How to Seek Knowledge

If the different ways that we justify our beliefs are the building materials for a structure of beliefs, then perhaps I can call the process by which we evaluate and adjust our beliefs the scaffolding for that structure. That is the topic of this post.

First, though, I want to make a couple of observations about the principle of critical trust that I presented in the last couple of posts. The essence of this principle is found in the two intellectual virtues that I explored earlier, critical thinking and epistemic faith. In using this principle, we extend initial trust to the beliefs that we naturally form from conscious experience, intuition, and testimony. Then we sift through those beliefs to try and discard the ones that are false, and strengthen the ones that are true.

The first observation I want to make here is that the principle of critical trust is not an “anything goes” principle. It does not mean we can believe whatever we want. Rather, it means we start by believing that things are the way they seem to be, on the basis of our conscious experience, intuition, and testimony – these initial beliefs must still have some kind of justification – and then we use reasoning to try and figure out the way things actually are.

The second observation is that, although the justification for these initial beliefs is not free, it is not all that costly either. We can usually find justification for what we want to believe, if we look for it. That is why it is important to weigh the justification both for and against a belief, rather than just looking at one side.

As a third observation, a corollary of the principle of critical trust is that when a belief is defeated by reasons to disbelieve it, it is rational to retain as much of the content of the belief as possible. For example, if you think you recognize your friend John on the street, but then later find out he was not in the city, you wouldn’t immediately think you had hallucinated the whole experience. You would think you had just seen someone who looked like John.

Worldview Formation

When I talk about constructing a belief system, what I principally have in mind is the formulation of a worldview, a comprehensive theory or perspective for categorizing, explaining, and interpreting all of reality. A belief system is essentially a worldview and the outworking of that worldview in the way that it interprets the world.

Everyone has a worldview, whether consciously or not – we all go through the process of creating an explanation and interpretation of everything we experience. We don’t have the capacity to deal with all of the diverse truths about reality independently. A worldview allows us to deal with all these truths in a unified way.

So, the principle of critical trust, and the need for a worldview, together suggest the following approach for forming, evaluating, and adjusting our whole belief system.

  • Presumptive Data: as presumptive data, accept that things are the way they seem to be, based on the sum of one’s conscious experiences and intuitions, and the testimony received from others. All of these data have a small amount of initial justification.
  • Initial Sifting: some of the presumptive data are likely to conflict with others. Sort the data into maximal consistent subsets (sets that contain as many beliefs as they can without containing a contradiction), and believe the subset with the most initial justification (rejecting the beliefs that conflict with this subset).
  • Worldview Formation: based on the beliefs accepted from sifting, form a worldview that explains as many of one’s beliefs as possible (principally, using inference to the best explanation).
  • Feedback Sifting: based on one’s worldview, some of the initially accepted beliefs may be rejected, and some of the initially rejected beliefs may be accepted. New presumptive data may be brought into consideration from re-interpretation of one’s experiences, intuitions, and received testimony in light of the new worldview. These two principles are important here:
    • Principle of Epistemic Defeat: reject beliefs that are in conflict with the best explanation of a vast number of other accepted beliefs. When beliefs are defeated in this way, retain as much of their content as possible that is undefeated.
    • Principle of Epistemic Enhancement: when an accepted belief is coherent with or suggested by the best explanation of a vast number of other accepted beliefs, its justification is enhanced.
  • Re-evaluation: New data or new understanding of old data may challenge one’s worldview, eventually forcing a re-evaluation and re-formation of the worldview when the data that cannot be fit into it acquires enough weight.

By going through this process, forming and re-evaluating our worldview and the whole system of beliefs associated with it, we will (hopefully) come to believe more and more truths, and disbelieve more and more falsehoods, in a validly justified way. That is, we will come to acquire more and more knowledge.

But the key to this process being successful in the long run is having the right worldview. If your foundational beliefs are true, you are more likely to interpret the data correctly, leading to further knowledge. If your foundational beliefs are false, your interpretation of the data is more likely to be wrong, impeding the search for knowledge.

Pilate’s Question

So I think I have adequately covered the nature of knowledge, and when it is reasonable to hold a belief. Now I come to a question that has been hovering in the background for a while. Belief is an attitude of accepting a proposition as true. Knowledge is validly justified true belief. I have claimed that it is important to hold a true worldview. And all throughout, I have been upholding the value of rationality – but the value of rationality is just that it allows us to know what is true.

But what is truth?

That is the topic I will turn to in my next post.

Building Materials (IV): Intuition and Testimony

So far I have investigated three different building materials that we can use to construct a belief system. In this first two posts on this topic, I examined deductive reasoning and abductive or inductive reasoning.

Then, last time, I suggested that it is rational to accept the principle of critical trust regarding our conscious experiences. According to this principle, we extend initial trust to the beliefs that we naturally form in the context of those experiences, and then we sift through them (using criteria of consistency and coherence) to try to discard the false beliefs and strengthen the true ones. One corollary of this is that we accept well-established types of experience to be grounds for properly basic beliefs.

Now I will look at the remaining two ways that we can justify our beliefs, and suggest that we should adopt an attitude of critical trust towards them as well.


Intuition is the capacity to acquire knowledge without conscious reasoning. In the last post I used an example to show that we naturally form certain beliefs in the context of our experiences, without using any reasoning methods to infer those beliefs from other beliefs. Intuition covers the beliefs we naturally form that are not directly grounded in conscious experience or inference.

Consider the fact that 1+1=2. When you understand the concepts of addition and the numbers 1 and 2, you just “see” that 1+1=2, and that it must be the case that 1+1=2. It is impossible for 1+1 to not equal 2. Or consider the laws of logic. When you understand the meaning of a statement like “if P is true, and it is true that P implies Q, then Q is true”, you can just “see” that this must be true.

Or, for an example that I have actually been using, when I consider the processes of reasoning or forming beliefs through conscious experience, I can just “see” that these are valid ways to acquire knowledge, if the right conditions are met. And I can just “see” that unless at least some of these cases of “seeing” are also valid ways of acquiring knowledge, then none of our knowledge can ever get off of the ground.

In this way, intuition supports itself – not through circular reasoning, where a belief is used to justify itself, but through a method of knowing without reference to other beliefs, where one of the things known through this method is that the method is valid. It is rational to accept intuition, because nothing else can be rational without it. We need intuition to come to know, and know how to use, all the other ways of justifying our beliefs.

It is also rational to accept intuition for the sake of being consistent. The line between intuition and conscious experience, and even the line between intuition and reasoning, is somewhat blurred. Many intuitions (though not all) seem like the result of an unconscious process of reasoning from the sum of many experiences and beliefs. So if reasoning and experience can produce knowledge, it seems reasonable to think that intuition can as well.

So I believe that the principle of critical trust encompasses beliefs we form through intuition, as well as beliefs formed through experience. Intuitive beliefs are often thought of as being self-evident. This does not mean that they are infallible – just like experiential beliefs, intuitive beliefs can be wrong, and they can be influenced by the way we interpret them. But like experiences, intuitions can be grouped into types based on what the intuition is about, and well-established types of intuition are grounds for properly basic beliefs.

Rational intuitions, which are about reasoning and the conditions for truth and knowledge, like the examples I gave above, are a well-established type of intuition. After all, a significant number of people throughout history have believed that they knew things, and had some idea of how they knew those things. And far more than providing a coherent picture of reality, these kind of intuitions appear to be necessary to have any coherent picture of reality at all.

Besides rational intuitions, a list of other well-established types of intuition could include:

  • Moral intuitions, about what is right and wrong,
  • Aesthetic intuitions, about what is beautiful or pleasing,
  • Interpersonal intuitions, about how other people would feel or act in a given circumstance,
  • Physical intuitions, about the everyday workings of the world around us, and
  • Metaphysical intuitions, about the fundamental nature of reality.

And all of these will come to play in some way or another as I build my belief system.

So, a belief formed through intuition is at least weakly inherently justified, and it is justified, and therefore properly basic, if the intuition is of a well-established type. Furthermore, such a belief is validly justified if the intuition is such that the belief it forms is appropriately connected to the truth (for example, if the cognitive faculties that produced the intuition are working properly).


Testimony is a person’s statement or assertion about what is true, a way of transmitting knowledge from one person to another. As a source of justification for a belief, it is “taking someone’s word for it.” Since we can only know so much through our own experience, we rely on the testimony of others for significant amounts of our knowledge. If you ask someone on the street for directions, or believe something because you heard it on the news or read it in a textbook, you are relying on testimony.

Though it is undeniable that the acceptance of testimony as justification for belief is ubiquitous and indispensable, how it is that beliefs can be justified in this way is an open philosophical question, which is why there are articles like this one here. The most common positions on this matter seem to be:

  1. Testimony is justified through a general abductive or inductive inferential principle, itself justified on the basis of experience and intuition, which is instantiated for each case.
  2. Testimony is justified a priori, no supporting reason needed.

It is certainly not the case that people consciously reason out whether each instance of testimony that they encounter is reliable, so the first position isn’t quite right. But since a reason to justify the use the testimony is available, it doesn’t seem like we need to jump all the way to the second position. I think a reasonable middle position here is that most of the time, we intuitively believe that we can extend critical trust to the testimony of others, and this intuition comes from an unconscious process of reasoning, similar to the principle postulated in the first position. And sometimes, we use that principle consciously.

A general principle to justify our use of testimony would be derived from premises such as the sum of one’s experiences with testimony, the fact that testimony is ubiquitous and indispensable, that it is often reliable, that the social norms of one’s community are generally structured so as to encourage reliable testimony, and that the sources of testimony themselves are generally rational and seeking to have justifiable beliefs. From these it would infer, absent sufficient reason to believe otherwise, that the source most likely reported the testimony in question because he or she believed it, and had sufficient reason to believe it. From such an inferential principle, it seems reasonable to conclude that testimony is capable of transmitting knowledge.

With such a principle, we can treat beliefs based on testimony as if they are properly basic. So, a belief based on testimony is justified, absent sufficient reason to disbelieve it, or to disbelieve that the source of the testimony is reliable. Furthermore, since testimony transmits knowledge, such a belief is validly justified if the source of the testimony knows that belief, and the knowledge is communicated in an appropriate way. (Testimony may be communicated in an invalid way if it is misunderstood, for instance; or if it fails to appropriately track the truth, such as if the same testimony would have been reported even if it were false.)

That concludes my survey of our building materials. In the next post I will briefly sketch how we put these materials together to construct a working system of beliefs.

Building Materials (III): Experience

In my last two posts I looked briefly at the different forms of reasoning, the ways that we justify beliefs using other beliefs. But as we found in my post on the regress problem, for us to be able to have any justified beliefs at all, we need to have some beliefs that are properly basic, that is, justified without reference to other beliefs.

So now I will look at the different ways we form properly basic beliefs, starting with experience.

Conscious Experience

In this context, experience refers to our conscious subjective perception of reality. All of us, pretty much constantly, take our perceptual experiences to directly justify beliefs about the world around us. For example, imagine that you walk into your kitchen one morning, and instead of seeing the table that you clearly remember being there last night, what you see is a small pile of sawdust. You immediately freeze, feeling quite incredulous, but as you walk into the space where the table should have been, and pick up a little bit of the sawdust in your hand, you form a number of different beliefs, in which you are quite confident.

  1. You believe that your table was here last night.
  2. You believe that your table is not here now, and it has been replaced by this pile of sawdust.
  3. You believe that you are presently in a state of great surprise and bewilderment.

If someone were to ask you how you knew these things, you would say, probably looking at them with some incredulity as well, “Because I remember the table being here, and I can see that it is not here now, and because I feel surprised and bewildered, of course!”

This imaginary example illustrates that, on the basis of our conscious experiences, we naturally form beliefs about the world around us, and we take those beliefs to be justified. When we see a tree, we naturally believe that there is a tree in front of us; when we feel happy or sad, we naturally believe that we are in fact experiencing that emotion; when we remember an event, we naturally believe that the event happened. And these beliefs are formed without any conscious inference from other beliefs.

In this example I have also illustrated that our experiences can be divided into three important subsets. These are external experience, introspection, and memory. External experience is any present conscious experience of something external to ourselves, and it of course includes sensory experience of the physical world. Introspection is our experience of our own thoughts, emotions, and other mental states. And memory, of course, is how we access our past experiences.

These three subsets of experience are not really different kinds of experience, however – for the most part, they are only differentiated by where the experience is directed. Certainly, our sensory experience of the physical world feels external in a way that introspection and memory does not. But internal mental experiences like dreams can also feel external, and when you think about it, our external experiences are still experienced internally in a fundamental way. Besides, at least to me, my introspective experiences feel closely woven together with my external perceptions. They can perhaps be separated in principle, but it seems more appropriate to handle all my conscious experiences as a seamless whole.

Two things should be noted, then, about our conscious experiences. The first is that they are fallible. The beliefs that we naturally form in the context of experience can be wrong. You can have a dream and mistake it for reality, or your memory can fail you and make you think the car you saw outside earlier was blue when in fact it was green. Even if we take these beliefs to be properly basic, they can be defeated by reasons to think they are false, or reasons to think the experience was not a reliable ground for true beliefs.

The second is that our experiences are theory-laden. The beliefs that we ground in experience (and even the experiences themselves) are influenced by our way of interpreting the world. Often this is helpful; for example, if you are familiar with the way that light is refracted by water, you do not form the mistaken belief that the straw in the glass of water is bent. But it is possible for our interpretation of things to distort experiences in the wrong direction as well.

The Problem of Skepticism

“Every man feels that perception gives him an invincible belief in the existence of that which he perceives.” – Thomas Reid

We do naturally take our perceptions to be good grounds for belief. But our perceptual beliefs are not quite as invincible as Thomas Reid here makes them out to be – sometimes our senses are deceived. Given that our experiences are fallible and theory-laden, how can we trust them? How can they really produce knowledge? What if all of our experiences are illusory, or part of a simulated reality like the Matrix? This quandary is known as the problem of skepticism.

My first response to this is to say, again, the mere fact that we might be wrong is not itself a good reason to think that we are wrong. And in order to know that we might be wrong, we have to accept that there are some things that we know – if not from our experiences, then at least from our intuitions about what is possible. But if we can accept on intuition that we might be wrong, we can accept on intuition that we might be right in the beliefs that we ground in experience, and that it is reasonable to believe we are right in these cases unless we have evidence otherwise.

My second response is that, when we are having an experience, the one belief we cannot be wrong about is that we are having that experience. Skepticism about whether our conscious experiences accurately reflect reality is one thing, but skepticism about whether we actually have those conscious experiences at all seems to be the height of irrationality. It just seems impossible to me that we could be wrong about that.

So I think that we have at least one kind of knowledge that is completely certain – the knowledge that we have the conscious experiences and memories that we presently have. From that, some philosophers have argued that we can use abductive reasoning to infer that our experiences do reflect reality and are generally reliable, as the best explanation of why we have the experiences that we do. On the other hand, some philosophers disagree that we can make this inference successfully, even if they think we are still rationally justified in trusting our experiences.

Whether the above inference is successful or not, I think we can definitely infer that our experiences reflect some kind of reality, whether it is a real physical world, or an illusory world. In that case, we can still know things about what is going on within our world, even if we are unsure about its ultimate nature.

For example, if we lived in the simulated reality of the Matrix, we would be wrong in thinking that our everyday beliefs were about real physical objects, but we could still have knowledge about those objects. If we were brought out of the Matrix, we would realize that what we thought were physical objects were actually simulated objects, but that would not mean that our beliefs about the relationships between those objects within the simulated reality were false. It would merely mean we were wrong in our underlying metaphysical beliefs.

The Principle of Critical Trust

That being said, I believe we are rationally justified in trusting our experiences, absent any good reasons not to. That is, we can treat our experiential beliefs as “innocent until proven guilty.” This is called the principle of critical trust, and it appears to me that it is better supported than any skeptical position.

The skeptic has to justify his skepticism from somewhere – he has to at least trust that we have experiences to think they can be wrong, and he has to trust his intuition about the fact that they can be wrong. But since our experiences are woven into a seamless whole, it is more consistent to extend this initial trust to all of our experiences, and then sift them to find out which ones are wrong, rather than discarding all but the few that conform with skepticism, and trying to discover which ones are right with nothing left to go on.

Here is how I think that works.

All of the beliefs that we form in the context of our conscious experiences have at least weak inherent justification. There is debate about whether we should consider this inherent level of justification sufficient, on its own, for the belief to count as properly basic. For example, if you see a ghost, you might not consider that experience by itself enough to believe in ghosts, even if you had no reason to disbelieve it. You might need some corroboration. However, with a foundation of inherent justification to build on, the coherence of your beliefs can play a role.

Experiences can be grouped into types based on the kind of things they are experiences of. Individual sense experiences are united by all being experiences of the external physical world, for example. When the experiences of a certain type cohere with each other, occur in different situations and to many different individuals, and in general weave together to produce a consistent and intelligible picture of reality, that type of experience is well-established. Types may be further established by their coherence with other types.

The coherence of well-established types of experience has the effect of combining, to some degree, the weak inherent justification of each token experience of that type. Coherence cannot manufacture justification from thin air, but it can combine the independent justification of mutually supporting experiences. The outcome of this process is that experiences from well-established types – such as our sensory experiences of the physical world, our introspective experiences of our own mental states, and our memories of the past – are grounds for properly basic beliefs.

So, a belief formed through conscious experience is at least weakly inherently justified, and it is justified, and therefore properly basic, if the experience is of a well-established type. Furthermore, such a belief is validly justified if the experience is such that the belief it forms is appropriately connected to the truth. For example, your belief that there is a pool of water ahead of you is validly justified if you are, in fact, seeing a pool of water – but the justification would be invalid if what you were seeing was a mirage (even if there still was a pool of water there).

In my next post I will look at the remaining ways that we justify our beliefs and acquire knowledge.

Building Materials (II): Abductive and Inductive Reasoning

I am exploring the different “building materials” that I can use to construct my belief system; by which I mean the different ways that we rationally justify our beliefs. Last time I looked at deductive reasoning. This time I will look at two other forms of reasoning.

Sometimes abductive reasoning and inductive reasoning are kept distinct, sometimes abduction is thought to be a subset of induction, and sometimes induction is thought to be a subset of abduction. (The latter option is my own view.) Both of them differ from deductive reasoning in an important way. In a deductive argument, the truth of the premises guarantees the truth of the conclusion, but in an abductive or inductive argument, the truth of the premises only makes the truth of the conclusion more probable. Nonetheless, both of these are indispensable in justifying our beliefs, both in our everyday life and in scientific investigation.

Abductive Reasoning

Abductive reasoning is more familiarly known as inference to the best explanation. In a way, it can be seen as deductive reasoning in reverse. Rather than concluding what logically follows from its premise, it concludes what that premise most likely follows from. Here is an example:

  • Premise: The ground is wet outside.
  • Premise: The best explanation of this is that it is raining.
  • Conclusion: Therefore, it is raining.

This is invalid as a deductive inference, but it can be valid as an abductive inference. That is, the truth of the conclusion is not guaranteed, but belief in the conclusion is justified.

Starting from some evidence and a set of hypotheses for that evidence, an abductive inference concludes that the hypothesis which is the best explanation for the evidence is the correct one, or that it is most likely to be the correct one from among the hypotheses considered. For the inference to provide good justification for the conclusion, the set of hypotheses should exhaust all the live options (the hypotheses that we are able to consider), and the chosen hypothesis should be sufficient to explain the evidence and should surpass all the other hypotheses in its explanatory merits.

There are a number of different criteria that are often used in assessing whether a hypothesis, let’s call it H, is a good explanation for some evidence, E. Here I will use the example of H, that it is raining, being proposed as an explanation of E, that the ground is wet outside.

  • Plausibility: H is more plausible if it is more likely given our background knowledge, that is, our knowledge with the specific evidence E removed from consideration.
    • For example, even if you hadn’t looked outside to see that the ground was wet, you know that it is spring, so it isn’t unlikely for it to rain.
  • Explanatory Power: H has more power to explain E if, given our background knowledge, H makes E more likely than it would be otherwise.
    • For example, even without knowing that the ground is in fact wet, we know that rain certainly would make the ground wet, whereas it isn’t that likely to be wet otherwise.
  • Explanatory Scope: H explains as many of the facts of E as possible.
    • For example, the hypothesis that it is raining fully explains the fact that the ground is wet. In contrast, the hypothesis that someone dumped buckets of water on the ground doesn’t really explain why all the ground is wet.
  • Accord: Adding H to our accepted beliefs does not entail contradictions or absurdities, nor does it make them more likely.
    • For example, the fact that the sky is dark and gloomy is not contradicted by our hypothesis that it is raining.
  • Parsimony: H is parsimonious if it is a simple and elegant explanation, if it postulates as few new facts as possible to explain the existing facts, and if it is not ad-hoc (that is, artificially tailored to the evidence). Effectively, H is parsimonious if it follows the principle sometimes called Occam’s razor.
    • For example, the hypothesis that it is raining is a simple explanation, whereas the hypothesis that someone dumped a large number of buckets of water on the ground in such a way as to make it look like it was raining is ad-hoc and not as simple.

One way to look at abductive reasoning is as a kind of qualitative or subjective generalization of Bayes’ theorem in probability. Bayes’ theorem computes the probability of a hypothesis after the evidence for it is taken into account, P(H|E). In this computation it uses the prior probabilities of the hypothesis and the evidence P(H) and P(E), and the likelihood of the evidence given the hypothesis, P(E|H). The prior probabilities are the probabilities given only the background knowledge, with the knowledge of E subtracted out.

P(H|E)=P(H) \times \frac{P(E|H)}{P(E)}

Here plausibility, accord, and parsimony can be seen as corresponding to the prior probability of hypothesis, P(H), and explanatory power and scope as corresponding to the ratio P(E|H)/P(E).

So, if an abductive argument is strong (that is, if the set of hypotheses considered exhausts the live options, the chosen hypothesis is sufficient to explain the evidence, and the chosen hypothesis better explains the evidence than all the other hypotheses) and if you justifiably believe the premises (the evidence to be explained), then the argument provides good justification for belief in its conclusion.

For this justification to be valid, the premises should be true and validly justified. Also, our reality has to be one in which abductive reasoning at least reliably produces true beliefs, so that the best explanation by the above criteria is most often the correct one. This seems to be the case, so I believe we are justified in using abductive reasoning.

Inductive Reasoning

The prototypical forms of inductive reasoning are generalizations, where the properties of all instances of a category are inferred from merely the known instances of that category, and appeals to uniformity, where predictions are made based on the expectation that observed patterns will continue. For example:

  • Premise: Every time we have burned a sample of sodium, it has produced a yellow flame.
  • Premise: We expect that burning sodium always produces the same colour flame.
  • Conclusion: Therefore, burning sodium always produces a yellow flame.

Many kinds of probabilistic inferences have been called inductive, making it difficult to clearly discuss all of them. Most forms of inductive reasoning that I have seen described either fall under the prototypical forms, or are a combination of them with deductive or abductive reasoning, or are actually better construed as a form of abductive reasoning.

My own view is that induction is actually a subset of abduction. Every generalization or appeal to uniformity can be construed as a simple inference to the best explanation. For example:

  • Premise: Every raven we have observed is black.
  • Premise: The best explanation for this is that all ravens are black. (Generalization construed as inference to the best explanation.)
  • Conclusion: Therefore, all ravens are black.

So inductive reasoning is really a form of abductive reasoning, or at least is closely related to it. So if an inductive argument is strong (for instance, when generalizing, the sample is appropriately representative of its category and you have a sufficient sample size) and you believe the premises (the data about the sample), the argument provides good justification for its conclusion. Validity for this justification has mostly the same requirements as for abductive inferences.

Now that I have looked at the different ways we justify beliefs using other beliefs, in my next post I will turn to the ways that we can form properly basic beliefs.

Building Materials (I): Deductive Reasoning

I’ve been surveying the ground on which to build a comprehensive belief system. Here are the main conclusions that I have come to so far:

  • A belief is an attitude of accepting a proposition as true.
  • Knowledge, the kind of belief that we want to hold, is validly justified true belief.
  • Justification for a belief is valid if it is appropriately connected to the truth of the belief.
  • Some beliefs are justified by other beliefs, and some beliefs are properly basic, that is, justified without reference to other beliefs.
  • Some of our properly basic beliefs are things we just know intuitively; among these are beliefs about what counts as knowledge and what doesn’t.

Now I am going to take a look at the “materials” out of which a belief system can be built: the ways that we actually use to justify our beliefs, in practice. I think the different forms of rational justification can be grouped into five categories:

  • Deductive reasoning
  • Abductive or inductive reasoning
  • Experience
  • Intuition
  • Testimony

In other words, there are five different ways to answer the question, “How do I know that?”

  • “It follows logically from these other things I know”
  • “It is the best explanation of these other things I know”
  • “I directly observe it in my conscious experience”
  • “I just intuitively see that it is true”
  • “Someone told me so”

The different forms of reasoning justify beliefs using other beliefs, while experience and intuition are ways of forming properly basic beliefs. Testimony can also be treated as a source of properly basic beliefs.

I am focusing here on rational justification, but I think in some cases there can be pragmatic justification for a belief. That is, sometimes you can be warranted in holding a belief, even when you do not have sufficient rational justification for it, because of the useful consequences of holding that belief. In fact, I think there are probably situations where you can’t actually find rational justification for a belief until after you have accepted that belief on pragmatic grounds. (You could argue that my project here of trying to discover the nature of knowledge and justification is such a situation.) But for now I want to try to keep pragmatic and rational justification distinct, and build my belief system using rational justification where I can. (And I also think that it is only rational justification, not pragmatic justification, that confers knowledge.)

Alright. With that preamble, the first “building material” that I will look at is deduction.

Deductive Reasoning

Deductive reasoning is the use of logic to derive a conclusion from one or more premises. Here is an example of a deductive argument:

  • Premise: If it is raining, then the ground is wet outside.
  • Premise: It is raining.
  • Conclusion: Therefore, the ground is wet outside.

Deductive arguments derive conclusions from premises using certain logical rules, called rules of deduction. If those rules are followed correctly, and the premises of the argument are true, then the conclusion of the argument is true. (Deductive reasoning can also be used where the premises are not certain, but where they are thought to have some probability of all being true: then a valid deductive argument guarantees the conclusion is at least that likely to be true.)

The above argument correctly uses the rule of conditional elimination (often called by its classical name, modus ponens) to derive its conclusion from its two premises. So, if those two premises are true, then the conclusion is guaranteed to be true. But the conclusion could still fail to be true if either of the premises is false. (For example, maybe it isn’t raining, or maybe rain doesn’t make the ground wet because the ground happens to be made of a super-absorbent material.)

Deductive arguments can also fail if they do not correctly follow the rules of deduction. For example:

  • Premise: If it is raining, then the ground is wet outside.
  • Premise: The ground is wet outside.
  • Conclusion: Therefore, it is raining.

This argument is invalid, because the truth of the conclusion is not guaranteed by the truth of the premises. (There could be another reason that the ground is wet; for example, maybe it isn’t raining but someone decided to dump buckets of water on the ground.) An incorrect use of the rules of deduction is called a formal fallacy. This example commits the fallacy called affirming the consequent.

Logical arguments can also suffer from informal fallacies. These have to do with the content of the argument, rather than its structure. For example:

  • Premise: Either pigs can fly, or the Earth is approximately a sphere.
  • Premise: Pigs cannot fly.
  • Conclusion: Therefore, the Earth is approximately a sphere.

This argument is logically valid and its premises are true. But the only reason you have for believing the first premise is that you already believe the conclusion; so this argument tries to support the conclusion using the conclusion itself. This is an informal fallacy often called begging the question.

So, if a deductive argument is free from any fallacies, and you justifiably believe the premises (or at least think it is likely that they are all true), then the argument provides good justification for its conclusion. Because the truth of the conclusion rests on the truth of the premises, for such justification to be valid, the premises should be true and themselves validly justified.

That is the first form of justification. In my next post, I will look at other forms of reasoning that we can use to justify our beliefs.