or Definition, facts, propositions, knowledge and truth
This is an attempt to clarify and synthesise ideas from my previous papers, and from contributions
to online discussion groups. But it can only ever be a work-in-progress.
As always, this is not an academic exercise, there is no scholarly apparatus, and I do not claim originality for any of it.
My aim is to make the glaringly obvious a little more visible.
3 Abstract things
6 Knowledge as justified true belief
7 The Gettier problem
8 Truth as correspondence
I assume we are part of a real universe, of which there can be objective knowledge, and about which we can make true assertions, given the way
we use the signs involved. It is possible to challenge each of these assumptions, but they are my starting point.
Given this, I think there are three separate things: features of reality; what we believe
and know about them; and what we say about them, which – in classical logic – may be true or false. And I think that to muddle these things up is a mistake.
For example, we can mistake what
we say about things for the way things are, with strange consequences, not least in philosophical theories of knowledge and truth, and in what is called metaphysics. And those consequences are my quarry.
believe the later Wittgenstein was after the same quarry, at least for some of the time. But I could be wrong about that.
The muddle we can get into – when we forget what language is and how it
works – begins with what we call definition.
We can use the word define in different ways. For example, to define a word is to show what it means – how we use it or could use it. But by contrast, to define a thing is to describe it, which is a completely different operation.
And yet, this is the first definition of the noun dog in The Concise Oxford Dictionary, ninth edition, page 399:
domesticated mammal … usually having a long snout and non-retractile claws, and occurring in many different breeds kept as pets or for work or sport’
In this example, a definition of a
word is a description of a thing. We explain the meaning of the word dog – the way we use the word – by describing the thing we name with the word. And this seems a natural thing to do.
But there is a two-stage process here, involving separate and different linguistic operations. First, we use the word dog to name and talk about the things we call dogs. And second, we can describe those things in some way or other –
using names of other things, such as mammal, snout and claw.
To put it another way: the two questions – ‘What is the meaning of the word dog?’ and
‘What is a dog?’ – are completely different. But the dictionary treats them as the same question. To define the word dog by describing a dog is to conflate the two separate and different linguistic operations: naming and describing.
And this may not seem to matter. After all, we know a name is not a description; naming and describing are different operations; a dog is not the meaning of a word, but rather a real thing – a feature of
reality; and a name is not the thing it names, any more than a description is the thing being described. Once pointed out, all this is obvious.
The problem comes if we forget these facts and mistake
the names we use, and the descriptions we produce, for the things we name and describe – if we forget that, people excepted, features of reality do not identify, categorise, name or describe themselves. And I think this happens more often than we realise.
We may think a definition of a thing is, somehow, a definitive description – one that captures the thing’s essence or fundamental nature.
But we can always describe things in different ways, for different purposes. And this applies to words, which are also real things we can describe in different ways, for different purposes: functionally, grammatically, etymologically, and so on.
There is no inherently essential or fundamental definition (description) of a thing. Talk of accuracy, precision, completeness, essence, fundamental nature, the absolute or ultimate reality – or just simple
talk of identity – is always within the context of one kind of description or another.
We can use different ways to explain how
we use or could use a word. For example, we can point at a dog to explain how we use the word dog – so-called ostensive definition. And in context, such an explanation may work, or at least help.
But there is no natural or necessary connection between a word and the ways we use it. The ways we use words and other signs – for example, to name and describe things – are always conventional and contextual.
3 Abstract things
The meaning of a noun is the use we make of it, not the thing it names. But
we do use nouns to name things. So it seems that what we call abstract nouns are the names of abstract things: belief, knowledge, truth, meaning, justice, beauty, goodness, identity, and so on. (Insert the abstract noun of choice.)
In fact, the expression abstract noun is a grammatical misattribution, because words are real things, such as sounds, marks on paper or screen, or signing gestures. So in the expression abstract noun, the modifier abstract
does not refer to the word noun. It actually refers to the unidentified thing that the abstract noun supposedly names.
But, pending evidence for the existence of abstract things, there is no
reason to believe they exist. More likely, we mistake abstract nouns for things which, therefore, we can try to describe. And this delusion has persisted for centuries – hence the perennially insoluble nature of philosophy’s so-called problems.
For example, the question ‘What is knowledge and where does it come from?’ can fool us into thinking knowledge is a thing with an origin and properties that can be described. So we produce rival theories
of knowledge, setting out necessary and sufficient conditions for knowledge to exist, or for something to count as knowledge. And we dignify this palaver with the name epistemology.
Or we conclude
such supposed abstract things as knowledge are difficult to describe. Or we fool ourselves by saying abstract things are concepts in minds – more abstract things – as though that explains anything. And then we deepen the confusion by distinguishing
between abstract concepts and concrete concepts – concepts of abstract things, such as knowledge, and concepts of real things, such as the concept of a dog. And this nonsense still passes muster in some circles.
The myth of abstract things is potent and pervasive. But they are just mysteries invented to explain mysteries of our own invention. A dog chasing its tail needs to re-think the premise.
We can use the word fact to mean ‘state-of-affairs’. But we also think of facts as things that are
true. And the following definition of the noun fact demonstrates the two different meanings.
‘a thing that is known to have occurred, to exist, or to be true’ (The
Concise Oxford Dictionary, as above, page 482.)
Now, if a fact is a state-of-affairs, then it has no truth-value. A state-of-affairs, and the things that constitute it, just
are or were, neither true nor false. And such features of reality can often be known and shown to be or have been the case.
But if instead a fact is a thing that is true, then the only thing it can
be is an assertion – typically a linguistic expression – because, given the meaning of the word true in the above definition, only assertions are true or false.
Of course, linguistic
expressions are themselves real things – features of reality, such as sequences of sounds, marks on paper or screen, or signing gestures. But assertions are the only features of reality that can have truth-value.
So we use the word fact in two completely different ways, to mean ‘a state-of-affairs’ or ‘a description of a state-of-affairs’. And while dual and even multiple word-uses are not uncommon, I think this example
is highly significant.
We can mistake what we say about things for the way things are – a description for the described. And the way we ordinarily use the word fact both demonstrates
and maintains the mistake. In effect, we can conflate the two meanings of the word fact, so that the description and the described seem to be one and the same thing.
And in practice, this is
what usually happens. If asked to identify a fact, what we actually do is produce an assertion that describes a feature of reality, given the way we use the words or other signs involved. If you disagree, try identifying a fact that is not an assertion of
For this reason, I define a fact as ‘a true factual assertion’: factual, because it claims something about reality that may or may not be or have been the case; and true, because
what it claims is or was indeed the case. (A factual assertion may be true or false.)
One advantage of using the expression factual assertion is that it allows us to distinguish between factual
assertions, which have truth-value, and non-factual assertions, which do not. For example, moral and aesthetic assertions are non-factual, because they express judgements or opinions, rather than make factual claims.
(I discuss the non-factual nature of moral assertions in An argument against moral objectivism (May 2019) at http://www.peasum.co.uk/441761408 . I think it demonstrable that moral assertions
do not make factual claims with truth-value; that there can be no moral facts, but only moral value-judgements; and that attempts by moral realists and objectivists to demonstrate the existence of moral facts are unsuccessful. But the debate goes on. And I
What we call objectivity is independence from opinion when considering the facts. So, in this definition, that there are facts
is a given. And this is why it matters to be clear about what we mean when we talk about facts and objectivity.
Aiming for clarity, many philosophers distinguish between states-of-affairs, which they call facts, and things said about them, which they call propositions. And this would be a useful distinction,
were it not for the problem of what a proposition is supposed to be.
A proposition is supposedly what a statement states, an assertion asserts, a declarative declares, and so on. So a proposition is
supposedly an abstract thing which can be embodied, expressed or represented by different so-called tokensentences, which are real things.
But, though ancient and pervasive, belief that abstract things
exist is irrational. And that, usually, we do not also speculate about the abstract things that questions ask, commands command and exclamations exclaim – points to the strangeness of thinking that what a statement states is an abstract thing of some
Like other abstract things, propositions are misleading fictions. We mistake the abstract noun proposition for a thing of some kind that can be described. But propositions supposedly
consist of subjects and predicates. And some are true or false, because they describe features of reality that may or may not be or have been the case.
So propositions are just like real declarative
sentences. And a symbolic representation of a proposition is nothing more than another linguistic expression – a codified translation. The so-called logicalform of an assertion is just another assertion.
But though propositions are merely linguistic expressions, they have played an ambiguous role in philosophy similar to the role played by facts in everyday language. And I call this phenomenon the myth of propositions.
The myth is so potent that knowledge of features of reality has been called propositionalknowledge. For example, omniscience has been defined as knowledge of all true propositions – as though knowing everything is
identical to knowing all true assertions about everything. And by extension, there has even been talk of propositionalbelief.
But propositions are just linguistic assertions that we can use to express
beliefs and knowledge-claims. There is nothing propositional about the existence and nature of things. Talk of propositional belief and propositional knowledge conflates what we believe and know about features of reality with what we say about them.
And this confusion is evident in the widely-used definition of knowledge as ‘justified true belief’, along with Gettier’s criticism of the definition.
6 Knowledge as justified true belief (JTB)
According to the JTB definition, the three necessary
and jointly sufficient conditions for knowledge are truth, belief and justification for the belief. And in one version of the definition, the so-called truth condition is as follows.
Analysis of Knowledge’ at https://plato.stanford.edu/entries/knowledge-analysis/.)
S knows that p if and
only if p is true.
The proposition p has two different functions here. The second p is an assertion with a truth-value. But the first p seems somehow to be the state-of-affairs
that the second p asserts. So the proposition p is both the thing described and the description, just as we use the word fact to mean ‘a state-of-affairs’ and ‘a description of a state-of-affairs’.
(I analyse the JTB definition more fully in Justified true belief: knowledge and the myth of propositions (July 2017) at http://www.peasum.co.uk/435531068.)
The JTB definition casually identifies a state-of-affairs with a proposition. And its truth condition is that we can know something is the case if and only if an assertion asserting it is true – an absurd
idea derived from the myth of propositions. It gets thing back to front.
And the confusion deepens. The justified true belief definition includes truth as a necessary condition for knowledge.
But the only things that can be true or false are factual assertions. So the expression true belief is another grammatical misattribution, rather like the expression abstract noun.
an acceptable shorthand, we do use the expressions true belief and false belief, and ordinarily their meanings are clear. But to believe something is to accept it, just as to disbelieve it is to withhold acceptance, or to reject it. And neither
acceptance nor rejection has a truth-value. Only factual assertions are true or false.
So in the expression true belief, the modifier true does not refer to the noun belief.
It actually refers to an unstated factual assertion. Simply believing (accepting) that a feature of reality is or was the case – in the JTB definition, ‘believing that p’ – has nothing to do with truth or falsehood, because
it has nothing to do with language whatsoever.
7 The Gettier problem
Recycling the confusion in the expression true belief, Edmund Gettier pointed out that, sometimes, justified true belief does not amount to knowledge, so that the JTB definition is incorrect or at least inadequate. And discussion of the so-called
Gettierproblem and possible solutions continues. But the assumed necessity of the truth condition often remains an unrecognised difficulty.
For anyone unfamiliar with Gettier cases, here is an example
I have used elsewhere.
A woman sees a group of people and mistakes one of them, a stranger, for her friend. So she believes her friend is there. And as it happens, her friend really is there, but hidden.
So what she believes is indeed the case. But does she know her friend is there?
The problem for the woman is not that she believes something that, unbeknown to her, happens to be true –
a proposition. She simply believes a state-of-affairs is the case for reasons that do not objectively justify the belief. The problem is her lack of knowledge of the actual state-of-affairs – objective knowledge that we Gettier-spectators do have.
In the post-match debriefing, a Gettier-protagonist would say they made a mistake. They would never say they knew the actual state-of affairs to be the case. That is not how we use the word know and its
To put it another way: the main condition for our knowing something is the case is that it is indeed the case. Our justification for believing it is the case is a separate, though disputed,
matter – and we can be mistaken. But none of this has anything to do with language, and therefore anything to do with truth or falsehood.
We may confuse ourselves by saying it is true that
a state-of-affairs is the case. But a state-of-affairs is neither true nor false. The truth is not out there any more than falsehood is. Outside language, reality is not linguistic.
8 Truth as correspondence
Like the word knowledge, the word truth is not the name of an abstract thing that can
be described. Signs such as words can mean only what we use them to mean, and there is no other court of appeal. For example, what we mean when we talk about factual assertions being true is what constitutes what we call truth. So what is called a theory of
truth can only be an explanation of how we use or could use the word truth, its cognates such as true, and related words, such as false and falsehood.
will furkle. And among the various theories of truth, the most prominent are called correspondencetheories. A claim common to such theories is that an assertion is true if there is or was a state-of-affairs that corresponds with the assertion, and false if
there is not or was not. And one formulation of the claim is: truth is that which corresponds with, or comports with, reality.
For example: the assertion snow is white is true, supposedly,
because real snow really is white – at least, usually. And this is an attractively simple explanation of what makes an assertion true, which is why correspondence theories are so popular. They seem commonsensical.
But in this context, The Concise Oxford Dictionary (as above, page 300) defines the noun correspondence, in the relevant sense, as follows.
agreement, similarity, or harmony’
And other dictionaries define correspondence as a close connection or equivalence.
Trouble is, the
words snow and white are not in any way similar or equivalent to, connected with, or in agreement or harmony with, the things we call snow and white.
A name no more corresponds with
the thing it names than an arrow corresponds with its target. And to pursue the analogy, it is not the arrow that identifies, defines or delimits its target. There is no natural or necessary connection between a name and what we name with it. The word correspondence
implies a two-way relationship. But there is no relationship at all, unless an arrow can be said to have a relationship with its target.
Arguably, words such as splash and smack,
are exceptions. But such onomatopoeias are conventional, idiomatic and language-specific. If they are exceptions, they only prove (meaning ‘test’) the rule, which holds for all other names. We can use the word dog to talk about dogs. But
there is nothing canine about the word dog, or the word canine.
To put this another way: the things we name using the words snow and white, could be any things whatsoever.
Features of reality exist, but they do not organise themselves in order to correspond with the ways we talk about them. To think that is to mistake what we say about things for the way things are.
there is no correspondence between names and what they name, there can be no correspondence between factual assertions and what they assert. There is no agreement, similarity, harmony, close connection or equivalence between the assertion snow is white
and the state-of-affairs that it asserts.
And, once pointed out, the circularity of the correspondence claim is obvious. Here it is again, expressed more formally: the assertion snow is white
is true if and only if snow is white. In effect, this is a tautology – a purely linguistic exercise. (And defining truth in a target language by using a meta-language is just another linguistic exercise.)
The idea that the whiteness of snow is a truth-maker, and the assertion snow is white is a truth-bearer, restates and reinforces the idea of correspondence.
But what we call truth can only be a function of the ways we use language in factual assertions. To repeat, truth is not out there any more than falsehood is. What we call reality is not the set of all possible truth-makers. (Nor is it the set of
all possible falsehood-makers.)
Correspondence theories of truth are mired in the myth of propositions, which is pervasive. For example,
the JTB truth condition – S knows that p if and only if p is true – is an expression of the supposed correspondence between a state-of-affairs and a proposition. Indeed, the correspondence assumed in the JTB definition is so perfect
that it amounts to identity: a state-of-affairs and a proposition asserting it are one and the same thing.
But a state-of-affairs can be identified and described in a limitless number of ways, for different
purposes. Within each kind of description, there can (classically) be true or false claims. No true description is truer than other true descriptions. And no one kind of description has an inherent priority over the others. And these facts have considerable
A true factual assertion
does not constitute what we call a state-of-affairs. It is just one way of identifying and describing a state-of-affairs. And there cannot be a description that is not a description. So, that what we call truth is relative to, or dependent on, a descriptive
context, is both true and inconsequential.
To say there is no such thing as absolute truth is to entertain a fantasy, if only to dismiss it. And this applies to what we call accuracy, precision, completeness
or perfection. Along with absoluteness, these are not things outside and independent from a descriptive context. That we can always say more does not mean we can never say enough.
To say all models
are wrong, but some are useful, is to entertain the fantasy of a model that is right – a perfectly or absolutely accurate, precise or complete description – relative to which all other attempts must fall short. If we recognise and dismiss the fantasy,
we can also dismiss or at least mitigate our superstitious anxiety about what we call the relativity of what we call truth.
Logic deals with the use of declarative sentences in arguments: what can be said consistently, without contradiction. But we also use the words logic and logical
to refer to reason or thought. For example, one entry in The Concise Oxford Dictionary (as above, page 802) defines logic as:
‘the science of reasoning, proof, thought or inference’.
But reasoning, proving, thinking and inferring are not linguistic activities. And yet logic, which deals with language, is supposedly the science of reasoning, proving, thinking or inferring.
So the assumption is that we can study what we call thought by studying the use of language in declarative sentences – which is what logicians do. It is as though a sentence expresses, or even simply is
a thought, and as though combining sentences in an argument constitutes thinking. But this is the myth of propositions at work – in this case, conflating what we think with what we say.
contrasting manifestation of the myth of propositions is the idea that logic deals not with language, but reality itself – and that features of reality conform to the rules of logic: identity, non-contradiction and excluded middle.
In one way this seems obvious. After all, a dog is a dog and is not not a dog. Or to put it another way, a thing cannot both be and not be a dog. As Bishop Butler put it: something is what it is and not another
thing. And the assertion this is and is not a dog is what we call a contradiction – a ‘speaking-against-itself’.
But reality is not linguistic, so in reality (outside language)
there can be no contradictions. There are no identities and categories in reality, but only things that can be identified and categorised in different ways for different purposes. And only assertions can be contradictory. So, that we say there can be contradictions
in reality again demonstrates the myth of propositions at work.
(I maintain that, even if quantum indeterminacy turns out to be fundamental, this would still not mean there are contradictions in reality.)
When it comes to what we call knowledge, we
want to be sure that what we believe is or was the case – is or was indeed the case. And, trying to define knowledge - what it is and where it comes from – we have often used the metaphor of a foundation – something that grounds what we know,
or at least claim to know. And the main candidates for the foundation of knowledge have been what we call experience – variously defined – and what we call reason – again, variously defined.
But, to repeat, there is no reason to believe the abstract noun knowledge is the name of a thing of some kind that, therefore, needs a foundation. Metaphors have their uses, but they can lead us astray. For example, if we think of
what we call knowledge as a building, we can fool ourselves into looking for its foundation, arguing over what it is, or concluding that it does not have one, so that knowledge is not what we think or say it is.
Or, to avoid the argument, we can change the metaphor. Perhaps knowledge is better thought of as a web, with unchallengeable certainties at the centre, and expendables at the periphery. Or, perhaps, knowledge rests on true propositions, or properly
basic beliefs. Or, perhaps, all we can hope for is a constructed coherence, because foundationalism is a mistake. And so on, down the rabbit hole where metaphysicians furkle. Metaphorically.
expressions abstract noun, and true belief, the expression certain knowledge is a misattribution, because, like doubt, certainty is a state of mind, not a property of what we call knowledge. And a metaphysical delusion about the
nature of knowledge underlies both epistemological foundationalism and its rejection, which really manifest our longing for certainty, and our uncertainty.
Like descriptions, explanations are conventional
and contextual. And like descriptions, explanations come to an end. The fear of circularity or infinite regress is related to the fear that absolute truth or a perfect description is impossible. And behind these fears is the fantasy that a description could
shake off its nature and meld with the described – become one with it – the fantasy of the saying that says it all.
I began by pointing out the need to distinguish between three things: features of reality; what we believe and know about them; and what we say about them. And I have tried to show why muddling those three things
up is a mistake – particularly when we mistake what we say about things for the way things are.
Trouble is, when we talk about anything – including reality, what we believe and know, and
what we say – we have to use language – so glaringly obvious a fact that we can forget it. We cannot use language to get beyond language, but we can fool ourselves into thinking that we can, particularly when we do philosophy.
Ordinarily, language is not usually in the way – something we feel the need to get beyond. Philosophical pastimes, such as logical or conceptual analysis, theorising about knowledge and truth, and metaphysics, while
entertaining for some of us, have no bearing on the everyday, practical use of language in multifarious contexts.
What we call facts
– in the sense of true factual assertions – constitute the objective knowledge we express using language. But those facts are nothing more than linguistic expressions. So, to use a metaphor, we build and repair this knowledge with materials of
our own making. Its foundation is in – and its coherence comes from the consistency of – our linguistic practices. But that does not mean the edifice must be shaky.
And I think this was
the later Wittgenstein’s profound and hard-won insight. His earlier picture theory of meaning, in Tractatus Logico-Philosophicus, is an example of the delusion of identity between factual assertions and the things they assert. But later he recognised
the mistake and set about painstakingly undoing it, by showing that the meaning of words is in the multifarious ways we use them.