Truth on the edge
A brief Western philosophy of the
This book offers a broad introduction to Middle Way Philosophy. The contents and first chapter (which has the same title as the book) are included on the web here as a taster. If you want to read further, please buy the book (available in paperback or as a pdf download).
Please scroll down for footnotes rather than clicking links.
Link to a diagram of the argument
When men have realized that time has upset many fighting faiths, they may come to believe…that the ultimate good desired is better reached by free trade in ideas – that the best test of truth is the power of truth to get itself accepted in the competition of the market, and that truth is the only ground on which their wishes can safely be carried out. That at any rate is the theory of our Constitution. It is an experiment, as all life is an experiment.
Oliver Wendell Holmes
1. Truth on the edge
Our lack of direct access to truth
The meaningfulness of truth on the edge
Why ‘truth’ should not be pragmatically redefined
A priori appeals to truth
Logic and hypothetical appeals to truth
From truth to justification
2. Metaphysical agnosticism
The features of metaphysics
Opposed dualisms in metaphysics
The definition of metaphysics
Metaphysics and falsification
Metaphysics and evil
3. Objectivity and integration
Moral and scientific objectivity
Provisionality and confidence
The conditions for belief
4. Maximising responsibility
The dualism of responsibility
Resolving the dualism
Belief and choice
Reward and punishment
5. Integrative practice
Theory and practice
Integration of meaning and the arts
Integration of desire and meditation
Giving priority to integrative practice
6. Moral judgements
Virtues and judgements
Use of principles
Calculations of consequences
7. Scientific progress
Science and scientism
Public replicability and types of evidence
Science and moral issues
The meaning of scientific theories
8. Riding religion
Religion and metaphysics
God on the edge
Religious symbols and inspiration
Integrative practices in religion
9. Uniting the state
Obedience to the government
10. Integrative solutions today
World food crisis
Chapter 1: Truth on the edge
In this chapter, I want to offer an account of truth that challenges most of the habitual ways of talking about it, both in Western philosophy and in wider Western discussion. It is essential to get this groundwork clear before moving onto discussion of ethics, politics, or any other more immediately practical topics, because it is our assumptions about truth that shape our understanding of all other topics.
Our lack of direct access to truth
Since ancient times, both in Greek and Indian traditions of thought, thinkers known as sceptics have been pointing out the hollowness of all our pretensions to know the truth about anything. Very often these thinkers have been perceived as destructive and negative in intention, or dismissed as impractical. Even when they have been recognised as valuably stimulating further thought by questioning accepted dogmas, the full value of what they offer has often not been recognised because it was regarded as peripheral. However, the recognition of our lack of direct access to truth is exactly where philosophy needs to begin, for ultimately practical reasons.
Sceptics have offered many arguments, but here are a few of the key ones in summary:
· As finite beings occupying a limited point in space, the information that we have access to is always necessarily limited.
· Given our limited mental capacities, it is unlikely that the concepts we form are capable of accurate representation of reality.
· Our senses are limited in what they can detect (for example, we only see objects that reflect light between certain wavelengths), so we cannot gain true perceptions of objects through the senses, since we might be missing crucial features.
· Given evidence and arguments for one belief, alternative evidence and arguments that appear to support opposing beliefs can always be found.
· Our conceptual frameworks for understanding the world are limited by our cultural and linguistic background.
· No conclusive proof can be offered that one’s current experience (or any given past experience) is not illusory. You may be dreaming at this moment.
· Given how often we have made mistakes in the past and had to alter our beliefs, it seems likely that we will make more mistakes and have to alter them again. At least some of our current beliefs thus seem likely to be mistaken, and we do not know in advance which ones.
Such arguments do not prove any given belief to be false, but they crucially challenge all claims to certainty.
There is nothing intrinsically negative about such arguments, for such arguments work just as effectively against negative assertions as against positive ones. If I take these sceptical arguments seriously, for example, I cannot claim to know the truth either about God’s existence or about his non-existence, for the statement ‘God does not exist’ can be just as easily subjected to these different kinds of sceptical doubt as the claim that he does. Similarly, the claim that there is not a table in front of me is just as much subject to sceptical doubt as the claim that there is.
Nor is it ‘negative’ merely to challenge any given belief, for such challenges are essential for the improvement of beliefs. Sceptical arguments do not (as is often assumed) imply that we should give up all beliefs, only that we cannot be certain that our beliefs are true. If we believe our beliefs are true we must be deluded, and scepticism thus helps us get closer to the truth by uncovering deluded versions of it.
Sceptical arguments have only ever been challenged either by dogmatically asserting a supposedly unassailable, foundational truth (such as my own existence, or God’s existence), or by changing the meaning of the term ‘certainty’ to fit everyday usage. Descartes provides an example of the former approach by building his philosophy on the foundational certainty of his own existence. However, he could only ‘prove’ his own existence as certain by assuming that his self-conscious experience at a given moment was conclusive evidence for a permanent and unchanging self – a massive over-interpretation of limited evidence. G.E. Moore can provide an example of the other kind of challenge. He asserted that he was certain of the existence of his hand because of the lack of positive counter-evidence (I have no reason to disbelieve in my hand’s existence). This is a moving of the goalposts towards a more everyday sense of certainty, but no proof of absolute certainty that he had a hand.
Our lack of absolute certainty requires recognition too constantly to allow any distraction from the sceptics’ central points. Contrary to the assumptions of many philosophers, our lack of absolute certainty is of great practical importance. If we are to bear in mind the real possibility of being wrong, even when there is no immediate evidence available to us that we might be, or nobody present casting doubt on our ideas, then the possibility of being wrong needs to be at the very basis of our philosophical attitudes. History is littered with examples of false certainty, from the Crusades, the Spanish Inquisition, Marxism, and Fascism to the long-held certainty (challenged by environmental circumstances more recently) that the planet has inexhaustible resources meant for our use and an infinite capacity to absorb the effects of our activities. It is only the recognition that we might be wrong that can potentially restrain the hand inspired by false certainty in the future.
We do not have direct access to the truth, and what we often take to be ‘truth’ is merely belief certified by the group in which we live, or belief for which there is some positive justification. It is all too easy to acknowledge intellectually that we have no such direct access to truth, but then perhaps five years later, in the press of whatever expectations society imposes on us, to once again, let this point slip and start thinking of our beliefs as ‘true’. Alternatively we might start using the concept of truth, perhaps with a few theoretical hedges, as the basis of judgement: a unjustified move given our lack of access to truth. We cannot afford to leave the sceptics on the edge of the discussion: their contribution needs to be a starting point, constantly borne in mind.
There is no practical alternative, if we want to operate with a consistently adequate degree of intellectual humility, to the complete abandonment of claims to truth, whether negative or positive, or of any other kind of appeal to truth. Quite simply, we are not God, and we do not have access to the infinite knowledge that God, if he exists, would have. Thus we are not justified in appealing to truth.
Nevertheless, this does not mean that we cannot use the concept of truth meaningfully, or that we cannot apply justification, short of claiming truth, to our judgements. Truth has not disappeared below the horizon, but it is on the edge of that horizon, in a position where we can talk meaningfully of it in a rather abstract way, but not grasp it or rely on knowledge of it. We need to constantly adapt to the full implications of truth being on the edge. I will return below to the argument that justification does not depend on truth, but more immediately, next, to the question of the meaningfulness of the idea of truth.
The meaningfulness of truth on the edge
Once our lack of direct access to truth is recognised, one relativistic or nihilistic reaction is to dismiss even the meaningfulness of truth. From this nihilistic perspective we seem to be left in a philosophical wasteland, with both science and morality apparently reduced to total ruins, babbling nonsense to each other about long-discredited notions. But however unjustified claims that appeal to truth may be, sceptical arguments do not imply that talk of truth is meaningless, nor do they imply that there might not even be some use in talking about a truth that we cannot justifiably appeal to (as I have been so far in this chapter).
However, to avoid the conclusion that truth is meaningless, we need to examine the concept of meaning. If your concept of meaning is truth-conditional, like that of many analytic philosophers, then it is clear that without appealing to a concept of truth you cannot justify a concept of truth, and thus ‘truth’ becomes meaningless in any other terms, because only made meaningful in terms of itself. I will say more later in this chapter about why a truth-conditional theory of meaning does require an appeal to the concept of truth, even though it is conditional.
Truth-conditional theories of meaning must also be rejected for other reasons. They do no justice to our experience of meaning because they class as meaningless any utterance that is not a proposition that might conceivably be true or false (such as “How are you today?” Or “Waah!” which are not propositions at all). They also belong to a wider category of theories of meaning that I call Representationalist, because they take the meaning of a piece of language to be based solely on a relationship with some other state of affairs that it represents. Sometimes, however, language is also meaningful to us not because of what it represents, but because of what it expresses, as in the example of “Waah!” Instrumental music, which represents nothing, can also be deeply meaningful to us. Whatever theory of meaning we accept must do justice both to representational and to expressive bases of meaning, and to a complex relationship between them. Meaning is potentially both a cognitive matter (e.g. the meaning of a scientific paper depends on the beliefs and justifications it is putting forward) and an emotional matter (e.g. a cry or a curse is meaningful because it expresses the emotions of the person uttering it).
Fortunately, an alternative approach to meaning that can meet these criteria has been developed since the publication in 1987 of George Lakoff’s book Women, Fire and Dangerous Things. Lakoff’s account of meaning is a genuinely pragmatic one that begins with human physical experience. The meaning of the word “in” as opposed to “out” for example, is first grasped preconceptually in relation to our bodies. Other commonly encountered objects in our experience (such as a “tree”) are labelled according to basic categories that become pre-conceptual. More complex concepts depend on metaphorical extensions of these basic physical meanings to form schemas – so, for example, my understanding of the concept of incorporation depends on the schematic extension of my preconceptual sense of “in”, so as to include not just taking things into a physical body, but by extension an organisation.
Such a general approach to meaning implies that the meaning of truth depends not on a circular relationship to the possibility of truth itself, but to its metaphorical extension of our basic preconceptual categories. The metaphor that I think it would be most useful to adopt in our grasp of ‘truth’ is the idea of truth being the ultimate state of affairs marginal to the universe we assume ourselves to inhabit. It is neither whole in it (for then we could claim a better acquaintance with it) nor wholly outside it (for then we could not adopt the concept). Truth is on the very edge of understanding. As with an object that occurs vaguely at the furthest horizon of our sight, we can talk meaningfully of its existence, but would be mistaken in speculating too closely as to the ultimate nature of the vague and ambiguous forms we can perceive.
Such a way of understanding truth relates to our physical experience of marginality: of objects being on the edge of a field of some kind. This marginality relates not only to our cognitive construction of the idea of truth but to our emotional appreciation of its mystery, our humility in relation to it. Of course, to choose to adopt one metaphor rather than another in our interpretation of truth implies that we do possess a choice in our selection of metaphors – a point that I will be returning to in chapter 4.
Why ‘truth’ should not be pragmatically redefined
One common response to the power of sceptical argument is one that has become associated with pragmatism and postmodernism – a relativisation of truth. If we think of ‘truth’ as what we find it pragmatically useful to accept, we are in effect reducing the meaning of ‘truth’ to that used in much everyday parlance; in effect, that of justified belief. For example, if Fred says to Mary “Is it true that you really love me?” and Mary replies “Of course it’s true!” the sense of truth (even if we assume Mary to be thoroughly honest) is that of what Mary assumes to be true so far because of lack of further evidence, not that of absolute certainty considered apart from the emotional impact of the situation.
William James was perhaps the first thinker to define truth in this sort of way, comparing a truth to a bank-note which “passes” so long as those who use it have sufficient practical confidence in it. More recently many postmodernists have adopted it, and also pragmatists such as Lakoff, mentioned in the previous section. There are important reasons why we should not respond to the challenges to absolute truth in this way by making it relative in meaning.
Perhaps the most important is that to do so deprives us of a needful concept. We need a concept for truth on the edge. However, we already have entirely serviceable concepts for acceptable beliefs in a given context: we call these beliefs, or justified beliefs, and to call them “truths” reduplicates the terms already available for belief and takes away, or confuses, the term “truth” for truth on the edge when there is no alternative term for it.
The meaning of a term, as explained in the previous section, is not only to be understood in terms of what it represents, but also in terms of its emotional significance. To use the term “truth” enclosed in scare quotes, or to pluralise it as “truths” tends to have ironic, mocking overtones. We cannot easily adopt an absolute term and use it relatively without such irony. Such irony is unhelpful, because it undermines respect for the concept of truth and confuses it with the lack of justification for truth-claims. So long as truth is a meaningful and useful concept, we undermine that meaning and usefulness by treating it nihilistically.
A priori appeals to truth
However, it might well be asked what exactly constitutes an appeal to truth, and where appeals to truth end. If, as I have argued, claims to know the truth are disallowed by sceptical arguments, but the concept of truth should be defended as meaningful even when absolute, what other kinds of appeals to truth, if any, can be justified?
The answer is that no appeals to truth can be justified, when an appeal to truth is understood as relying on either claimed or hypothetical truth as the basis of further claims about meaning or knowledge. The reason for this is the reliance that all these claims place on dogmatic assumptions about either what truth is like or what role it plays. The assertion that truth is meaningful, however, depends not on a concept of truth itself for its justification, but on a Lakoffian pragmatic understanding of meaning. To assert that truth is meaningful only requires an appeal to our experience of it being so, but to appeal to it in any other way involves a circular appeal to the nature of truth itself.
A common philosophical appeal to truth is made a priori in the rationalist tradition from Plato to Kant and beyond. Earlier Rationalist appeals to the a priori involve the dogmatic appeal to essences or to essentially infallible methods of understanding. However, in Kant’s more compelling formulation, a priori claims to truth are those required for any possible experience. Kant’s approach was qualitatively different from those of Plato and Descartes, because he did not appeal to an absolute foundation of truth to justify further beliefs, so much as deduce what conditions were required for the experiences and judgements we encounter. Kantian arguments thus do not involve an appeal to truth from the beginning (though we may find other questionable assumptions in them). Kant’s beliefs are simply justified through a priori reasoning, and subject to further investigation or criticism by the same method, but just because they have been established a priori does not mean that we have to regard them as absolute truths.
This can also be shown through the fact that Kant’s a priori theories are still subject to many of the sceptical arguments listed at the beginning of this chapter. Just because a certain framework of understanding provides the only possible method for us to access the universe, this does not prove that such a framework shows us how the universe actually is. It also does not show that our framework is not conditioned by our linguistic and cultural background. Intelligent aliens, if there are any, might have quite different sorts of frameworks, even different frameworks for time and space through which we most basically structure our world. Their language and culture, if these are appropriate terms to apply at all to aliens, might also be unimaginably different from our own.
The same arguments can be used for the status of mathematical and geometrical ‘truths’. Mathematical and geometrical theories are proven a priori because they relate to the way in which we organise our experience, using numbers, lines, and points in theoretical space. However, their being proven within the framework of mathematical reasoning does not show them to essentially have the status of absolute truth-claims, for they are our constructions through which we interpret the world. Similarly with scientific claims that appear to have the status of universal laws, such as
Logic and hypothetical appeals to truth
Slightly different problems are created by the Aristotelian logic basic to Western philosophy, and its hypothetical appeal to truth. I have already criticised truth-conditional theories of meaning as inadequate to the way in which we actually encounter significance in experience, yet the appeal to truth conditions appears to be central to logical deduction, the process of claiming that if one claim is true, then another one must be true.
This process requires the use of truth-conditions and an assumption of representationalism, as a logical deduction involves the line of thought that the truth-conditions of one claim are equivalent to those of another. Thus if x implies y, this means that x represents a state of affairs that, if true, also be represented by y. For example, if Fred is a bachelor, this proposition represents a state of affairs that, if true, could also be represented as the claim that Fred is an unmarried man.
Logical deduction is very widely used, including very frequently in this book, so I am not about to attempt to undermine the validity of all logical deduction. Nevertheless, our understanding of what is involved in logical deduction must change if we are to avoid being dependent on representationalism. What representationalism misses out in its reduction of the meaning of a term to its truth-conditions is the emotional significance of that term as equally part of its meaning to a flesh-and-blood person in a particular context. So, to revise our understanding of logical deduction to take this into account we would have to say that x implies y if x has a significance, both cognitive and emotional, equivalent to y for those who produce and understand the logical deduction. This means that logical equivalence can no longer be assumed to be an absolute and universal matter, and though the rules of logic that Aristotle and his successors have devised may apply to a large extent within a zone of shared meaning-assumptions, they do not apply in all cases
In practice, however, where the expected audience for a logical deduction is predictably to be found amongst a certain class of people with shared expectations, and the participants’ experience of the meaning of a statement is overwhelmingly cognitive (as in most books of philosophy) this account need not make a huge difference to our understanding of the validity of logical deductions. However, it does leave them open to challenge when the emotional aspects of meaning do have significant practical importance and have not been taken into account in the implications that are claimed.
This approach does make an appreciable difference where the subject-matter of logical deductions is beyond people’s experience entirely, as is the case in some branches of analytic philosophy that deal only in remote hypothetical truths. The further those possibilities get from experience, the less confident we can be of the claimed equivalences being valid, as there is no clear way to check the meaning of either the premises or the conclusions in relation to experience.
This mode of deduction with the qualification of its dependence on shared meaning is one that I use throughout this book, and should be taken to be the nature of all arguments expressed in deductive terms, even when this is not explicitly stated. To avoid confusion with deduction as used with truth-conditional assumptions, I call this pragmatic deduction.
Another reason often given for continuing to appeal to truth is the idea that justification depends on truth, so that in order to justify a belief one is in effect appealing to truth. Of course, this is one possible, and indeed common, way of justifying a belief, where a truth is assumed from which one could deduce other beliefs. Using the metaphor of building further certainties on firm foundations, this view is known in epistemology as foundationalism. However, foundationalism has already been found unacceptable due to its reliance on truth-claims, and it does not provide the only possible way of justifying beliefs.
The other recognised types of theory in epistemology are coherentism, in which beliefs are justified because of their coherence with other beliefs we already hold, and reliabilism, in which justifiable beliefs have been gained by a reliable mechanism. Neither of these alternatives is a sufficient account of justification by itself, however. First let us deal with reliabilism. We can only judge a mechanism for checking our beliefs as reliable if we have some other way of checking it, if we want to avoid the circularity of justifying a reliable method by the same reliable method. The analytic version of reliabilism in which we do not need to know that the mechanism we are using is reliable can be dismissed immediately, because it provides no way of helping us justify our beliefs and is thus of no practical relevance. But reliabilism in any other terms collapses into another form of foundationalism, in which appeal is made to a particular basis of truth, which happens to consist in a method of finding the truth.
Coherentism, on the other hand, is not sufficient for justification because it is possible to have a coherent set of beliefs that are totally deluded. Some mental illnesses illustrate this within the realm of ordinary experience. We can add to this the criticism that coherentism is only appealed to because it is the only remaining alternative when foundationalism has been ruled out, so it amounts to an anti-foundationalism. However, the dubiousness of claims to truth does not support an equally dubious rejection of claims to truth.
We are thus driven to the conclusion that justification depends, not just on foundations (whether of absolutely certain beliefs or of absolutely certain methods) or on coherence, but on both. To be justified, a belief must be coherent with other beliefs we hold, and it must also be consistent with a foundational assumption – but how can we have a necessary foundational assumption after rejecting truth claims?
The answer to this lies in what I call negative foundationalism. Negative foundationalism is the belief that, to be justified, beliefs must be consistent with the recognition that no certainty about the truth can ever be reached by finite beings. The only certainty we can rely on is an absence of certainty, whether negative or affirmative. Negative foundationalism is a necessary counterweight to coherentism, which can easily slip into a dogmatic denial of foundations and a reliance only on what is already conventionally accepted.
Negative foundationalism is far-reaching once its implications are explored, as they will be more fully in the next chapter. For example, the belief that God exists and the belief that God does not exist are both justifiable in terms of coherence with human experience, since all our experiences can be interpreted in a way consistent either with God existing or not existing. However, negative foundationalism requires us to reject both these beliefs, because they are both incapable of justification through experience.
However, to pursue the implications of negative foundationalism too far here will take me into the themes of the next chapter, which is concerned with exactly how we justify our beliefs and on exactly what grounds we can identify and avoid absolute truth-claims. Both of these points need considerably more explanation to make a full and convincing case for negative foundationalism.
For the moment, however, enough should have been said to establish justification as an alternative to truth that is not dependent on claims to truth. Instead, I want to argue, justification of our beliefs is dependent on (a) coherence with our other beliefs and (b) consistency with a recognition that we have no direct access to truth. Both of these conditions are necessary for justification, and a reliance on one alone is not adequate for justification.
It also needs to be stated here that the kind of justification I am talking about is objective justification. This is not just a piece of analytic philosophy that confines itself to analysing what we already assume by ‘justification’; it is considering the grounds on which we could accept new beliefs. The next two chapters should make it clearer in what sense I want to argue that justification can be objective without appealing to truth-claims.
From truth to justification
Abandoning claims to truth, but nevertheless retaining justification, has immense implications, that much of the rest of this book will be exploring. Here are some of the most crucial implications:
 For the ancient Greek sceptical arguments see Sextus Empiricus, trans. Benson Mates (1996) The Skeptic Way: Sextus Empiricus’s Outlines of Pyrrhonism Oxford University Press,
 See G.E. Moore ‘Proof of an external world’ from G.E.Moore: Selected Writings ed. T. Baldwin, Routledge 1993.
 Some philosophers also consider the assumptions of scepticism to have been undermined by Wittgenstein’s private language argument. For a refutation of Wittgenstein’s argument please see A Buddhist theory of moral objectivity 4.e.iii: Wittgenstein relies on dogmatic assertions of what is “meaningful” or “meaningless”. Scepticism in any case does not require the solipsistic assumption of an isolated self that Wittgenstein wants to attack, only a recognition of the lack of ultimate justification for our assumptions, howsoever formed.
 For a more detailed summary of Lakoff’s theory of meaning, see A Buddhist theory of moral objectivity 2.c.iii. For full details, of course, see Lakoff’s own works, starting with Women, fire and dangerous things (University of Chicago Press, Chicago 1987)