L2: Truth
(The casual younger reader of the thrilling adventures of the JugginsVerse will probably wish to jump back to here and start again. This episode is about a lesson on philosophy but, if it is any consolation, you cannot be more disappointed about this than Lottie.)
“I’m sorry” says Hipparchus “But you can put that poem away. I think we should first think a little about truth itself before we consider Ms Dickinson’s terrible advice. I cannot think what I was thinking when I suggested that.”
“Bother! You don’t mean we are going to do boring philosophy again?” She had endured a whole two weeks of tutorials on the philosophy of thought and language a year earlier.)
“Yes. Just a little philosophy for one morning only and tomorrow we can return to poetry, if not that poem. Read this paragraph. Read it aloud and tell me what you think.” There can be an imperious tone to these tutorials but Lottie likes to think she gives as good as she gets.
Lottie reads the text, which is from one of her tutor’s old
intelligence colleagues, who also had a day job at Oxford.
What is truth?’ said jesting
Pilate, and would not stay for an answer. Pilate was in advance of his time.
For ‘truth’ itself is an abstract noun, a camel, that is, of a logical
construction, which cannot get past the eye even of a grammarian. We approach
it cap and categories in hand: we ask ourselves whether Truth is a substance
(the Truth, the Body of Knowledge), or a quality (something like the colour
red, inhering in truths), or a relation (‘correspondence’). But philosophers should
take something more nearly their own size to strain at. What needs discussing
rather is the use, or certain uses, of the word ‘true’. In vino, possibly, ‘veritas’,
but in a sober symposium ‘verum’. (Austin 1961: 85)
There is a pause.
“That sounds just so stupid. It sounds artificial. Or ‘arch’ as you would say, archly!”
“Now now. I won’t hear a word against dear John. But what
preliminary thoughts do you have about what truth is? I should say that the
arch language suggests that Austin himself preferred not to carry out the kind
of discussion I want us to have today. Austin thought that most philosophical
problems are distorted by a lack of attention to the sort of uses of words we
actually make. In this case, concerning ‘truth’ and, better, ‘is true’. But I
want to risk doing what he would not support. What sort of ‘thing’ do you think
truth is? Or to start more gradually, to what do we apply the predicate – the
descriptive word – ‘is true’?”
“Well it is true that I had coffee, fruit juice and marmalade for breakfast. And it is true that my name is ‘Lottie’. And it is true that 2+2=4. It is also true that football is the best game – the best activity – in the world. These are all truths.”
“Perhaps. Your list certainly suggests the diversity of type of subject
matters to which ‘truth’ in some sense can apply. But starting more modestly,
to what sort of thing or entity does ‘is true’ apply?”
“Well, what I said. That I had toast for breakfast. That
2+2=4 etc etc.”
“Yes. It is true that you had toast for breakfast. Now
think back to your eating breakfast. Is there anything there and then to which
you can apply ‘is true’ in the way you could apply ‘is strong’ to the coffee we get served?”
Lottie says nothing.
“How about this question: why did you say that it is
true that you had toast for breakfast?”
“Because I did! I mean: I did have toast!”
“Yes but there are many things that happened to you, or
things that you did in one sense of that word that you’ve not mentioned.
You might have offended your tutor by arriving late, for example. (That is, in one one sense, not something you did but it may be the result of what you did do and thus in that sense something you did, too. Oh dear. Let us talk about Elizabeth Anscombe next term.)”
“But I wasn’t aware of any of that!”
“So to what does ‘is true’ attach, then?”
Lottie thinks. “… What... I... thought?”
“Yes and in virtue of what you thought, what you said. This
gives a sense of what can be true or false. Thoughts can be true or false. Including beliefs,
guesses, hunches. And expressions of thoughts in utterances. Sometimes it might
be right to say that sentences can be true – perhaps signs – though only relative to a language
and perhaps even only relative to an occasion to use that sentence (saying this
might get me back into Austin’s good books!). So going back, is your eating
toast for breakfast true? Your eating was no doubt hurried and perhaps a little
messy. But was it true?”
There is another pause. Hipparchus continues:
“Let me suggest the following idea. The sorts of things that
can bear truth or falsity are things with mental contents such as beliefs or
thoughts, and things with linguistic content or meaning such as utterances or,
perhaps, sentences, though always only relative to a context (such as being in
an English-speaking community). That you had toast for breakfast is
the sort of thing you can think. And what you thus think can be true, or false.
But what you think, when you think truly, can also be a fact: that you
had toast for breakfast. That you had toast for breakfast is a bit of the
world’s history. In this role, it is not something mental. Does that make sense? So there is
a danger in saying that facts are true, or that there can be ‘true facts’. What
might that danger be?”
“Well it suggests that there can be false facts. Which
sounds a bit like that oaf Donald Trump. Facts cannot be false. There are no other sorts of facts to true ones. Oh, but does that mean
that they cannot be true, either?”
“Not so fast. Knowledge, too, cannot be false. But while the description ‘true knowledge’ would be pleonastic, it is an implication of the claim that Lottie knows that p, that p. Or ‘p’ is true. It may help to think about what someone else – say your friend Masongill – commits himself to if he says that you, Lottie, know that Jupiter is the largest planet in the solar system. He, Masongill, also commits himself to Jupiter being the largest planet. (He also commits himself to you arriving at that idea properly, but let us ignore that.) To judge that someone else knows that p commits the judger, too, to holding the fact that p. Of course, Masongill might be wrong about Jupiter, if, for example, another larger planet is suddenly discovered and thus he is wrong about you too because, then you would not then know that Jupiter is the biggest planet. But the conceptual implication is that if you know something, then that something is true. Of course this is just a rough analogy. We are not talking about knowledge today but truth. But like knowledge, saying it is a fact that p implies it is true that p, or simply p.
Still, I like your thinking. Even if it follows from it being a fact that p to it being true that p, or just: p, it is best not to think of facts as themselves true (and certainly not to think of false facts!). This has led some philosophers to say that while facts cannot be true they, or something like them, can make things – mental things, truth-bearers – true. But I am not going to worry about ‘truth-maker’ theory today.
So to sum up. ‘Is true’ applies to truth bearers. Not every entity in the universe can be true (or false). This chair cannot be true. Perhaps a friend can be ‘true’ but only as a metaphor. To be true requires the sort of content that can be, though it does not have to be, specified in a that-clause. Earlier you expressed the thought that you had had toast for breakfast. There’s the that-clause. But now we can call that ‘Lottie’s thought’ if we want to and that is just a singular term referring to a thought. We can then say that Lottie’s thought is or was true. Or, if we wished, we could name that thought ‘McBain’ and say that McBain is true…”
“You do realise that that’s just weird!”
“…Still, the thought itself – our friend McBain in this case – can, like any truth-bearer, be unpacked in a that-clause.
Good. So we have made progress. We can say what can
be true or false. So now: if a thought is true, in what does its truth consist?
What sort of property is truth when we say, for example, that Lottie’s thought (‘McBain’) is true? Is it an intrinsic property like being red? Or a relational
property like being the same shape as?”
“Well it looks intrinsic.” Lottie replies. “We say that France is the
same shape as a hexagon. But we don’t say my belief is the same trueness as
anything. (We might say that things were equally true but that is obviously quite different.) But its being true does seem to depend on a connection. My belief is
true because of what I had for breakfast. It gets right what I ate. Or
it is right because of what I ate. Hmm. I assume that this is a trick question so
I’m going to say: ‘relational’.”
“Good, though it is a pity you think that philosophy involves trick questions. Still, it is a sensible thing to say. Let me speed ahead and
say that this is the basis of the correspondence theory of truth. We could think of that in capitals! The ‘Correspondence Theory of Truth’. It says that a true belief
– or utterance or sentence – corresponds to a fact. ‘Correspond’ is a two-place
relation like ‘the same shape as’. Now… let me ask you a couple of questions
about this and also about another theory and we can break for coffee.
If we wish to explain truth using ‘correspond’ and ‘fact’
what account can we give of those terms always assuming we cannot here use the
word ‘true’? (Why not? Because we are trying to explain ‘true’.) And, giving you a hint, if one thought instead that truth wasn’t a
relation to a worldly fact but was a coherence relation with, or to, all
one’s other beliefs, – you’ll guess that this is the ‘Coherence Theory of Truth’ –what account of ‘coherence’ could we give?”
***
Lottie goes off for coffee in the main hall of the Potting Shed leaving her
tutor alone in his study. While she does this, she meets a couple of the
younger adults who live there: Masongill the mole and pTravis the pterodactyl, one of whom, pTravis, studied some philosophy at college. But let us not
eavesdrop on their conversation.
When she returns, half an hour later, to her tutor’s
study, there is a confidence in her stride.
“Well, Teach” – Hipparchus shudders but she continues – “I’ve been chatting to some of the others over coffee and we all think you’re pulling a fast one. We cannot have a correspondence theory of truth because we cannot explain either of the basic notions it relies on. The best we could say about what a fact is is that it is what a true utterance says or what a true belief concerns or contains or is about. And we cannot say anything about ‘correspondence’ because it is obviously a weird relation: a relation that science could never discover. It is not a geometrical relation. It is not like a physical force. It is not like being heavier than or being the same weight. It doesn’t seem to be a real relation or property at all! The best we can say is that correspondence is the relation of a true belief and a fact. And that won’t help explain truth.
And while we’re at it, to cohere is for it to be possible
for something to be true alongside something else. So we can only explain
‘cohere’ if we already know what truth is or what ‘is true’ means.”
“Very good. Congratulate your friends for their good advice too! pTravis, no doubt, has a fine mind. Also, to cohere seems too weak a notion
to capture truth. It might be a good check. One should never hold two beliefs
that are inconsistent. But merely possibly being true together does not imply
that both actually are true. And even if ‘cohere’ means something a bit
stronger – such as one belief being evidence for another – rather than merely being
consistent with (as you and your friends, and I, too, seem to have assumed), it is
still hard to see how far this would get us.
So where do you think that leaves us? What do you, and your
friends, think?”
“Oh I don’t know. We didn’t think.” Lottie replies a little dismissively. “We just assumed that if
truth wasn’t correspondence or coherence then we didn’t know what it was. We
didn’t get any further. We only had half an hour! We didn’t want to waste even all of that on philosophy!”
Hipparchus sighs. Young people! “First, let me say that just because we cannot explain
truth in these two ways does not disbar us from using these slogans. There’s
nothing wrong with saying that a true belief corresponds to a fact or that all
true beliefs cohere. These are perfectly fine. It is just that they cannot be
used to explain ‘truth’. People often criticise the very idea of truths
corresponding, but they should have nothing to fear. Still, to persuade you of that, I
need to tell you a bit more.
Let’s start with the T-schema. It is the following simple formula where ‘s’ stands for a sentence such as ‘snow is white’ and ‘ ‘s’ ’ refers to that sentence as a sentence, as when one say that the sentence ‘snow is white’ has three words. By contrast, that snow is white contains no words: perhaps it ‘contains’ snow! The apparent triviality of the T-schema motivated the philosopher of language Donald Davidson to call it a ‘snow-bound triviality’.
‘s’ is true iff (if and only if) s
The left hand side mentions a sentence (‘s’). The right hand uses that same sentence to say something, such as that snow is white. The whole schema presents an equivalence. The ‘iff’ means if and only if.
It is a logical connective (like ‘and’ or ‘or’ or ‘if…then’). It says that when
the left hand side is true, so is the right hand side, and if the left side is
false, so is the right side. So the schema says that the left hand side and the right hand side have equivalent truth values. They are alike as to truth.
The T-schema says something about truth. The Polish
logician Alfred Tarski argued that generating every true instance of the T-schema would be a test for a formal specification of the truth condition for a
language. I would love to tell you more about Tarski. If you recall, we
discussed him a little in a previous tutorial. Tarski provided a logical
machinery for generating a specification of the T schema for every sentence of
a logical language using standard first order logic of the sort that Frege
developed. But that is not today’s topic.
The radical post-Tarski suggestion is that the T-schema says
almost everything one needs to know to know what truth is, or to grasp the
concept of ‘truth’. Let me say that again. If a passing Martian were to ask
what truth was, or what we understood it to be, this account says that it is
completely set out – or rather almost completely set out – by just this schema.
It is not that this is a test of truth, or part of a theory of
truth, it is almost everything about truth. One grasps the concept of truth if
one knows how to apply it to sentences under the condition set out in the T-schema. What is that?
Well we can pass either way across that logical connective
‘if and only if’. If one is prepared to assert that snow is white, then one
should also be prepared to say that the sentence (relative to a language) ‘snow
is white’ is true. Or, if one is prepared to assert that the sentence
‘snow is white’ is true, one should also be prepared to assert: snow is
white. And that is – almost – all there is to it.
Let me rush on and say why I keep saying ‘almost’. We do seem to be able to understand something else about truth. I can grasp the idea that I had a chat with Donald Davidson and that everything he said was true – he was Davidson, after all! – but that, because it took place in a university bar over several pints of bitter, I cannot recall what it was. (I am ashamed to say that this actually happened!) In such a case, I cannot strip off the ‘is true’, take off the inverted commas and re-assert what Davidson says because I have forgotten it. So truth also serves as a device of compendious endorsement.
I think that the best way to think of this is a kind of ‘genetic’
account. If one were to start with the redundancy approach, then expanding it to
include compendious endorsement in such cases of forgetfulness would not require adding a substantial element
into the idea of truth. It would not make truth into a property linking
thoughts and facts via correspondence, for example. It would merely be an expansion of the initial trivial rule. Had one been able to strip off the inverted commas and assert what remained, one would have been prepared to do so. That’s still quite minimal.”
***
“What do you mean by the idea that truth isn’t a substantial
property?”
“I mean that we can teach its whole use (almost!) via conventions about how
to cite sentences and commend them and how to take those sentences and use them
instead, summarised in the T-schema. And when we do that, our pupil is not inferring some more dramatic
power of truth. There is no dramatic power of truth. They, the pupil, are/is learning to talk
about sentences as well as using them.”
“But surely being true is a substantial property? It makes
all the difference? We can’t just go around asserting anything!”
“Good. Let me say two things about that.
First, in the 1980’s, there was an argument in philosophy of
science that the then sometimes influential view – Instrumentalism – that one should regard
scientific theories as merely practical instruments, and not think of them as
claiming to be true, was wrong. The counter argument was that, unless they were
true, then it was a weird miracle that they worked. That they worked
was an argument that the theories were actually true, or largely true. Success
implied truth. This is not a watertight argument. False beliefs may sometimes by chance guide one to some success. Even a stopped clock gives the right time twice a day. But consistent success is surely a reasonable sign of underlying truth?
That argument does not work in this context, however. Or rather, it is not truth as a substantial property that is doing the work. Think of these two claims:
- The theory that nothing goes faster than light works well because it is true.
This suggests that it is the substantial property of being true that explains the success of theory. (Compare with: the lorry crashed through the ice because it was so heavy, where being heavy is a genuine explanatory property.) But it need not. We could say instead:
- The theory that nothing goes faster than light works well because nothing goes faster than light.
Truth is not a distinct explanation of the success of our beliefs. It is transparent. The underlying explanation is the relevant fact to which ‘is true’ in the first sentence points.
Second, as the philosopher of language Michael Dummett argued: if all one knew about the use of the truth predicate (‘... is true’) was the triviality of the T-schema, then one would be missing a key fact. Truth is also the aim of assertion. Compare this idea – knowing the T-schema – with knowledge of the rules of a game such as football – what constitutes a goal, off-side etc – but shorn of the concept that winning (by scoring goals) is the whole purpose of the game. (I am imagining someone not knowing that goals are the goal, if you like!) To know what constitutes a goal or offside would be almost useless unless one also knew that winning, via scoring goals, was the point of the activity. Similarly merely knowing how to move one way or another across the T-schema would miss a vital feature of truth which is that we aim at truth. Not only can we switch between asserting that snow is white and saying that the sentence ‘snow is white’ is true but we also know that the aim is to say that snow is white only when snow is white and thus to say that ‘“snow is white” is true’ only if snow is white.
Following Dummet’s thought, the idea that truth is the goal of inquiry has been used to argue that more has to be said about truth than the T-schema provides. We would need to teach our Martian student more than just the T-schema. Thus, it seems, if truth is a norm or a goal as well as what is captured in the T-schema, then truth ‘inflates’ into something else. There must be some substance to truth. It is not almost redundant.
This inference can be blocked, however. Truth is not an extra
goal of enquiry. Rather, any normative properties – any properties having to do
with what we ought to think – apparently accruing to truth in general really
belong to each particular truth in question. Thus, in each case, one should aim
to:
- assert that snow is white iff snow is white
- assert that grass is green iff grass is green …
These can be gathered together by saying that one should
- assert s iff ‘s’ is true.
But in doing this, one is not adding a distinct norm of
truth, one is simply codifying the prior norms of assertion, the aims of that
game, but using ‘truth’ to speak of them compendiously.”
“So…” said Lottie “You seem to think that truth is really
important and not important at all? Typical!”
“Well one way to think about this may be what Austin meant by the contrast between veritas and verum. Truth, as an abstract property, causes philosophers problems and could, perhaps, be better ignored. But each truth, each ‘verum’, is important. But the truth in each verum does not inflate into some interesting property of truth in general: veritas. Snow being white has almost nothing in common with grass being green or murder being wrong or the Hawaiian pizza being an abomination. Each truth is its own particular and independent state of affairs.
Let me say one more thing before we call it a day. Earlier you suggested that an example what might be true was that you had toast for breakfast which I have suggested is best thought of as a bit of world history, a fact.”
“Yes but then I withdrew that!”
“And on my picture, the ‘truth-bearers’ are in some sense mental. They are thoughts or utterances with contents or meanings that can be given in that-clauses. For example, the thought that you had toast for breakfast. I’d like to suggest why you suggested what you did and why it was not silly, even if I do not think it quite right. What underpins minimalism about truth is what also causes just this kind of confusion. It is that we individuate thoughts and utterances using the very same that-clauses that we also use to individuate facts, or non-mental world history. I can say that Lottie thought that she had toast for breakfast and also just say that she had toast for breakfast. The first is her thought. The second, a fact. Further, it is because she did have toast for breakfast that her thought that she did is true. And then, it is because we use the same that-clause that it is then tempting, but misleading, to think that just as ‘... is true/false’ applies to the thought, which is articulated in a that-clause, it also applies to the fact, which is articulated in the very same that-clause: that she had toast for breakfast.”
“I am still here! It is me you’re talking about!”
“Indeed! So, to conclude, it is just this use of the same language – the that-clauses – which Wittgenstein says constitutes the harmony of thought and reality. It is that, too, that is expressed in the T-schema. The sentence on the left fits the fact on the right. Of course, that fact is expressed by a sentence on the right. The harmony of thought and reality lies in this nexus of language. But it is also why it is tempting, unhelpfully, to think that bits of the world itself, aspects of world history, are themselves true. Better, though, to keep a firm distinction between the order of thought and world history. You will recall that we talked about this last year.”
Lottie waits to make sure that he has finally stopped before she looks with some disappointment at her tutor. “Can we talk about Emily Dickinson tomorrow? I actually liked her poem.”