What is truth ?

Fallibilism rejects the correspondence theory of truth; attachment to the theory is one of the defining features of the opposing, infallibilist camp.

It is appropriate to pin the label ‘infallibilist’ on the correspondence theory since any claim of truth under it necessarily implies a claim of finality and perfection: Under the correspondence theory, the honorific ‘truth’ is reserved for claims that are thought to have achieved a perfect match, finally locked into reality.

Most people today do not actually believe this naïve version of the correspondence theory, but some vague notion of eventual convergence towards the truth.

However, the supposedly more sophisticated theory of convergence towards correspondence is no better than the theory of brute correspondence. In many ways, it’s only more confused, and worse. Talk of convergence is used by certain interested parties as a propaganda instrument, a way of making notions of correspondence palatable without having to sound too implausible.

Under the convergence theory, claims of metaphysical certainty about being in contact with reality are replaced by claims of metaphysical certainty about making progress towards reality – being on the right track, a glide path of inevitable convergence towards correspondence truth. False certainties about having reached reality are thus replaced by false certainties about measuring the remaining distance to reality. What this forgets is that unless we knew the final position, we have no way of triangulating the distance to various positions in the approach.

Science works. Its theories tend to become safer and safer bets. However, today’s safest bet does not have to look anything like tomorrow’s safest bet. As much as hagiographic historians of science may try to tidy up the process, the progress of human knowledge is not significantly more linear than political history – unpredictable revolutions, counter-revolutions and counter-counter-revolutions do happen. Paradigm changes can come out of directions no one was expecting. There is always the outside chance of convergence being derailed, and then rerouted to a different track. We can never know that derailment is not an option.

Let’s say we exchange theory B for theory A, believing that to be progress. In future we may come to prefer a theory C to theory B. Yet theory C, while retaining many of the advantages of theory B, may in some respects again be closer to A than B. And before we know theory C we can’t know which aspects or components these are.

Everything remains on the table, all of the time. We may move some assumptions towards the edge of the table, and almost forget about them. But they still rest on the table, forever.

The commonsense idea of truth is a mangled notion, badly distorted by long association with the correspondence theory. We need to separate out from the idea of truth the illusory notions of finality and perfection, and keep only the valuable and natural concept of justification.

There is no sense in defining knowledge – as many analytical philosophers like to do – as true justified belief. This definition is either tautologous or wrong. Tautologous because there is no accounting for truth independent of justification; any useful definition of truth must include an account of justified knowledge. Wrong because knowledge does not have to be true in a correspondence sense, ie indefeasible, in order to be knowledge.

The correct definition of knowledge is just ‘sufficiently well justified belief’, with a rider reminding us of fallibility. Knowledge does not have to be ‘true’ in the false sense of perfect or final in order to be useful. The excess meaning that ‘truth’ has over ‘currently best justified belief’ is simply our fallibility, the well-known propensity of even our best justified beliefs to be defeated by unexpected evidence.

You say that there are no universally agreed, wholly uncontroversial criteria of justification ? That disagreements about the truth of claims persist despite best faith efforts at resolution ? Welcome to the pluralism of life. At the danger of sounding like Sartre, there is no way of getting out of your existential responsibility to make up your own mind, of choosing which claims you want to accept and on which grounds. Or your responsibility to decide which authorities you want to trust and for which reasons: Delegating your choices to an authority does not absolve you from the responsibility of having delegated. Here’s the evidence, here are the disagreements. Where’s the “truth” ? Make up your own conscience, no one can do it for you.

Truth, naively understood, is not the same as best justification because we remain morally free to take any attitude to the claims we are presented with. It is up to us to opt for those criteria of justification that look and feel – to the best of our ability, head thoughts and gut feelings combined – like being able to help us place our bets in the most responsible way we know. The final criterion of justification is to keep all the criteria up in the air and try to make the best sense of the whole that one can.

This makes truth a moral concept. The fallibilist theory of truth is therefore not a correspondence theory; it is a moral theory.

The correspondence theory has tried to hide the moralising effect of its claims behind the mechanical image of copying reality. What might be called the argument from certainty made it possible to moralise without being seen to moralise.

The argument from certainty short-circuits the loops from being presented with to acceptance, from limited confidence to absolute confidence, from acceptance to execution, loops which could otherwise have filtered out many totalitarian impulses. The appearance of metaphysically removed objectivity made statements all the more moral.

Fallibilism, on the other hand, by recognising a gap between claiming and accepting leaves room for personality, history and tradition. Truth being the radically fallible outcome of the least unreliable procedure yet devised cannot ever become totally impersonal, ahistorical, or decontexual.

Under a moral theory of truth,

P is true.

translates as

You ought to believe P.

Belief here is intended in a thick sense – not just entertain intellectually, but fully accept P, act on the basis of P, incorporate P into the web of assumptions according to which you live your life.

From here the moral interpretation of truth can go in one of  two directions. Let’s look at the traditional way first: Claims of truth have always had moral force. To claim truth for a statement is to claim that all reasonable people ought to respect in their actions what the statement says; is to claim, conversely, that whoever disregards the statement is acting unreasonably, and at any rate cannot be “one of us”.

P is true.

understood in the bad but usual way means something like

You are deficient as a person if you do not (with us) accept P.

Calling on this sense of truth is a way of turning the argument ad hominem, of turning it away from the evidence and towards the person; the effective meaning is little different from a personal insult. It is a move towards escalation, potentially explosion. Value-free objectivity is a worse than value-less illusion.

That said, there are good ways of using notions of correspondence, apart from this malignant root of pompous rhetoric. Asserting the existence of a single, true reality to which our ideas must correspond has no useful literal meaning (depending, as it does, on the correspondence theory), but it does have effective meanings.

It can in effect be a way of saying:

a) Synthesis: Look, we can’t both be 100% right about this. Let’s talk it through and try and find an agreement, or even better, some sort of synthesis that we can both agree to be superior to our starting positions.

b) Progress: An injunction to think of perfectibility, to suggest that there might be a chance to improve on our current set of beliefs.

c) Virtue: A call on people to practice the virtue of objectivity, i.e. focus on facts and arguments, not emotions and interests.

d) Caution: A sceptical warning that even a unanimously held consensus can be horribly wrong.

e) Egalitarianism: Good evidence offered by a low status witness should trump bad evidence offered by a high status witness.

f) Possibly more.

Language is brilliant: even literally meaningless mumblings can be used pragmatically to get things done.

Because of the baggage of correspondence thinking it carries, it would perhaps be best if we simply stopped talking about truth. This is unlikely to happen, so the next best thing we can try to do is usurp the old word ‘true’ for a new usage in order to drive out the overtones of correspondence.

We are now on the second path a moral interpretation of truth can take. To get there all we have to do is follow up with the obvious questions: P is true. So I ought to believe P. Says who ? And why ? All we need to do is demand that the claims become embodied, by naming a speaker, and listing evidence.

P is true

if used at all should be used as a shorthand for a more
cautious kind of claim that does not scare off dialogue:

My friends and I recommend P as the best option currently supported by our evidence (for reasons A,B,C,...)

This is long-winded. So ordinarily I will just abbreviate it to ‘P is true.’ When you catch me saying ‘P is true’, this is what I mean. (It is not an argument against this definition that it does not, in all regards, respect current usage. Capturing current usage is not the aim of this definition; its explicit aim is to reform current usage.)

The practical meaning of truth is exhausted by a presentation of evidence. Once that is done we can do only two things: shrug our shoulders or search for new, more convincing evidence. Of course everybody believes their own position to be superior; that goes without saying. The impoliteness, the pretentious claptrap comes into it when people claim that their own presentations are something qualitatively different from presentations, namely representations; whereas their opponents' presentations are mere, well, presentations.

Your only evidence for the superiority of your position is your evidence. So leave it at stating your evidence. Nothing beyond follows. The only way for anyone to transcend the evidence is to search for yet better evidence. There is nothing left to do except shut up, or go search.

Truth thus reduces to an ethical commitment to seek the best available evidence while acknowledging the continuing fallibility of even the best justified beliefs.

Everything else is rhetoric. Divisive, unconvincing, autistic rhetoric. A form of preaching to the converted that may bore even them.

I am sure about not wanting to hide behind any term less vulnerable than ‘recommend’. I accept that you are free to reject whatever I say. I would consider it unwise if you did, but there we go. I can’t force you to accept any of my presentations. Words do not compel, or at least not in this sense. And I have no interest in using other means to “persuade” you. I’m merely giving dialogue a try, betting on the off-chance that you might be persuadable; I know that discourse can fail, and more often than not will.

Also note how ‘recommending’ that we all say ‘recommend’ is perfectly consistent – self-ratifying and not threatened at all by begging its own definition. The proposal of reforming our habits is itself a first example of the new habit.

Contrast this with the inability of the correspondence theory to account for its own truth. What kind of a fact could the truth of a correspondence theory of truth possibly correspond to ?

Ultimate explanations are inevitably self-referential. A theory of truth, if it is to be universal, has to account for its own truth. In the case of the correspondence theory this arguably leads to a regress – we need a theory of correspondence TOC₂ in order to account for the truth of TOC₁. So while the main reason for rejecting the correspondence theory are the uses to which it tends to be put, we can also give a pure coherence argument against it: The theory cannot avoid a vicious regress in its justification.

By contrast, the fallibilist theory of truth, a recommendation to recommend, is safe from vicious circularity; likewise, the meaning rule, basis of the fallibilist theory of value, is able to satisfy itself. Consciously adopting the principle of minimising exposure should help to minimise exposure. Taking on the meaning rule could be approved by it.

Language is a system of presentation, not representation; and dealing with truth claims therefore a special branch of moral activity. Epistemology, the theory of truth and knowledge, becomes a subdiscipline of ethics, the theory of morality. Epistemology is concerned with regulating truth-related action, the making of truth claims as much as reacting to them.
Truth in no strict sense concerns representations of reality, in statements about truth we judge acts of presenting. Representing reality can mean only the most desirable mode of presenting, desirable because it is open, honest, inspiring.

Scientific standards of evidence are in practice standards of honesty in presentation. This is to say they are quietly moral.
The scientific ethos which holds that, with a proper reason, no assumption ought to be unquestionable has a wider applicability than the practice of science. The ways of science are paradigmatic ways of reducing exposure. These evidence-focused practices represent a core commitment of the fallibilist ethic, and apply to all areas of life, not just the sciences.

Truth is a moral concept; so is reality: Truth is what we ought to believe, reality is what we ought to respect. Facts are relatively unsuspicious truth claims; values are strategies for minimising assumptive exposure; character is an assembly of such strategies. By saying ‘ought’ you recommend a course of action you reckon makes the most sense, because it would, for all concerned, minimise exposure best.
Fallibilism believes that no hard and fast boundary can be drawn between fact and value; that the difference is one of degree, not kind. Factual assumptions are often narrowly focused, evaluative assumptions more sweeping; factual assumptions are mostly concerned with things one cannot influence, evaluative assumptions with things one can; factual assumptions tend to concern detached descriptions of the past whereas evaluative assumptions declare intentions for the future; and so on, in several other dimensions.

The moral theory of truth has nothing to fear from the idea of knowledge being a social construction, historically situated and evolving. The how of constructing goes naturally with the how of presenting, just as the whether of correspondence went naturally with fixed essences.
There is relativism here in the sense of a healthy respect for the full variety of approaches. There is at the same time absolutely no relativism in the sense of being indifferent between competing validity claims. The moral character of the theory encourages, even demands, that people become judgmental about the quality of claims they encounter.
To repeat the obvious: Some theories are solidly constructed, and stand up. Others are a danger to their inhabitants.

Certainty is a useful organisational category in handling assumptions – certainties are those assumptions expected to repay doubting least. Certainties in this sense need not be eternally true because assessments of the payback can change. Not infrequently, the certainties of one age are the laughing-stock of the next. We have simply no need for eternal verities, substantively defined. What we need is what we have: often reliable procedures.
The point where doubting becomes frivolous and time-wasting, or the weighing of alternatives becomes paralysis is for moral judgement to define. No one doubts the existence of certainties. The argument is about what it means to be one.

Because the correspondence theory required statements to objectify easily when projected – which then, backtracking, originated the claim that the statements mirrored reality –, it was biased in favour of the propositional form, A is B.
Knowledge that could be assimilated to propositional form was made to appear more solid than knowledge that could not. Factual truth seemed normal, moral truth odd; cognitive knowledge normal, intuitive knowledge odd; knowing that normal, knowing how odd; language-based reasoning seemed normal, emotions odd.
Because morality rests in the halo of preferences or factual propositions, in the inexpungable possibility of their being transcended, it could not been seen by a vision focused on the substantive statements themselves. In rushing towards the limit of total objectivity, the halo was lost.

Under a moral theory of truth by contrast, there is a sense in which morality could be said to underlie and so be more certain than factual truth.