so doesn't cohere with the existing views of the scientific community - it also fails to provide any reason why it is a superior model to the one we already have, and so doesn't justify rethinking all the previous answers on the crossword. What data does it explain that the current, realistic, scientific, view doesn't? What reason does it gives for us to overturn all of the other evidence we've got, or at least drastically reassess it?
Archaeologists aren't being dogmatists in saying that Borobudur wasn't built by aliens. They're just going by what all the evidence says - not depending on a naturalistic worldview to support the belief, but rather accepting it as a matter of course. And if you've got a reason to believe something that is supported - not foundationally, but foundherently - by a large set of other beliefs and data, then it isn't dogmatic to say that some ideas are wrong. It's just how you deal with beliefs that are wrong, even if that assessment can only ever be provisional.
And a final point: it could be claimed in response that such an answer depends on the a priori - foundational - belief in evidence. In this view, science can't independently justify anything, because it depends on the foundational belief in evidence, itself unjustified. But this isn't true. All evidence means is 'a reason to believe something'. It's nothing more than that, and you can't get rid of the idea of having a reason to believe something, even if you adopt a coherentist position.
Next time: Ontology and epistemology - how they relate to one another, and the presupposition of the existence of reality.