Why belief in
the supernatural is irrational
I first published Exorcising Religion in 2015. In this revision, there is more detail on the case against belief in the supernatural. I posted this version here on 12 January 2017. As always, it’s a work-in-progress,
so any helpful comments are welcome.
2.1 Rationalism 2.2 Irrationality 2.3 Ring-fencing belief 2.4 Rational criticism
3.1 Defining words 3.2 Defining things 3.3 Abstract things 3.4 Propositions
4.1 Assertions 4.2 Facts 4.3 Mathematics and logic 4.4 Non-factual
4.5 Other kinds of truth? 4.6 Factual or non-factual? 4.7 Belief 4.8 Objectivity
4.9 Justification 4.10 Certainty 4.11 Evidence 4.12 Foundations?
4.13 Exclusions 4.14 Authority
4.15 Scripture 4.16 Personal experience
4.17 Probability 4.18 Proof 4.19 Demonstration 4.20 Factual proof
4.21 Scientific method 4.22 Absence of evidence 4.23 Burden of proof
To be rational is to have or seek sound reasons for what we do and believe.
2 We can express a belief by making or endorsing
an assertion, such as the earth orbits the sun or there is a god.
3 Assertions are either factual or non-factual.
The assertions the earth orbits the sun and Bach’s music is sublime have different functions.
4 An assertion
may have both a factual and a non-factual use. The assertion Jesus lives may be factual or, for example, metaphorical.
There are two kinds of factual assertion: those about the way things are, including scientific assertions; and those about how things were and what happened.
6 Features of reality just are, neither true nor false. It is what we say about them that can be true or false. The truth is not out there any more than falsehood is. Reality is not linguistic.
7 In a given context, a factual assertion is true if it correctly asserts a feature of reality, given the way we use the words or other signs involved.
8 A factual assertion is either true or false, regardless of what anyone believes or claims to know. (We call a true factual assertion a fact.)
9 The truth or falsehood of a factual assertion is independent of its source, popularity, longevity and utility.
10 Personal beliefs and knowledge-claims are what we call subjective. But factual knowledge is what we call objective, which means independent of personal
11 True factual assertions (facts) constitute the objective knowledge we express in language. We build and repair this
knowledge on foundations and with materials of our own making.
12 If there are sound reasons to believe a factual assertion, such
as there is a god, the belief is justified. But whether the reasons are indeed sound is open to rational debate.
experience, perceptions and understanding are necessarily subjective. And anyone can be mistaken, for a variety of reasons.
feel justified in a belief is not always to be justified. And how well our minds and perceptions are functioning is irrelevant. To free ourselves from subjective limitations, we need objective justification.
15 Spoken or written testimony that a thing exists or an event occurred may be accurate. But it consists of factual assertions needing objective justification in the form
16 The evidence for a factual assertion is the feature of reality that it asserts. The assertion records but does
not constitute the evidence. A factual assertion does not verify itself.
17 To prove a factual assertion means to test
it against the evidence. The evidence may justify belief that the assertion is true. But it may not guarantee its truth. Factual proof is different from deductive reasoning.
18 We could be wrong about a factual assertion. We may lack sufficient evidence, or misinterpret the evidence we have. It is rational to be open-minded.
19 If there is no evidence for a factual assertion, we cannot test it, so we cannot show if it is true or false. That may not mean it is false. But to believe it is true is irrational.
20 There is no evidence for any factual assertion of the supernatural. Such assertions have no objective justification, there is no sound reason to
believe them, and to do so is irrational.
21 A burden of proof is the responsibility for providing evidence for a factual
assertion. To provide evidence is to meet the burden of proof. Someone who denies an assertion, such as there is a god, has no responsibility for showing it is false.
22 To define a thing is to describe it. But we cannot define a thing into or out of existence. And if a thing does not manifest in reality, its presence is indistinguishable from its absence. When we describe
it, we can (and do) make up anything we like.
23 A factual argument consists of two or more factual assertions, the last of which
is a conclusion. But an argument is not evidence for its own conclusion. An argument for the existence of a god is not evidence for the god’s existence.
24 The arguments for there being a god are abductions (arguments to the most likely explanation). But as there is no evidence for the existence of supernatural beings, nor for their role as causal agents, the
supernatural cannot be the most likely explanation for anything.
25 A causal explanation works only if there is evidence of the
cause and of how it caused the effect. If there is no evidence, there is no sound reason to believe the explanation. The absence of a rational explanation does not make an irrational one viable. Sometimes, the only rational conclusion is: we do not know.
26 If we have no natural explanation for an event, we can either keep looking, or give up and conclude it must be a miracle. The first approach
leads to greater knowledge. The second explains nothing and keeps us trapped in superstitious ignorance. The assertion we do not know the cause, so it must be supernatural is an argument from ignorance.
27 Our moral rules and values are rational, because there are sound reasons for having them. There are facts about well-being, so morality has an objective basis. We did not need
a rule-giver. And anyway, an action is not good just because someone (or some god) says it is. The argument from authority is as false for morality as it is for factual knowledge.
1 We must be free to be irrational, as long as we
do no harm to others. But irrationality is foolish. And it is wrong to bully or seduce others, particularly children, into our folly.
We must be free to practise any religion we choose, as long as we do no harm to others. But all religions have the same supernatural justification: none whatsoever.
3 Others must respect our right to believe anything. But if our belief is irrational, they do not have to respect what we believe or our judgement. It may be better for us if they do not, but rather help
us to understand our mistake.
4 Religion is big bizniz. Religious leaders and teachers may always have been sincere. But still,
to gain wealth and power on the basis of unjustified claims is wrong. And to do it knowing the claims are unjustified is also fraudulent.
5 We can be rational and good without religion. And the fewer reasons we have for being irrational or wicked, the better.
Whatever is good about religion can survive its demise.
Many years ago, at a summer camp led by charismatic young adults, I became a Christian. My enthusiasm lasted a few weeks, then shame began to take over. How could I have been
so easily taken in? It was hard for a supposedly clever boy to acknowledge his loneliness and insecurity.
Since then I have been non-religious,
but also fascinated by religion in general, and Christianity in particular. Or perhaps haunted is a better word. And the cure for a haunting is an exorcism – in my case, long overdue.
Great thinkers have tackled the arguments over religion for millennia, so I cannot hope to come up with new ones. My aim is just to assemble and clarify those that seem to me the most telling. And to begin, I
think it important to make two distinctions.
The first is between religion and theism (belief in gods or a god). Belief in the supernatural
does not entail religious commitment. For example, believing that Jesus was the son of a god does not make you a Christian. And religious faith does not entail belief in the supernatural. For example, there are Christian atheists.
Religions are historical institutions, open to anthropological and sociological study, and moral appraisal. But claims about supernatural things and events are about reality
or existence itself. And my main aim is to deal with those claims.
And the second distinction is between religion and morality. Whatever
relationship there may be between them, they are different things, and it is a mistake to muddle them up.
People must be free to do and
believe what they want, as long as they do no harm to others. And religion can motivate people to do good things. But there are several related reasons for thinking we could be better off without belief in the supernatural – and the religions that peddle
First: we could end the personal and social harm that religion causes. People have believed and done irrational and wicked things
because of their religion. Of course, people have also believed and done rational and good things because of their religion. But what matters is being rational and good, which we can be without religion. And the fewer reasons we have for being irrational or
wicked, the better.
Second: we could use the time, energy and wealth we devote to religion in more useful ways – for example, to
solve practical social problems.
Third: it is rational to care about the truth. Free of irrational beliefs, we could face our future
as individuals and as a species more clear-sightedly, and so perhaps more successfully.
In what follows, I concentrate on what most concerns
me: getting clear about what we are really saying when we make or deny claims about the supernatural. But this requires some groundwork on what it means to be rational, and what we think makes an assertion true or false.
What I have written is not an academic study, so there are no footnotes or references. I hope the arguments I put forward persuade by their own merits, or at least stimulate
you to argue and think things through for yourself.
Rationalism is the theory
that reason is the foundation of certainty in knowledge. And in theology, to be a rationalist is to treat reason as the ultimate religious authority. These grandiose claims for reason have ancient origins: reason (or more precisely intellect) was thought to
be our divine light, our god-given guide to knowledge and truth.
This confidence in reason as the divine source of certainty has largely
disappeared. We must still use reason, secularised and qualified, to reach the conclusions that constitute a large part of what we call knowledge. But we can now explain what that means in practical, everyday terms.
To be rational is to have or seek sound reasons for what we do and believe, and to be willing to change what we do and believe if the reasons turn out to be unsound. To put
it another way: actions and beliefs are rational when there are sound reasons for doing and believing them. But there are three provisos.
reasons are explanations. But we use different kinds of explanation for what we do and believe: scientific, historical, psychological, economic, political, and so on. Which kind of explanation is appropriate may vary according to what we are trying to explain
and the purpose of the explanation. And these choices are open to rational debate.
Second: what we regard as the appropriate explanation
and evidence has sometimes varied from place to place and through time. An example is the shift from a supernatural to a scientific understanding of nature. This does not mean our explanations always will change, so that we can never be sure they are correct.
It only means we must be aware they may change in the light of new knowledge. It is rational to be open-minded.
Third: what we think
is a sound reason for a belief may not in fact be sound. For example, we used to believe the sun and stars move around the earth, because the earth seems to be stationary. So being able to give a reason does not ensure that what we believe is true. Rationality
just gives us the best test we have, given the state of our knowledge.
Given this account of rationality, we can define its opposite: to be irrational or non-rational is not to have or seek sound reasons for what we do and believe. If this is a
matter of choice, we would probably describe it as irrational. But we would probably describe very young children and some mentally disordered people as non-rational.
Of course, a great deal of what we do is unquestioned routine based on unexamined beliefs. But that is not irrational or non-rational or unintelligent behaviour. We just get on with life, because it is rational
to do so. We act rather than reflect. But to refuse to at least address arguments challenging what we do or believe would be irrational.
given this account of rationality, the question is whether belief in the supernatural is rational. Are there sound reasons for believing in such things as fairies, ghosts, souls, spirits, devils, angels, gods, the one god, miracles, or the afterlife, or any
one or any combination of them?
Most people who do believe in one or some of them do not knowingly have unsound reasons, or no reasons
at all. On the contrary, they think they have very sound reasons. So the debate is about the soundness of those reasons. It is a rational debate.
2.3 Ring-fencing belief
I emphasise this because one response to rational enquiry is the claim that belief
in the supernatural is not answerable to reason. The attempt to ring-fence such belief has taken various forms, some ancient, some mutually inconsistent or contradictory, but all of them persistent. Here are some of the more prominent.
Our belief is true, so you must believe or be punished, in this world or the next.
We cannot hope
to understand or talk sensibly about the divine mystery.
Religion and science are non-overlapping magisteria, with different concerns.
have always been religious, so we always will be.
Religion is something you do, not primarily a set of beliefs.
Unless you practise a religion,
you cannot understand it.
Religious belief is subjective, so not open to objective appraisal.
Criticism of religious belief is offensive to
2.4 Rational criticism
In different ways, all of these arguments aim to deflect or neutralise rational criticism of religious beliefs and practice, or to end the discussion. But no one can have the last word in this or any other rational
There are always explanations for why we do or believe things, even if we cannot or do not want to explain them. For example,
to say the divine is inexplicable may be to offer a reason for exempting religion from rational appraisal – a reason which, like any other, is open to rational appraisal.
And to say religious faith is not the intellectual endorsement of a creed or set of doctrines, but rather a commitment to a way of life, leaves why we should make such a commitment, and why to one religion rather than
any other, open to rational debate.
As disillusioned adults, we may be nostalgic for childhood innocence and faith. Sometimes we are
told to believe in a god in the way little children do. Sometimes we feed our children fairy-tale falsehoods: only believe and your dreams will come true. Sometimes we long for irrational certainties.
The metaphor leap of faith suggests attractive qualities: impetuosity, trustfulness and courage. But (to labour the metaphor) we need solid ground from which to leap. There is no traction in the void.
The leap of faith has grounds, which are the reasons for taking it.
Rationality involves having or seeking sound reasons for our beliefs.
And we express those beliefs and reasons in the form of assertions that we regard as true. So we have to clarify what makes an assertion true.
But again, some groundwork is necessary. The nature of truth is bound up with the nature of belief, knowledge and justification – all of them studied in the branch of philosophy called epistemology. And first, they all require definition.
We use the word define in (at least) two different ways. To define a word is to explain how we use it, or could use it. By contrast, to define a thing is to try to describe it definitively, which is a radically different operation.
3.1 Defining words
Words do not define themselves. They do not tell us how we must use them. Nothing in the nature of what we talk about compels us to use words in a particular way. And this applies to the word definition. We decide what defining involves.
Words mean only what we use and understand them to mean. We have to agree on the ways we use them, and confusion results if we do not. Their meanings
come out in the various contexts in which we use them. Those uses can be more or less precise. And they can and sometimes do change.
Things do not define themselves. They do not tell us how to describe them, let alone definitively. (People
are an exception. They may say how they want to be described. And we may well agree with them.)
We produce different kinds of description,
with different purposes. So the idea of a definitive description may be misleading. A definitive description of a thing, a definition of that thing, can only ever be in the context of one kind of description.
We can describe a real thing by locating it, listing its properties, recounting its history, or explaining its function, and so on. Which type of description is appropriate
in a given context is always open to rational debate.
The idea that a thing has an essence or an essential nature arises from our mistaking
a definition of a thing for the thing itself. A definition is just a description. And once we have named a thing, an assertion of its identity is similarly just a description.
The temptation to mistake what we say about things for the way things are is often over-powering. And it can lead to the illusion that reality is linguistic, so that, for example, things conform to logical or mathematical
When we describe a real thing, we can usually check if the description is accurate, if our assertions about it are true, by examining
the thing itself. But if the thing is not real, this is impossible. We can make up anything we like.
The properties we ascribe to an
unreal thing are not real properties. For example, Pegasus did not have real wings. In the realm of the unreal, everything is unreal.
may try to define a supernatural thing, such as a god. If it is real, we must in principle be able to check our description is accurate. But if by definition this is impossible, or if the god is unreal, we can (and do) make up anything we like.
We cannot define a thing into or out of existence. The existence of things is independent of the names we give them and the ways we describe them.
3.3 Abstract things
We can get confused when we try to define what are misleadingly called abstract things, such as belief, knowledge, justification and truth. For example, to define knowledge is not to describe a thing.
To assert that abstract things exist, and that therefore we can try to describe them, is to use the word exist in a strange and, so far, unexplained way. And to call
abstract things concepts does not solve the problem. Concepts are just more abstract things that exist in the mind – another abstract thing. Our metaphors fool us into imaginary mysteries.
To assert or deny that abstract things exist may be to mistake abstract nouns for things which therefore do or do not exist. But words are just words. All we can do is explain how we
use or could use words such as knowledge, belief, justification and truth, and their cognates, such as know, believe, justify and true. In what follows I try to do so as clearly as I can.
In philosophy, a proposition is supposed to be what an assertion asserts, or a statement states, or a declarative declares, and so on. (Strangely, we rarely speculate about what a command commands or a question questions.)
Like concepts, propositions are abstract things. And like the myth of concepts, the myth of propositions is a philosophical distraction. So I prefer to talk
about factual assertions, which are spoken or written linguistic expressions, and which are real.
We want to know if the things
we believe or claim to know are true. For example, we want to know if belief in the supernatural is justified. The next section, on truth, is the heart of my argument about belief in the supernatural.
Things, situations and events (features of reality) just are, neither true nor false. Truth is an attribute of some assertions, not of what they are about.
(Assertions about assertions are an exception.) The truth is not out there any more than falsehood is.
Because we think of nouns
as naming words, we may think an abstract noun such as truth is the name of something. And we talk about discovering the truth, which adds to this confusion. The way things are or were is what we discover. Then it is what we say about them
that can be true or false.
To say I am the truth is to use the word truth in a strange way. The truth is not a thing,
and so not a thing that something or someone can be. Of course, we are free to use words in any way we choose. But a strange use of words is likely to obscure the meaning of what we say.
If instead I am the truth is a metaphor, we have to interpret it, to find what if anything it is asserting, and whether that assertion is supposed to be factual, and therefore true or false.
What we call a factual assertion is one that is true or false. As it happens, we use the word fact to mean true assertion. And there are two kinds of fact: those about the way things are, including scientific assertions;
and historical assertions about how things were and what happened.
A fact is a spoken or written linguistic expression. It is a real
thing, consisting of sounds or marks on paper, and so on. And like all linguistic expressions, it does not inhere in reality. It is an assertion about a feature of reality.
In a given context, a factual assertion is true if it correctly expresses or describes a feature of reality, given the way we use the words or other signs involved. (But how well the assertion does this is always
open to rational debate.)
Although we produce them, factual assertions are either true or false, regardless of what anyone believes or
claims to know. And facts (true assertions) constitute what we call objectivity, which in turn informs what it means to be rational: to have or seek sound reasons for what we do and believe.
Before we go on, a quick digression on mathematics and logic, which may also seem to be factual.
4.3 Mathematics and logic
Arguably, mathematical assertions are also true or false. But 2 + 2 = 4
is true simply because it follows the rules of number and addition. And rules do not have what is called a truth value – they are not true or false. They are just instructions.
Given the way we use these figures and symbols, there is no situation in which 2 + 2 = 4 could be false. It does not try to describe something, and no kind of evidence could falsify it. So arguably it
is redundant to say it is true.
Also arguably, logical rules can be true or false. But the so-called law of identity, a = a,
is not a description. For example, the assertion a rock is a rock does not say what a rock is, nor why it is different from all the not-rocks. And no evidence could falsify it. So it is not really about identity. Instead it is a rule for talking about
anything. Logic analyses language, not reality. Reality is not linguistic.
Rather than true or false, it may be better to call mathematical
assertions and logical rules correct or incorrect. End of digression.
4.4 Non-factual assertions
Many assertions are not true or false in the way factual assertions are. For example, there are aesthetic assertions such as Bach’s music
is sublime, and moral assertions such as killing people is wrong. We may say these assertions are true or false to show that we agree or disagree with them. But this is to use the words true and false in a non-factual way.
Non-factual assertions often matter profoundly to us, and moral assertions certainly do. But they do not have the same function as factual ones.
For example, it would make no sense to try to prove, or assess how probable it is, that Bach’s music is sublime. That would be to misunderstand the assertion, to treat it as factual.
Moral assertions are not true or false descriptions of what are or were features of reality, so they are not factual. Instead they are judgements about what should happen, which is why there can be rational debate
about them. But they are based on facts about the well-being of living things, so morality has an objective basis. And this may explain why some people believe there are moral facts. It is an understandable misunderstanding.
4.5 Other kinds of truth?
often talk about different kinds of truth: poetic, emotional, subjective, metaphorical, allegorical, symbolic, mythical, spiritual, and so on. But truth is not a thing, and so not a thing of which there are different kinds.
To say something is emotionally or spiritually true is usually to show how much we admire the way it captures or expresses something. We may say this is true to life
or this expresses a profound truth. A demand for evidence of the factual truth or probability of these assertions would usually be inappropriate.
It is a mistake to confuse the normal, factual use of the word truth with these quite different, non-factual uses. Factual truth is not one among many other kinds.
4.6 Factual or non-factual?
Whether an assertion is factual or
non-factual depends on its purpose: what we mean by it, and how we understand it. And this means we can use many assertions either factually or non-factually.
For example, how we understand and therefore judge assertions about the supernatural depends on whether they are factual or not. Here are some examples.
The Bible [the Qur’an] is the inspired word of a god.
The angel Gabriel commanded Mohammed
to recite the Qur’an.
Two days after his execution, Jesus was alive again.
If these assertions are historical, and so factual, then they are true or false, and we need evidence in order to decide. If they are not meant to be historical, their significance and consequences are radically
Mathematical assertions and logical rules are in a special category of their own. All other assertions are either factual
Some factual assertions, such as those about feelings, may be hard to assess. For example, evidence for the assertion
Helen loves her husband may be hard to find, difficult to assess, or inconclusive, and so on. But that does not mean it is not a factual assertion.
We can express a belief by making or endorsing a factual assertion, such as there
is a god. But the assertion itself is not a belief. In the same way, the expression the earth orbits the sun is not a belief. It is (usually understood to be) a factual assertion, which is therefore true or false.
A belief is an attitude towards something, for example towards a factual assertion. And an attitude has no truth value. What we believe may be true or false, but the
belief itself is neither. To put it grammatically: in the expression true belief, the modifier true is misattributed.
this is not trivial. We do use the expressions true belief and false belief as a kind of shorthand. But it can cause confusion. For example, it is one of the problems with the traditional definition of knowledge as justified true belief.
To decide if a factual assertion is true or false, we need to unframe it, to separate it from the beliefs held or expressed about it, because
they are irrelevant.
To believe a factual assertion usually means to believe it is true, not that it may be true. For example,
believing there is a god is different from believing there may be a god.
Personal beliefs and knowledge-claims are what we call subjective. But factual knowledge and its justification are impersonal, or what
we call objective – independent of anyone's opinion. There is a distinction between what anyone believes or claims to know, and what we call knowledge, which is objective.
The factual assertions for which there is evidence constitute the objective knowledge we express in language. And to be rational about factual assertions is to be objective. But it is also to be open-minded.
We could be wrong about a factual assertion. We could lack sufficient evidence and be unaware that we do, or we may misinterpret it, or new evidence
may contradict our conclusion, and so on. Ours is a long history of being wrong and often only slowly and painfully finding out.
objective is to use a method for getting at factual truth. It does not guarantee that we always will. But to be subjective about factual truth is to have no rational method at all: this is true because I think or feel or believe it is.
The reasons we have for believing a factual assertion, such as there is a god, or claiming to know it is true, constitute our justification for doing so. If they are sound reasons, we are justified. But whether the reasons are indeed sound
is open to rational debate, such as the one between theists and atheists.
To be rational, both beliefs and knowledge-claims need justification.
So the difference between beliefs and knowledge-claims is not to do with justification itself. A knowledge-claim is just a kind of belief. So it is simpler just to talk about beliefs.
When it comes to believing a factual assertion, feeling justified is not always the same as actually being justified. And how well our minds and perceptions are functioning is irrelevant. That has no bearing on
whether the assertion is true or false. All we can do is show why we believe it is true. And to free ourselves from subjective limitations, we need objective justification.
How certainly someone believes a factual assertion is true,
how sure they are, is a personal matter. If someone says I believe there is a god, but I know the earth orbits the sun, that may show a different level of confidence, perhaps because they feel there is less evidence for the first assertion than the
But still, these are simply two factual assertions: there is a god and the earth orbits the sun. So both need
objective justification, regardless of what anyone believes, or how certain they are. Justification for a factual assertion has to be objective, and that requires evidence.
We use the word evidence in two very different ways.
Unfortunately, it can refer to spoken or written testimony, such as in court, consisting of factual assertions: this was the case or this happened. A witness swears: the evidence I shall give shall be the truth, the whole truth
and nothing but the truth.
But this use of the word evidence for what is really testimony can be confusing. For example,
if someone says I have been abducted by aliens, that is a factual assertion. It is not evidence that an alien abduction has occurred. A factual assertion does nothing to establish its own truth. It may be true, but it needs objective justification.
What we more properly call the evidence for a factual assertion is the feature of reality that it expresses or describes. And a fact is
a linguistic expression which records that feature of reality. A fact does not constitute its own evidence. Rather, a fact is an assertion for the truth of which there is evidence.
For example, if someone says today the sky here is blue, that testimony provides no evidence that it is true. The actual evidence is the blueness of the sky.
To make a factual argument, we refer to evidence by asserting a fact or facts. But an argument is just a sequence of two or more linguistic expressions, the last of which is a
conclusion. The purpose of an argument is to assert that the cited evidence justifies the conclusion. And we have to judge whether it succeeds.
In itself, an argument does not constitute evidence for its own conclusion. Evidence, on the one hand, and on the other hand the facts and arguments we produce, are quite separate. For example, an argument for the existence of a god is not evidence
for the god's existence.
The amount and kind of evidence needed for a factual assertion varies with the nature of the assertion. For
example, the assertion I have a dog is unremarkable, and usually inconsequential. But the assertion a god speaks to me is remarkable and highly consequential. Much hangs on the likelihood and implications of an assertion.
We have standards of evidence for a factual assertion, though new knowledge or techniques may change our standards and so our judgement of the assertion.
As always, it is rational to be open-minded.
And though the evidence for a factual assertion, such as there is a god, may warrant
only a probable rather than a definite conclusion, we are sure the assertion is true or false. Even if we will never know the answer, it could in principle be decided. Usually, we know what kind of evidence would settle the matter.
To be rational is to believe only those factual assertions for which there is evidence. This is what Thomas Huxley meant when he coined the word agnosticism, though
its meaning has changed a little since then. Nowadays we tend to use expressions such as rational skepticism or critical thinking.
We may express the evidence for a factual assertion by producing other factual assertions
for which we may in turn have to provide evidence, and so on. But justification comes to an end when we reach assertions which we have no sound reason to question – though that does not mean those assertions must be true.
It is rational to want a foundation and coherence for the factual assertions we believe or claim to know are true. But a fact is a linguistic expression. So we build and repair
factual knowledge on foundations and with materials of our own making.
Deciding what makes a factual assertion true involves excluding the factors that are irrelevant. The truth of a factual assertion is independent of:
who asserts it – the argument from authority
many people assert it – the argument from popularity
for how long it has been asserted – the argument from tradition
how sincerely it is asserted – the argument from sincerity, and
how useful or comforting it is – the argument from utility.
These are, no doubt,
reasons why some people believe certain assertions to be true. But they have no bearing on whether those assertions are indeed true.
The arguments from popularity and tradition – this assertion is true because many people believe it,
or because it is part of a tradition – are really specific types of the argument from authority, which is a pervasive fallacy.
a factual assertion is true or false has nothing to do with its source. For example, if the assertion life evolved by natural selection is true, that has nothing to do with Darwin. It is purely a matter of the way things are. And the assertion E=mc2
is true because Einstein asserted it is obviously false.
The argument from authority is false even if a god is the authority. If
the god makes a true assertion, its truth is not a consequence of the god having made it. And the god cannot alter the truth or falsehood of an assertion. To claim that the god never speaks a falsehood is (usually) to make a factual assertion requiring evidence.
Belief is irrelevant.
Incidentally, the claim of papal infallibility is another argument from authority which is therefore false. What
the pope says, ex cathedra or otherwise, is not true just because he says it, or because a god tells him what to say.
The argument from scriptural (written) authority is also false. The assertion this is true because it
says so in this book assumes that the assertion everything in this book is true is true. (An assumption is a stated or unstated assertion on which the argument depends.)
Of course, the assertion everything in this book is true may be true. But the assertion does not establish its own truth. We need evidence. Accepting such a factual assertion without objective justification
- accepting it on faith - is not rational.
The argument from scriptural authority may be circular: everything in
this book is true because in the book it says that everything in this book is true. But this means: in scripture it says that scripture is true; so scripture is true.
This reasoning provides no justification for its conclusion. So an argument from scriptural authority often relies on three assumptions: there is a god; that god wrote - or inspired the writers of - the book; and that
god always speaks the truth.
But these are separate factual assertions needing objective justification in the form of evidence. Without
that, the assertion everything in this book is true because a god wrote or inspired it is unjustified. But the argument from scriptural authority is tenacious, despite its unjustified assumptions.
4.16 Personal experience
By definition, personal experience is
subjective. Our experiences, perceptions and understanding may lead us to believe factual assertions that are true. But they may not give us the means to recognise others that are false, or to tell the difference.
Anyone can be mistaken, hallucinating, experiencing mental disorder, suffering from indoctrination, and so on. Personal experience is like eye-witness testimony in court. It may turn
out to be reliable. But it needs objective examination and corroboration.
Someone may sincerely believe they hear a god speaking to them.
But that is not evidence that what they believe is actually the case. If there is evidence, it is elsewhere, independent of any beliefs. And in this case, other explanations are more likely. For example, auditory hallucinations are quite common experiences.
The assertion this is true because I believe it is has no place in a rational discussion. But it is a seductive fallacy with which we can
all deceive ourselves.
The evidence for a factual assertion may be conclusive or inconclusive, though what counts as conclusive evidence is open to rational debate. If it is conclusive, that justifies our believing the assertion is
true. If the evidence is inconclusive but still valid, it contributes to the probability that the assertion is true.
A probability is
a numerical assertion about the likelihood of something being the case or happening or having happened. To calculate it accurately, we have to know all the facts that influence or influenced the outcome. So we need evidence.
Probabilistic reasoning is informed guesswork, so the better the information, the better the reasoning. For example, when we roll a die, we need to know where its centre of
gravity is, and whether all its surfaces and edges are uniform and equally polished, and so on. The crude probability of 1/6 that a number will turn up may not be accurate.
To try to calculate the probability that there is a god, or that a god created and sustains the universe and life on earth, we need all the relevant factual information. The more unjustified assumptions we rely on, the
less accurate and useful the calculation.
The truth or falsehood of a factual assertion is independent of whether it can be proved true or false. But we use the words proof and prove in two confusingly different ways.
A proof in mathematics is called a demonstration that a conclusion is necessarily true. And traditionally we ended a mathematical proof with Q.E.D (quod erat demonstrandum - which is what had to be proven.) For example,
in geometry we prove that the internal angles of a triangle add up to 180°.
A triangle is not evidence for the conclusion,
but rather contains or implies the conclusion, which is why the conclusion must be true. As in the case of 2 + 2 = 4, it may be better to call the conclusion correct rather than true.
In logic, this kind of reasoning is called deduction: in any situation in which the premise is (or the premises are) true, the conclusion must be true. A premise is usually
a factual assertion, but it can also be a mathematical or logical one.
4.20 Factual proof
But a factual proof, such as in science, is completely different. Here we try to reach a rational conclusion using the available evidence. In logic, this
kind of reasoning is called induction: the premise or premises provide a sound reason or reasons for the conclusion. But they may not guarantee the conclusion is true.
With induction, to prove the conclusion means to test it against the evidence. And this older use of the words proof and prove to mean test comes out in expressions that can
be confusing nowadays. Here are some examples.
the proof of the pudding is in the eating – we test the pudding by eating
the exception proves the rule – an exception tests how strong the rule is
40% proof spirit – a spirit shown by testing to have 40% alcohol
Originally, to be rational meant to apportion (or ration) our beliefs by testing them against the evidence: the better the evidence, the stronger the belief.
4.21 Scientific method
Science progresses in a muddled, ad hoc,
all-too-human way. But the method of induction is a useful simple model for how it works, as follows.
1 We observe the natural
evidence and record it by means of factual assertions.
2 We generalise those facts to produce a provisional explanation, which
is called a hypothesis.
3 Then we test the hypothesis against the natural evidence by making predictions.
4 If the hypothesis passes the test, it becomes what is called a theory, which is a (so far) confirmed explanation.
The problem is that, like the word proof, we use the word theory in two very different ways. In everyday speech, a theory is a suggested explanation awaiting
verification. But a scientific theory is a confirmed explanation based on evidence, used to make predictions, and, so far, verified in practice.
If a prediction fails, the new information will influence our understanding accordingly. We combine practical certainty with open-mindedness. That is rationality at work.
And leaving aside scientific progress, such practical certainty makes our everyday lives possible. For example, our practical beliefs about matter, temperature and gravity are beyond question. For sound reasons,
we are not unsure about them.
We have evolved on a world in a universe with certain physical properties. As there are no sound reasons
to doubt them, to do so would be irrational, and perhaps an indication of mental disorder.
In philosophy, the so-called problem of
induction is that inductive conclusions are not deductively, necessarily, true. But they never try to be. That is not their function.
Absence of evidence
Sometimes, those who make claims about the supernatural quote two maxims: absence of evidence
is not evidence of absence and you cannot prove a negative. But these are both misleading.
If there is no evidence for
a factual assertion, there is no sound reason to believe it is true. Of course, the absence of evidence may not mean the assertion is false. But it is a sound reason, at least for now, to believe it is not true. However, a simple example shows that things
may be clear-cut.
In a given situation, the assertion there is a chair in this room is true or false. We can prove (which here
means test) the assertion by investigating. If there is a chair, the assertion is true. If there is no chair, it is false.
that case, the absence of evidence does indeed confirm a negative: it is not the case that there is a chair in this room. And belief is irrelevant, because there is a fact of the matter – which is why belief that the assertion is false is objectively
(Obviously, I am assuming we are real beings in a real universe. I believe these are factual assertions we have no sound reason
to question. Idealism, solipsism and computer-simulation hypotheses make a virtue of having no evidential basis, which means they provide no justification for factual claims about anything, let alone the supernatural.)
Back in our real universe: in a given situation, the factual assertion there is a supernatural chair in this room is also true or false. But we cannot test this assertion
against the evidence, because there is no evidence, no feature of reality that could objectively verify or falsify the assertion. The presence of a supernatural thing is indistinguishable from its absence.
For this reason, supernaturalists who wish to be rational have to claim there is indeed natural evidence for the supernatural. Mere arguments for its existence are not enough. The debate must always come back
to the evidence.
4.23 Burden of proof
To prove a factual assertion is to test it against the evidence. And if the assertion is historical, this means to provide evidence for its truth. The burden of proof for a factual assertion
refers to the responsibility for providing evidence for its truth. To provide the evidence is to meet the burden of proof.
example, many Christians believe the factual assertion two days after his execution, Jesus was alive again. To be rational, they need at least one sound reason to believe it, at least one piece of evidence to provide objective justification.
Someone who denies that a factual assertion such as there is a god is true has no burden of proof, no responsibility for showing it is
false. If you make no claim, you have nothing to prove.
The burden of proof for factual assertions of the supernatural has never been met. As far as we know, there is no evidence for such assertions. So those who believe them have to
do so on faith, which in this context means belief without evidence.
It is not the case that factual assertions of
the supernatural must be false. But the absence of evidence for them means it is irrational to believe they are true.
Belief in the supernatural
cannot rationally be exclusive. To believe without evidence in one supernatural thing or event is to have no grounds for denying other supernatural things or events. The denial would be as irrational as the belief.
In the absence of evidence, arguments for the existence of a supernatural being, such as a god, have to be causal: this feature of reality, or reality itself (the cosmos), can have no
natural cause, so the cause must be supernatural.
But a causal explanation works only if there is evidence of the cause and of how it
caused the effect. If there is no evidence, there is no sound reason to believe the explanation. And the absence of a rational explanation does not make an irrational one viable. Sometimes, the only rational conclusion is: we do not know.
A so-called argument to the most likely explanation (an abduction), or an explanation based on probability, must first show its conclusion
is possible. There is no evidence for the existence of supernatural beings, nor for their role as causal agents. So the supernatural cannot be the most likely explanation for anything.
If we have no natural explanation for an event, we can either keep looking, or give up and conclude it must be a miracle. The first approach leads to greater knowledge. The second explains nothing and keeps us
trapped in superstitious ignorance.
If assertions of the supernatural are not factual, then they are not factually true or false, and
they do not have the implications which have been claimed for them. But we must be free to believe anything, as long as we do no harm to others, or try to impose our beliefs on them.