Susan Haack

I was recently invited by the editor of a literary journal to contribute to an issue on coherence – in epistemology (theory of knowledge), philosophy of science, chaos theory, the human sciences, political theory, economics, cultural studies, art, music, literature, rhetoric, composition teaching, and/or “advanced scholarly argumentation”. “Is the concept of coherence value-neutral or value-laden?” the letter continued. “Can one think of coherences in the plural, or is there something in the logic of the concept that overrides differences and assimilates its instances to itself?”

My first reaction was that, like the famous peppermint burgundy of Monty Python's “Wines of Australia” sketch, such questions might best be “laid down – and left.” But on reflection I was moved to suggest a less dismissive answer: “coherent” (like “rational,” “reasonable,” etc.) has many meanings and many objects, and in some uses, including the logical and the epistemological, it is a term of favourable appraisal; nevertheless, while coherence has sometimes been under-valued, it has also, sometimes, been over-valued. Most to the present purpose, coherence is only one element of rationality in belief or action.

Besides being used to describe the texture of batter or cement, “consistency” denotes one dimension of logical appraisal. A set of propositions is logically inconsistent just in case a contradiction – the conjunction of a formula and its negation, “p and not-p” – can be derived from it; otherwise, it is logically consistent. Outside formal-logical contexts, “inconsistent” is often used in a broader sense in which it connotes the mutual incompatibility of a set of propositions; “Tom is a bachelor” and “Tom is married,” for example, though not formally inconsistent, are mutually incompatible.

Of course, mutually incompatible propositions can no more be jointly true than formally inconsistent formulae can; though there are many different truths, there can't be incompatible “truths” or “knowledges”. Incompatible propositions can be accepted as true, but they can't all be true; and, though sometimes we say that something is “true for you but not for me,” this is just a misleadingly elliptical way of saying that you believe whatever-it-is, but I don't, or else that whatever-it-is is true of you, but not of me. Inconsistencies are often an important spur to intellectual effort (as Russell's Paradox was to the various developments of set-theory after Frege); but this is precisely because the presence of an inconsistency in a theory means that it must be false at least in part.

Since mutual compatibility is a necessary condition for the truth of a set of propositions, some philosophers have hoped that the addition of further requirements such as “comprehensiveness” or “mutual entailment” would yield a concept of coherence that could serve as a necessary and sufficient condition of truth. F H Bradley wrote that “truth is an ideal expression of the Universe, at once coherent and comprehensive”; and coherence theories of truth were defended not only by other idealists and sympathisers, but also, among the logical positivists, by Otto Neurath. Brand Blanshard took coherence to constitute, not just the criterion, but the definition of truth; usually, however, interest in coherence was more epistemological than metaphysical. And it is in epistemology – where coherentism, with its traditional rival, foundationalism, is one of the standard theories of justified or reasonable belief, of good or adequate grounds, reasons, or evidence – that coherence concepts have of late played their most prominent philosophical role.

Taking empirical justification to require anchoring in the world, foundationalism explains justified empirical belief as belief either directly justified by the subject's experience, or supported by beliefs which are so justified, and so on. Taking justification to be a relation among beliefs, coherentism explains justified empirical belief, instead, as belief that coheres (in some epistemological, consistency-plus, sense) with the subject's other beliefs. Some foundationalists, however, allow that basic beliefs are “defeasible” (in principle revisable), and/or that there is mutual support among derived beliefs; and some coherentists allow that perceptual beliefs generally have a distinguished status. And as the traditionally rival theories have leaned towards each other, the leaning has destabilised them: the foundationalist assumption of a robust distinction of basic vs. derived beliefs, and the coherentist assumption that justification is a matter of relations among beliefs exclusively, have come to seem less and less well-motivated.

The “foundherentist” theory I proposed in Evidence and Inquiry requires neither of these assumptions. Like foundationalism and unlike coherentism, foundherentism acknowledges the relevance of experience to the justification of empirical beliefs; but, like coherentism and unlike foundationalism, it posits no privileged class of basic beliefs justified by experience, supporting but not supported by other beliefs. Experiential evidence and reasons work together, as clues and already-completed crossword entries do. How reasonable a crossword entry is depends on how well it is supported by its clue and any completed intersecting entries, how reasonable those other entries are, independent of this one, and how much of the puzzle has been completed; analogously, how justified a belief is depends on how well it is supported by experiential evidence and reasons, how secure those reasons are, independent of the belief in question, and how comprehensive the evidence is. Supportiveness is tied to explanatory integration, which is close kin to the concept of explanatory coherence on which some coherentists rely; but in foundherentism supportiveness is only one of the multidimensional determinants of evidential quality.

There is legitimate mutual support among beliefs, but no vicious circle of reasons; the requirement of independent security prevents it, here as in a crossword. Nor is there an infinite regress of reasons; experiential evidence – which consists not of beliefs but of perceptual events, and so stands in no need of justification – serves as anchor, as clues do in a crossword. For, as I argue in Defending Science – Within Reason, though perceptual events, not being propositional (meaning that they are not statements in language), cannot stand in logical relations to propositions believed, they nevertheless bear on the justification of beliefs because of the connections set up as language is learned: connections which, however, are fallible and revisable; as the child discovers when he learns “looks like a dog,” “appears to be striped,” and such.

The foundherentist resolution of the impasse of foundationalism and coherentism is less radical and more plausible than the contextualist alternative, which reconstrues justification as consonance with the epistemic practices of the subject's community or, in Richard Rorty's tribalist version, with the epistemic practices of our community. Though he sometimes describes himself, very misleadingly, as a coherentist, the concepts to which Rorty actually gives the starring epistemological role are consensus and social cohesiveness. “The only thing exemplary about science is that it is a model of human solidarity,” he writes, apparently assimilating the achievements of four centuries of scientific investigation to – what? a really good trade union? Science is a deeply social enterprise, relying on the work, cooperative and competitive, of generations of inquirers, and their pooled evidential resources. But that scientists in the relevant field accept it doesn't warrant a theory; killing off dissenters, or playing a tape repeating “the earth moves” under the pillows of holdouts while they sleep, won't make the claim in question any more likely true. The epistemological significance of consensus in the scientific community is rather that, on the whole and in the long run, often enough, the good evidence that warrants the theory also explains scientists' agreement.

Whether construed in the usual, relativist style, or in Rorty's tribalist fashion, contextualism really is a desperate measure; abandoning the idea of objectively better or worse evidence, it would, among other things, knock away the epistemological underpinnings of the entire legal system. Happily, no such desperate measures are necessary.

We appraise not only people's beliefs, but also their thinking, speech, writing, and actions for coherence. What counts as thinking coherently depends on the context: a physician checks whether a patient in shock knows his own name, what day it is, who is currently president, and so on; but the academic we describe as incoherent can pass that test – the complaint is, rather, that his thinking is muddled, fuzzy, scrambled, perhaps contradictory. It's normal for one's first thoughts about a difficult question to be inchoate, and to shift up and back between one conclusion and its opposite; but sometimes, rather than working through this frustrating initial stage, a person seizes on a confused or half-baked idea and relies on it in blithe disregard of its inability to take the weight. And then, as Peirce observed, the consequences can be disastrous: “It is terrible to see how a single unclear idea, a single formula without meaning, lurking in a young man's head, will sometimes act like an obstruction of inert matter in an artery, hindering the nutrition of the brain, and condemning its victim to pine away in the fullness of his intellectual vigor and in the midst of intellectual plenty.”

Sets of propositions can be inconsistent; situations or states of affairs can't, but they can be chaotic or confused. And, as there can be a sober description of a drunken man, there can be a consistent description of an inconsistent set of propositions; of Frege's inconsistent logic, for example. But often, when we speak of the coherence of a person's speech or writing, we have in mind, not its logical consistency, but something more pragmatic: we praise a colleague‘s or student's paper or presentation for its cogency, or complain that it is lazy, muddy, jumbled; we describe the speech of someone drunk, drugged, or mentally disturbed, or of an academic undone by too much Theory, as “incoherent” – rambling, garbled, a glossogonous word-salad, high-toned gobbledygook.

“Incoherent” is sometimes also used, again of someone's speech or writing, in the specialised sense of “self-undermining”: for example, when someone claims that truth is relative to culture, or that there are no beliefs – though neither the proposition that truth is relative to culture nor the proposition that there are no beliefs contains any hidden contradiction – his asserting this undermines what he asserts; for one who sincerely makes a categorical assertion expresses his belief, and makes a non-relative claim to truth.

When it is actions that are being appraised, “consistent” means something like “behaving in the same way in similar circumstances”; and so applies to a person's (or institution's) practice over time. Whether such consistency is desirable depends on whether you are consistently following ill-considered ways of behaving simply because this is how you behaved in the past, or consistently following a well-considered policy of action – precisely the point of that famous observation of Ralph Waldo Emerson's, that “a foolish consistency is the hobgoblin of little minds, adored by little statesmen and philosophers and divines.” Emerson rightly scorns the mental rigidity of one who insists on doing things as he has always done them: “He may as well conform himself with his shadow on the wall.” But the distinction between a wise consistency and a foolish, often forgotten when Emerson's observation is quoted only in part, is essential.

Since fairness demands that people similarly situated be similarly dealt with, consistency in practice is sometimes a legitimate, even an essential, concern – felt by the conscientious teacher grading papers, or the jurist looking to precedent. But here too its desirability depends on the wisdom of the practice being consistently applied: it would be worse than foolish to continue adding up the marks wrong so as to maintain consistency with the mistake I made the first time; and, though we want the law to provide stability and predictability, we don't want it to be stagnant, unresponsive to social change.

Cognitive psychologists have their own word for a kind of incoherence that especially interests them, tensions between a person's beliefs and his actions and preferences: “cognitive dissonance”, the concept explored in Leon Festinger's theoretical writings and in his studies of millennial sects which react to the failure of their predictions of the End of the World, not by acknowledging that they were mistaken, but by reinterpreting their prophecy and proselytising more energetically; and the key concept in Alison Lurie's wickedly funny fictional variations on these real quirks of human nature in her Imaginary Friends .

For our appraisals of coherence and incoherence also extend to literary texts and fictional characters. I once heard Peter Geach argue, against the analogy of possible worlds to novels, that there can be inconsistencies in novels – for example, geographical inconsistencies in War and Peace . He meant “inconsistent”, obviously, in the quasi-logical sense. But in literary contexts the point is more often, not the consistency or otherwise of the chronological, geographical, forensic, etc., details of a fictional work, but the consistent or inconsistent behaviour of its characters, the congruence or incongruity of its themes, or the unity or disunity of its mode of presentation or language.

In The Way of All Flesh Samuel Butler writes of Ernest Pontifex (whose zigzag path to maturity the book chronicles) that “when he rose he flew like a snipe, darting several times in various directions before he settled down to a steady, straight flight,” but that “once he had got into this he would keep to it”; sending me first to my bird-book, and then back to Aristotle's observation that characters should be consistent, but “if the model for the representation is somebody inconsistent, and such a character is intended, even so it should be consistently inconsistent.” Here “consistent” is used in the same sense in which it applies to the behaviour of real people, now extended to fictional characters.

In Daniel Deronda , there is a satisfying congruity of intertwined narratives, unified by George Eliot's theme of the power of ignorance: Deronda, originally unaware of his origins, discovers that he is Jewish, and explores what that means to him; Gwendolen Harleth, too self-confidently ignorant to realise how ignorant she really is, makes a disastrous marriage when her family faces financial ruin. Here “congruous” means something like “illustrating the same theme”. In a work of literature, however, not only such congruities, but also, sometimes, the well-chosen incongruity may be pleasing. We enjoy contrasting intertwined plots, skilfully sliced and spliced narratives, even agreeably shocking narrative disruptions. And we find well-chosen grammatical incongruities a rich source of verbal humour, and sometimes, wonderfully effective literary devices, such as the opening line of chapter five of Ursula Le Guin's The Left Hand of Darkness : “My landlady was a voluble man”; the startling verbal incongruity is exactly right, given her cast of hermaphrodite, but otherwise very human, characters.

Sorting out the coherence family of concepts – “consistent”, “cogent”, “congruent”, “cohesive”, etc. – has been challenging enough; a thorough-going scrutiny of the umbrella concepts of rationality and reason would require, just for starters, an exploration of such contrasts as the rational versus the arational, the rational versus the irrational, reason versus experience, reason versus emotion. Another time, perhaps!

This is a modified and abridged version of a paper entitled “Coherence, Consistency, Cogency, Congruity, Cohesiveness, &c.: Remain Calm! Don't Go Overboard!”, forthcoming in New Literary History.

Susan Haack is Cooper Senior Scholar in Arts and Sciences, Professor of Philosophy, and Professor of Law at the University of Miami. Her most recent book is Defending Science — Within Reason: Between Scientism and Cynicism (Prometheus, 2003).