Truth: the Final Word

My Thoughts on Truth and the Philosophical Truth Discourse Accumulated over Years of Contemplation

Foreword:

A lot of the material here dates back to a bigger Google Doc project I worked on with my best friend for more than a year, maybe two years, and which we both essentially abandoned around two years ago. Other things are more recent. This information is not important and you wasted your time reading this. Enjoy.

I’ve always been a bit mystified by the philosophical debate over so-called “theories of truth”. One reason is that I don’t really think we need a ‘theory of truth’ per se, and another is that it seems to me a failed enterprise, given that all the candidates for stand-alone “theories of truth” seem not to match our intuitions about the role the lexeme TRUTH plays in a large set of sentences in natural language (each one fails on a different large set of sentences in natural language). This forces the proponents of these theories to declare large swathes of natural language somehow infelicitous, which is confusing to me, as I can’t understand why you would propose a ‘theory of truth’ that doesn’t account for large swathes of language. The ‘resolution’ to this problem is to hold that humans are getting it wrong somehow, a take which requires that truth is a real thing: some entity in an external reality of abstracta which binds to some other set of abstract entities - theories or propositions or theorems or some other class of things, which therefore must also exist in this abstract realm. But even if one was inclined to believe such a thing (for what it’s worth, I personally tap out of philosophical inquiry once I hear so much as a mention of the ontological status of abstracta), it doesn’t seem like anyone has any arguments as to why this truth stuff in this abstract realm would only choose to bind to the class of entities that their theory favours, i.e. why it would only choose propositions which ‘represent’ or ‘picture’ aspects of reality (Correspondence Theory), or why it would only choose coherent groupings of propositions (which I imagine as propositional goo in the abstract ether) (Coherentism) or why it would refuse to bind to anything exactly but instead kind of float around certain propositions and theories that prove useful to human endeavours (Pragmatism). There’s certainly no mechanism that anyone has identified, so far as I can tell, because nobody knows the scientific laws of this postulated abstract realm! (Also, if truth is a real entity, does accuracy also exist in a similar fashion? Correctness? Validity? Rightness? Do they somehow reduce to one real entity? I’m sure that someone has examined these questions, but I would bet good money the answers are wack.)

Deflationism is the ‘theory of truth’ that supposedly one turns to if one does not have such commitments to abstract reality. But despite its anti-metaphysical positioning, it also proposes a way of understanding TRUTH that tries to explain away common usage as some kind of mistake. Deflationism basically says that “is true” is a redundant predicate, because you’re already “asserting” whatever it is you’ve uttered before you add “is true”. This wonderful trick goes by the name the “Equivalence Schema”, and it’s literally a triviality (at least in its most common version: https://stanford.library.sydney.edu.au/entries/truth-deflationary/). This theory accounts for usage cases where people describe someone as “speaking truths”, or say “That is a true statement”, by saying that the people in question are endorsing (or asserting(?what does this mean?)) the propositions behind those utterances. 

Incidentally, it kind of seems like this proposition entity plays a big role in the Deflationist theory, and, whilst I could be wrong, it just struck me that there’s something problematic about this. The very idea behind Deflationism – in particular, the idea behind showing the possibility of removing truth-language - is similar to the Expressivist move in meta-ethics to try to show that statements about “good”, “bad” (etc) can be translated into sentences which merely express sentiments, and other similar moves in philosophy. The thought is, “If I deny that X is a real entity, I have to show that communication makes sense without it.” Mutatis mutandis (as the great Jerry Fodor used to say), the proud placement of this “proposition” word in the Deflationist theory seems to suggest a metaphysical commitment to the reality of propositions. And yet, it seems kind of weird that if you don’t believe truth is a real entity, you’d nevertheless believe in the abstract reality of propositions, given that the proposition is usually defined as the truth-bearing entity. Maybe there’s some sense to the idea of propositions existing and truth not, or maybe there’s some way of re-describing proposition in the Deflationist theory using other terms that imply ‘ontological commitment’ to entities that are less troublesome. But it sure seems a mess!

So, basically, not really a fan of Deflationism either.

You might reasonably ask: if I’ve rejected all these theories, then what could I possibly believe? Does this not leave me in the foggy hinterland of Nowhere, stranded, lost, directionless?  My own view, which I accept may be irrational somehow, is that we don’t need to propose an alternate ‘theory’. I think that the project is flawed! 

The following undertaking is a related project which I think is slightly more useful: an analysis or account of the TRUTH lexeme/concept that explores the many different functions served by the word/concept in different contexts, and the different things that can be said about what justifies an attribution of “truth” in these different contexts. As the Deflationist theory suggests, truth, like its sibling words – accuracy, correctness, validity, rightness – is used in the context of speech acts relating to endorsement or approval, and I believe it is interesting to examine the different rules necessary to induce this endorsement in different social games. It’s also interesting to explore what happens when you approach certain areas of discourse without the traditional philosopher’s penchant for linguistic imperialism (it’ll become clear what I mean by that in the course of events).

So, without further ado, here’s my own analysis of TRUTH.

Preliminaries:

It seems to me that having an anti-metaphysical take on truth renders it impossible to separate your analysis of TRUTH from your ethical beliefs about when people should deploy this TRUTH lexeme. If you are one of these ‘realists’ about truth (as in, one of these believers in the realm in which theories and propositions gallavant and frolick, conjugally bonded for eternity to either Truth or Falsehood corpuscules) then many deployments of TRUTH become ‘wrong’ in the profound sense that they are factually incorrect. On this view, when people attrbute truth to a proposition that has none, they literally misdescribes the abstract reality (incidentally, this raises the question: is there are a meta-abstract realm which grounds statements like “It is false that the proposition The moon is made of cheese is true.”? What about “It is true that it is false that … is true””? An infinite regress of such realms?). But if you’re less inclined to join this tribe, then the only ‘wrong’ you have at your disposal is the familiar ‘wrong’ from ethics, rather than this more intuitively powerful normative force of misrepresentation of abstract reality. That is why I must begin by laying out my own view on the ethics of TRUTH usage.

The only way I can think to express my own normative stance on these special words is using Wittgenstein’s term “language game”, which I define as any “form of life” (roughly, set of social rituals enacted in a certain cultural or physical context or set of such contexts) with a characteristic associated language/discourse. Here are some suggestive examples: number theory, calculus, propositional logic, conversations about sport, being at church (/worshipping), small talk, philosophical seminars, philosophical ethics, metaphysics, physics, chemistry, university lectures, literature discussion groups, gaming, Dungeons and Dragons, making love. (No requirement of non-overlap.) I define a language game to be “epistemically complex” if it commonly involves lexemes like TRUTH, ACCURACY, CORRECTNESS, VALIDITY, and if it relies on a sophisticated set of non-random principles or rules of justification/proof to warrant the invocation of these lexemes (that is, it has a sophisticated internal logic for deciding the utterances/statements/assertions/theories which get the “truth” label (or similar labels such as “correct” or “accurate”), versus those that don’t). I also rely on a further distinction between “formal” and “informal” language games of this type (this distinction is bimodal but continuous), where the “formality” refers to the precision of the proof rules, so that the best example of “formal” language games are the language games of mathematics or logic, and all other languae games are on a continuum away from these. Making heavy use of these categories, I can sum up my normative/ethical take as follows:

i) for any statement, assertion, sentence or theory produced within the context of an epistemically complex language game (ECLG), the question of its truth (its warrant to be called “true”) might as well be determined by the conditions/rules appropriate to the language game in which these utterances/notions are produced, whether these rules be formal, not-at-all formal or in-between;

ii) that for informal ECLGs, any statement acceded to by a large majority is clearly not false (at least, from the point of view of someone making her judgment using the appropriate rules), whereas in formal ECLGs, this need not be true because a large majority could (in principle, at least) be following the rules incorrectly, or at least not seeing the potential for new rules that everyone will ultimately agree match the spirit of the game (where the game is in this case a mathematical/scientific enterprise) (this statement makes most sense in mathematics and begins to make less sense as we move into the sciences, but I hope the idea can be understood));

iii) that some ECLGs ought to be altered, replaced or even destroyed on the grounds that this may help bring people towards a more scientific understanding of certain aspects of the world around them.

The last of these requires exemplification. A good example is any community of believers in ‘woo’ or abject mysticism – e.g. believers in ghosts – who, necessarily, aren’t relying on scientific or scientifically adjacent justification rules. If possible without causing strong negative outcomes, it seems to me morally desirable to try to alter or dismantle the language games in which these people promulgate these ideas, because I think the world would be a better place if fewer people believed in such things. 

Ethical stance articulated. Now onto the mains...

Overview

It seems to me that one of the most crucial things to note about TRUTH is that there are reasonable and socially effective invocations of this lexeme that, in their semantic context, suggest divergent things about relationship of the concept of truth to justification and epistemology. I think this is part of the reason as to why these different ‘theories of truth’ cropped up to begin with.

This can be observed even within the context of scientific discourse: I will show that standard ways of attributing the property of “truth” to a scientific statement or a theory can’t be unified in terms of the underlying ideas about how truth (the concept) connects to epistemology. I will also illustrate that in ethical discourse the TRUTH lexeme/concept has a very different, and still interesting and valid, role. In illustrating this, I will defend my (ethical) view that philosophers have no right to declare that all applications of TRUTH in ethics are wrong or bad. Later, I’ll try to clear up some confusions left over by this discussion. 

As the dessert course, I’ll supplement the criticism I made of theories of truth earlier with a full-frontal assault on the ‘Correspondence Theory of Truth’ specifically, by arguing that many commonsense ‘truths’ which purport to be about the external world are merely conceptual (meaning that the class of statements for which a correspondentist analysis is at least potentially appropriate is smaller than is commonly believed).

Part I: Science is True and not True, but not Both at Once

I had to start my analysis by laying out normative principles because people wield this TRUTH lexeme in mutually incompatible ways, and one needs a way of adjudicating between them. As you’ve seen, I tend to have a chill ‘peace and love’ vibe about this whole thing. My aim is to show that truth talk is so messy that any alternative would lead to truly crazy conclusions about the constraints we ought to impose on language.

When I say that people wield TRUTH in mutually incompatible ways, the “incompatibility” I’m referring to is the implication of the relevant utterances about the connection of truth to epistemology. Crucially, I’m not just saying that people from different walks of life exhibit this incompatibility when compared against each other, but that it is fairly natural for even the same person to use the word “true” in these mutually incompatible ways. And I can illustrate this with an example of how science people talk about scientific theories.

The following seems to me an unremarkable sentence: “Einstein’s General Theory of Relativity is true – it has survived over 80 years of experimental test and produced predictions that are being confirmed even today (e.g. gravitational waves)”. One can easily imagine a physicist or someone with an interest in physics, saying something like that to introductory students, or maybe to a crowd of people who are ignorant or distrustful of modern physics. In fact, I believe I’ve heard utterances very similar to this. The following also seems to me an unremarkable sentence: “We have good reason to think that Einstein’s theory isn’t absolutely true, because in current form it can’t be brought into harmony with quantum mechanics and because of the ‘pessimistic meta-induction’.” One can imagine a similar person – even the same person, perhaps – saying this in the context of a philosophy of science seminar. Again, I think I’ve heard utterances very similar to this (because in fact it is true to say that GR is incompatible with QM, because there is no gravity in QM and there are apparently technical reasons why we can’t just dismiss gravity as “an emergent property”).

But obviously, if you’re being pedantic about truth, the two sentences are at odds! In the first situation, the use of this imprimatur, “true”, supposedly has to do with the strong empirical status of the theory (giving correspondentist vibes). In the second situation, “true” is being used in a way that’s favourable to a kind of Peircean account, which says that a true theory is a theory that’s unbetterable at the “limit of inquiry” - on which account there are arguably no true scientific theories at present! 

Karl Popper, the philosopher of science whom scientists tend to like the most, was a Peircean about truth in this way, arguing that scientific theories aren’t true exactly, but that they merely approximate the truth in the sense that they logically imply a greater number of true propositions than their predecessors (https://en.wikipedia.org/wiki/Verisimilitude). I somehow doubt, though, that all scientists who admire and claim to agree with Popper would want to put away the truth language when talking about scientific theories, especially given that we all, as humans, tend to apply the truth predicate to many statements and notions that don’t remotely have the epistemic status of our best physical theories (I’ll take that as an obvious observation, though it’ll also be demonstrated amply by later discussions).

Again, my point in raising this is normative: it seems to me that it would be very silly and wrong to try to correct people when they talk about our best scientific theories in a given field as being “true”. These uses of truth are both fine, in their respective contexts! This kind of incompatibility might become weird and confusing if a person mixed up their use of TRUTH in the context of a single lecture, say, but otherwise it doesn’t pose a problem.

Part II: Contra the Philosophers, Ethics is a Complex Game Worthy of Ethical Respect

Outside of the Ivory Tower of academic meta-ethics, it would be extremely peculiar to describe as either “false” or “not true” the sentence “You shouldn’t murder your own children”. Supposing you put that proposition to the average person (in a way that didn’t make you come across as mentally unwell or whatever) and asked whether it was “true” or “false”, I suspect that the response, idealising away the existence of trolls and other assorted louts, would be some variation of “Yes, that is certainly true!!!” (Perhaps this claim ought to make you angrily cry Unscientific appeal to psychological speculation! But somehow I suspect that will not be the reaction. And I think everyone can accept the basic point that this statement seems quite obviously morally ‘right’ or ‘correct’ or, at the very least, ‘conventional’.)

And yet many philosophers think that people are ‘wrong’ to think this way!

One class of philosophers who think this way go by the name Expressivists. Expressivists claim that when people make moral judgments like this, the literal meaning of the words is basically irrelevant, because they’re really just expressing an attitude or emotion or both. Since expressions of attitudes or emotions are not the kinds of things to which we could conventionally attribute “truth” or “falsehood”, they argue that the best approach is to declare all moral judgments to fall in the class of utterances which are “neither true nor false”.

I don’t agree with this view at all. Whilst the logic makes sense, the claim that the only thing the average person is doing when they make such a statement is expressing an attitude or emotion seems to me an empirical claim falling within the domain of social psychology, and, interpreted as such, it seems totally silly. Even ordinary people get into arguments about morality; my experience is that humans overwhelmingly think there are rights and wrongs to evaluating rights and wrongs. And ordinary people clearly have no problem using the words “true” and “false” in evaluating moral judgments like the one mentioned. This suggests that, in addition to being expressive or sometimes expressive and commissive speech acts, expressions of moral judgment are typically assertions or assertoric, i.e. put forward as “facts”.

Further, it is clear, as innumerable authors have pointed out, that it is common and perfectly felicitous to conditionally embed ethical judgments as if they were propositions (“If killing elephants is wrong, then killing whales is wrong”), which obviously poses problems for expressivism. Clearly, you can’t successfully ‘translate’ this sentence into a form where the speaker is just talking about their emotional states! 

This seemingly takes away the reasons to say that moral judgments are “neither true nor false”.

Another clan, going by the name Error Theorists, take the view that, since people think they are asserting “truths” or “facts” but are actually using language that has no place in any modern science, the correct reaction is to think, “They are wrong, they are stating falsehoods”. I think that this kind of argument makes a kind of intuitive sense; I also think that to declare oneself an Error Theorist is sort of sexy (you are part of the small community of people who realise that people are actually talking nonsense when they talk about morality! And everyone else is a naive fool!). The issue I have with it is that I don’t believe that the fact that Darwin banished cosmic purpose from the scientific worldview means that you have to be so supercilious and snooty about moral language games. In fact, I think it’s a bad look. 

… You might say this is not an argument, and you would be right, but that would be to miss the broader point that Error Theorists’ arguments don’t themselves attempt to argue that we should call ethical judgments “false”, or that we should think of them as false; they make no argument that we need to collapse all language games into one. They merely argue that ethical statements don’t, under the harsh light of logical analysis, seem to be talking about experimentally detectable or otherwise perceptible properties of the fully mind-independent world. They think the normative implications of this are obvious presumably because they are ‘realists’ about truth, meaning that these ethical statements are as bad as any other factually incorrect statements, in that they clearly don’t express propositions which match the propositions bonded to the truth substance in the abstract realm. But without this assumption, this line of reasoning doesn’t at all imply the normative claim that we should call or think of ethical judgments as “false” (an Error Theorist knows she can’t justify any normative claim, by definition.). Even if I agree - and I do - that Darwin showed that the already logically questionable theory of agential design of life is completely unnecessary to explain the wondrous phenomena of life, and that “good” and “bad” do not exist in the fully mind-independent world, it doesn’t force me to judge negatively the usage of words like “true” and “false” in ethical/moral contexts. In fact, this insight makes no difference just so long as I am able to keep in my mind that ethical language belongs to a different language game or games to, say, fundamental physics, and that it doesn’t undermine science for different parts of natural language to have their own truth conditions appropriate to the game in which we use the language. This is, of course, similar to a standard objection to Error Theory and Realism made by expressivists.

Clearly, Error Theorists don’t think it’s a problem that they would come across as batshit insane were they to say to their family at Christmas “I believe the claim that “You shouldn’t kill your own babies” is false.” But the fact that saying this definitely would come across as nuts, or at least a bit kooky, is indicative of the fact that the Error-Theoretic position involves applying standards to moral language games which are totally inappropriate to use. Why can’t we have our ethical debates and discussions even if good and bad aren’t properties that have reality in the fully mind-independent world? We clearly can! After all, even those of us who do have this metaphysical view still engage in ethical debates and discussions.

I can imagine someone potentially pushing back on this way of thinking using the ethical stance I myself laid out in the Preliminaries, in particular, (iii). If one were to successfully demonstrate that the language game of ethics is epistemically primitive, i.e. that ethics is not an ECLG, or that it’s a highly informal ECLG – that is, that the rules for determining felicitous attributions of “truth” and “falsehood” to ethical statements/judgments are haphazard, random and highly subjective – then this might give one an (ethical) reason to try to undermine this language game entirely.

But I believe this view is closed off in ethics, because at least certain kinds of ethics language games can be very complex and rather formal. To put it another way, the other problem with both these Expressivism and Error Theory is that they make it seem as if the rich system of norms that sophisticated, intellectual moral discourse involves – with logic playing a central role – is just some kind of “illusion”. That is, I think that both these theories of ethics are bad at ‘accounting for’ the language game of academic ethics - in particular, what I believe to be the true proposition that some ethical arguments are good (rigorous, solid, plausible), whilst others are bad (incoherent, unsound, reliant on intuitions that almost nobody shares). The word “true” is the main word academic ethicists, and other people involved in sophisticated ethical debates, use to demarcate those moral stances that have on their side the strong weight of relevant considerations from those that don’t. Consider the following.

Analytic ethics survives because (crudely) some ethical judgments are more logical than others. In slightly more precise terms, I have developed the following simple model of ‘rational ethical discourse’:

“Person A agrees with person B that x (where x is some ethical principle or a highly general ethical judgment). Person B points out that person A is violating x in the case of y (where y is some specific issue: women’s rights, race, animal welfare, abortion, whatever). Person A argues that position on y is not a violation of x because of an error in Person B’s argument, or because of empirical considerations which Person B has overlooked. The debate either continues with furthe discussion of the merits in each other’s arguments or onto further discussion of the empirical considerations.”

(A similar model of rational belief-resolution, modified so that Person A and Person B become two perspectives within a single individual’s mind, goes by the name “reflective equilibrium”.)

It may be argued that, even if such a model of civil debate captures what occurs in the ‘best’ cases of ethical disagreement, the level of ethical disagreement that exists in our world suggests that typically such debate doesn’t work very well in helping people change their mind. This might in turn suggest that this style of discourse does not face adequate constraints of precision to meet an intuitive standard for discourse that can reasonably be called “truth-seeking”. In other words, it is not a stable-consensus discourse, where this term refers to a discourse with sufficiently precise rules of inference that we should expect rule-following participants in the discourse to eventually reach the same conclusions (or, in the language of the Preliminaries, it does not belong to a sufficiently formal ECLG). Even if it is possible for people to sometimes make some degree of ‘progress’ on ethical conflicts by working out that one of the views would require a conflict with a principle that they both hold dearly, it doesn’t happen often (or so the imaginary critic says). More damningly, it seems like a process that is inherently slippery or gameable, given that ethical principles, unlike mathematics theorems, are vague, natural-language formulations which do not have completely precise implications. For example, “Murder is bad” may seem an unambiguous principle - a bit like a mathematical theorem or logical tautology - until you recognise that:

(i) people don’t always agree in classifying events as “murder” (e.g. there is disagreement about whether abortions can be cases of “murder”, and there is often ambiguity between manslaughter and murder); 

(ii) the word “murder” is stretchy, especially in a war context, where if people view certain acts of killing as righteous, they tend to not want to call them “murder”, even if they’re probably or obviously outside of the laws of war (e.g. there are certainly disagreements in our society about whether dropping bombs which predictably kill civilians makes the orchestrators complicit in “mass murder”, and I believe many would bristle at calling the assassination of Bin Laden “murder”, even though it was an extra-judicial killing outside of the context of a war); 

(iii) it’s not immediately obvious how to weigh the “badness” of murder against other things (so we might imagine people having different intuitions about how much non-lethal maiming an individual would have to mete out to have done something equally as “bad” as committing a murder); 

(iv) people can wriggle their way out of their apparent commitment to a principle like “Murder is bad” by saying “Yes, murder is always bad, except in this case”; 

and so on. 

So the very words we use to describe states of affairs in the world on which we’re casting ethical judgment - e.g. “muder”, “charity”, “nurturing”, “genocide”, “rape”, “liar”, “caring” - are themselves debated endlessly due to their fuzziness and temporality and emotion-ladenness. 

Of course, the problem with ethical discourse being gameable in this way is that it would seem to undermine the idea that we can reasonably make judgments about the “rationality” or “well-motivatedness” or “validity” of ethical judgments, because it suggests that the reasons people believe they have for their ethical judgments or claims could always be called into question by someone operating according to the most ‘proper’ standards of ethical argumentation. It suggests, in fact, that two people both abiding by the most elite standards of ethical argumentation might nevertheless never reach an agreement on a case, because, e.g., they keep refining their principles to account for counter-examples, instead of changing their mind on cases, or define words in different ways and refuse to budge. None of this is even to mention disagreement about overarching ethical frameworks, which lead to subtly different principles and ways of thinking being invoked. In making the comparison to logic or mathematics, people often compare these frameworks to differently axiomatised systems in those fields. This might almost seem an ‘out’ for the idea of ethical reasoning as analogous to formal reasoning, except that I think that this form of persistent disagreement I’ve been describing occurs even between people who have a similar or the same overarching framework. Evidently, this is unlike the reasoning of logic and mathematics, where the proper standards of argumentation are sufficiently constrained (within the given proof system) so as to leave no ambiguity about the right conclusions, once the relevant proofs have been arrived at. (Another way of putting the disanalogy is that, in ethics, we never arrive at decisive proofs at all!)

The best response to this rebuttal may be to concede that whilst ethical discourse, even at its best, is not a stable-consensus discourse – in other words, the ECLG that is academic or sophisticated ethics is only moderately formal at best - it may still be a discourse with sufficient constraints that people can be shifted along the ‘landscape’ of ethics by the perturbations of reasoned argumentation, i.e. that instead of one equilibrium state, there are finitely multiple. Different ‘initial conditions’ may lead to different results, but it doesn’t mean that the discourse isn’t consensus-achieving in certain areas, or among communities of people with similar dispositions, or whatever. There are at least reasons of a kind involved in ethical discourse, and people can definitely respond to them (as I know I have on this or that issue). I think this is enough to make it justifiable in many cases for people to use words like “rational”, “valid” or “well-reasoned” with respect to ethical arguments or claims.

And, as the philosopher Crispin Wright has noted, certain ethical statements do meet one interesting criterion of discursive equilbrium that he has dubbed superassertibility. As I understand it, a superassertible statement is one that is warranted by the standards/rules/proof procedure of the discourse or language game, and could never become un-warranted unless those rules substantially evolve. Superassertible statements in a less formal ECLG are sort of like the axioms of formal ECLGs. “Murder is bad” is perhaps one such example in ethics; despite how fuzzy that statement actually is, it’s inconceivable that saying it could ever become seen as “false” or “distasteful” or “repugnant” or whatever, without a substantial shift in probably human nature itself.

To sum up, whilst I don’t subscribe to the view that all, or even most, ethical controversies do admit of an obviously more rational verdict – nor Parfit’s view in On What Matters that we should be seeking a general theory of ethics – it is an important fact to bear in mind that the tools of logic and generalisation are often meaningfully deployed in ethical discourse, and that ethical discourse can be consensus-promoting at least some of the time.

This seems to me both interesting in itself, and also an ethical justification for not sermonising that people are making a ‘mistake’ or ‘error’ in using truth language in these contexts. Respect the complexity!

(It’s an interesting question to what extent similar remarks as those above might be applied to aesthetics. I will not explore it here because it’s actually a kind of massive topic. I do at least believe that there is such a thing as sophisticated aesthetic discourse which relies on reasons, but we think it’s probably significant that many fewer people call themselves aesthetic realists than call themselves moral realists. I think there are not any statements exactly like “You shouldn’t kill your own children” or “Murder is bad” in aesthetics… (maybe “Michelangelo was a brilliant artist” comes close… but clearly it’s not quite the same).)

Part III: This is Not Relativism

I understand that many philosophers would be deeply concerned by the way I have defended “ethical truths” (even if they agreed with the bit about rational ethical discourse), since the reasons I adduced in favour have at least a ring of “relativism” or extreme scepticism to them. The idea that (as they would put it) what is true is determined by the utterances that people overwhelmingly agree to call “true” may strike the average hard-headed analytic philosopher as worryingly Rortyan (and the layman as worryingly Postmodern NeoMarxist). 

As it happens, I do not see my remarks on the possibility of “ethical truths” in the relativist or extreme sceptic light at all. Although these views I have expressed may have a lot in common with Rorty’s (I’m not sure – I read parts of Philosophy and the Mirror of Nature before casting the book aside because it was too vague, unclear and unpersuasive), I am very much much in the camp of the “anti-relativists” in terms of tribal loyalties and identify as a scientific naturalist. In terms of my views about metaphysics, I think of myself as a neo-Positivist or neo-Positivist-Pragmatist, and my main philosophical influences are Hume, Quine, Dennett, and Ladyman and Ross . Not only that, I am partial to speech acts like “It is a form of insanity to deny objective truth”, “There is no such thing as personal, subjective truth”, “Science converges on truth” (etc), and I tend to be hostile to speech acts like “There is no objective truth”, “There are other ways of knowing apart from Western science”, “Il n’y a rien dehors de la texte”, “Nietzsche was the greatest philosopher to have ever lived” (etc).

‘But,’ our imaginary maddog realist asks, ‘How can this be?’ 

I don’t think that the above implies I have an incoherent set of beliefs, but I can understand why some philosophers might think so, because it feels that way, even to me (I worry myself that I have fallen into inconsistency). The solution, in my mind, is this:

(I) I do believe that language is useful for exchanging information about mind-independent phenomena – that it can and often does serve this purpose. My rejection of the Correspondence Theory of Truth doesn’t rule out the belief that some locutions ‘represent’ reality in at least the minimal sense that they give us information about the world that we can use to make accurate predictions. But there’s a whole other class of locutions that merely give us information about our interlocutor’s mental state. I think that truth language, because it’s related to endorsement, falls into the latter class.

(II) I think that one can fully justify the epistemic supremacy of this theory over that – e.g. of contemporary fundamental physics over Aristotelian or Newtonian physics - without leveraging this truth notion directly. Obviously, it may be common in philosophy of science to say such things as “Einstein’s Theory of Relativity implies more true propositions about the universe than Newtown’s mechanics”. And in doing this kind of epistemology – in making discriminations between empirical theories - one clearly must lean lean on the notion of a successful/accurate/true prediction. But the centrality of the truth language here does not mean that truth exists!

First, with respect to the Einstein claim, I think one can say essentially the same thing in a different way: “Einstein’s Theory of Relavity implies more physical implications that have been verified by experiment.” Crucially, this move effectively leaves us only having to deal with the notion of a successful or true prediction. Meanwhile, I believe I can analyse this notion without painting myself into a corner, as follows: a successful prediction consists in some event or set of events where some scientist precisely describes some model or makes some precise statement, and some later event or set of events where some kind of observer observes a certain configuration of phenomena that match the model or earlier statement. Done. Nothing I have said rules out my ability to consistently use the language I just used. (As always in philosophy, there is an endless amount to be said about the words used in trying to analyse other words.. For example, can I analyse this invocation of the word “match”, or was that just a covert replacement for “successful” that really gets us no further? Well, it just so happens that I think I can partly unpack the notions of ‘matching’ or ‘representing’ in terms of information theory, but we shan’t go there. My point is that I don’t think my view’s anymore susceptible to attacks on these sort of ontological grounds than the ‘theories of truth’.)

This does not mean, as also indicated before, that the word “truth” is dispensable. It’s an immensely powerful imprimatur of ideas we want to endorse, and gives us shortcuts for agreeing with people (e.g. we can say “Everything X says is true” rather than asserting each statement separately). But the point is this: denying that “truths” are ‘out there’ (implying that huge swathes of language, and many forms of life, are ‘wrong’), does not mean being a relativist or extreme sceptic. 

Well, that’s the start of the explanation at any rate; I understand that the concerns may persist. Here is an obvious concern. There are contemporary communities, and throughout history there have no doubt been societies, where the majority of people, even the ‘smart’ people or ‘intellectuals’, apply the word “true” (or its nearest-equivalent in their language!) to sentences or theories which no intellectuals today would apply this word to (“The Earth is flat”, “The body has four humours”, “X is a witch”, “God created the world in 7 days”, “God gave us souls”, etc). And the logic I have relied on hitherto implies that these sentences are true within these communities (for these people), or were true within these communities or societies (for these people), if the majority of these communities do/did find it felicitous to call them “True” (or, to put it in a way more concordant with the spirit of our thinking, that they at least weren’t misusing any language, and that their contemporaries probably didn’t have good reason to try to correct their language). If you reflect on this for a second, there’s actually nothing scary about this logic at all, because within our community we know that these beliefs are false; they are not true for us, and we have good reasons for declaring them false. We have good reasons for perhaps even calling these beliefs “absurd” and so on when we talk amongst ourselves; depending on the circumstances, we may have good reasons to try to tell the superstitious communities among us directly that their beliefs are false according to our, superior epistemic standards (the restriction is that we ought to stay in accordance with the principles I laid out in the Preliminaries). And is this really such a traumatic conclusion? Are you realists not satisfied? …

Actually, there is one more thing I must deal with to make sure that I am understood. As I made clear much earlier, I regard logic and mathematics, and also physics, chemistry (etc), as “language games” in themselves (again, the ‘language game’ concept is for me very stretchy - for simplicity we can think of “mathematics and logic” as one language game in this context). You can only play these language games properly if you obey the rules (the rules in logic and mathematics are usually formal and precise, obviously (axioms and simple rules of inference, or more complicated deduction rules that everyone agrees on), whereas the rules in the sciences are more complicated), and people are right or wrong about their statements or claims and (trivially) therefore also right or wrong about their ascriptions of “truth” and “falsity”. The ethical principles I have put forward do not imply that there is something unethical or misguided about a mathematics professor telling a student that they are wrong when that student is indeed wrong. And it does not imply that mathematical truths are relative in any way shape or form. In fact, in mathematics, the rules are sufficiently precise (even before deduction rules were formalised in mathematical logic by figures like Frege, Russell and Hilbert, and even before we realised that the ‘top-down’ Euclidean approach to founding mathematics doesn’t work [Lakatos] because we don’t have enough absolutely self-evident axioms) that (as far as we’re aware) the major results in the history of mathematics conjectured to be true and later disproven were actually false according to the same deduction rules and technical-term definitions by which mathematicians were operating when most of them thought they were true. 

Obviously, science language games are much different than mathematics and logic: the whole system of standards, the rules of the game, have evolved massively over time (especially, of course, with the Enlightenment and Galilean Revolution); standards of rigour continue to shift and change (consider the internal battles in the social sciences, or the String Theory Wars in physics); and technical-term definitions change with every big theoretical upheaval. For these reasons, it’s hardly surprising that when scientists and philosophers of science talk about science, they will use the word “true” in ways which imply different things about the relationship of truth to justification. But, again, similar would apply to mathematics, insofar as the rules in certain science language games are sufficiently constrained that one needn’t be a relativist about game-specific falsehoods – e.g. I can happily endorse labelling as “falsehoods” the ramblings of a crank physicist, because what makes him a crank is that he is not following the rules of rigorous physics (the language game of academic physics)!

Part IV (Dessert): “Grass is green” does not Represent Reality

I have a critique of the Correspondence Theory of Truth that I believe is somewhat novel (probably not really novel, but I think unconventional to be sure). It turns on the observation that the number of felicitous, socially effective and natural applications of the word “true” for which the Correspondence account, despite all its vague generality, makes no sense whatsoever is much larger than people conventionally think. 

I used to be heavily invested in the field of cognitive science and its various sub-fields and offshots. There’s a simple theory in cognitive linguistics which goes by the name Prototype Theory (https://en.wikipedia.org/wiki/Prototype_theory). This Theory basically states that human concepts of commonsense entities such as household objects and animal species are (a) internally graded and fuzzy, and (b) that we have strong intuitions of the belongingness of entities to given categories based on the perceptual ‘distance’ (morphological/chromatic/textural difference) of those things from the prototypical exemplars of the categories in our heads, almost as if we model belongingness to a category against a single central ‘model’ or ‘pattern’. In my view, this Theory seems pretty accurate and is decently well-supported by experimental studies. 

A very common syntactic form in English is the subject-predicate sentence form where the subject is some simple taxon and the predicate is some simple everyday property. Examples of such sentences are “Grass is green” and “Trees have leaves” or “Bananas are yellow” or “Elephants are big”. I mention this subsequent to my introduction of Prototype Theory because I think that what makes these sentences seem true is that the prototype of the subject-concept in our head has the property mentioned, even if the instantiations of the taxon in the external world often do not really have the property in a robust or consistent way.

As I just implied, those examples I inserted above seem to me good examples of sentences that people intuitively and overwhelmingly believe to be true. I cite as evidence, for the first at least, the fact that this sentence is commonly used in logic courses as an example of a “true proposition”. Despite this status as True Sentences, it seems to me that the more you reflect on such sentences, the stranger their intuitive truthiness seems. For starters, it is evidently not true that all grass is green, nor that all trees have leaves. In fact, it seems to me that most people would agree that the assertions “All grass is green” and “All trees have leaves” are false (the third of our original examples, altered in this way, seems also false and the last one becomes at the very least not obviously true). One could try to defend the ordinary truth evaluations on the grounds that the first two assertions are empirical generalisations - although one must concede that, functioning as generalisations, their truth conditions are extremely vague and probably elude precisification altogether (the majority of the surface area of every single identifiable, living organism falling within the Family Poacaea is green? This is silly for a million and one reasons.). The key to solving this mystery is understanding the psychological explanation for our attribution of Truth to these sentences, which goes as follows: 

Mental prototypes have determinate properties. Sentences which have a very superficially similar grammatical form are evaluated for truth against different propositional ‘architectures’. We evaluate the truth of the claims “All grass is green” or “All trees have leaves” against a broader background or situating `architecture' of alternative propositions – drawing on more background knowledge – than we do for the claims “Grass is green” or “Trees have leaves”. “All grass is green” comes out false because the universal quantification forces us to think beyond the prototype, about all the instantiations of the taxon “grass” and their various forms, and the same for “All trees have leaves”. The claims that “Grass is green” or “Trees have leaves” are, nevertheless, we say, true. Why? The answer is Prototype Theory. The first comes out true because the GRASS prototype, which seems to be describable as something like “healthy, lawn grass”, can take only one colour property, and it is GREEN, not red, not yellow, not blue. Similarly, “Trees have leaves” comes out true because the prototypical tree in our head has leaves. There’s no ambiguity because the prototype can’t have and not have leaves; the prototype has determinate properties. Despite what philosophers have thought since time immemorial, these propositions are really conceptual truths, not truths about the world. 

And so now we see where the Correspondence Theory comes in. Most advocates of the Correspondence Theory would claim that Grass is green is a classic example of a proposition whose truth is ‘explained’ or ‘accounted for’ by their theory… but it’s not, because the correct explanation is merely psychological! Furthermore, I think that this point cuts across a very large amount of language. Sure, it is true that grass is a real enough entity and greenness is a reliable property to attribute to things, so thinking about grass and greenness is not like thinking about unicorns epistemically, but insofar as what could possibly explain why we endorse these clearly true sentences with the word “true”, the “Correspondence Theory” is a non-starter.

I can imagine an objection to this which concedes that the proposition expressed by the statement “Grass is green” actually pertains to that subset of grass in the world which is close to the prototype in our head (i.e. when we say “Grass is green” we express something closer to Healthy lawn grass looks uniformly green, at least from a distance) and yet that this is compatible with the Correspondence account because the Correspondence account explains the truth of the actual proposition/s being expressed by the statement. This is probably the best move (rather than saying Grass is green is actually false, though I can imagine that move being made also), as it comes close to forcing me to retreat to my other critiques of the theory. The way I would push back is by re-asserting that this truth is in fact merely conceptual. When we say Healthy lawn grass looks uniformly green, we’re still just talking about a concept in our head, which is inherently defined by greenness. It’s like saying “Green grass is green”. 

I’m sad that I had to finish by revealing my belief that the category of “conceptual truths” is far larger than most philosophers believe – which hints at my broader perversions in the area of semantics – but philosophy takes us to strange places. Sayonara.