1. The poverty of philosophy as a science

Throughout its history philosophy has been thought to be a member of a community of intellectual disciplines united by their common pursuit of knowledge. It has sometimes been thought to be the queen of the sciences, at other times merely their under-labourer. But irrespective of its social status, it was held to be a participant in the quest for knowledge – a cognitive discipline.

Cognitive disciplines may be a priori or empirical. The distinction between what is a priori and what is empirical is epistemological. It turns, as Frege noted, on the ultimate justification for holding something to be true.Footnote ¹ If the truths which a cognitive discipline attains are warranted neither by observation nor by experiment (nor by inference therefrom), then they are a priori. Otherwise they are empirical. The natural and moral sciences (the Geisteswissenschaften) strive for and attain empirical knowledge.Footnote ² The mathematical sciences are a priori.

Cognitive disciplines have a distinctive subject matter, concerning which they aim to add to human knowledge. Physics deals with matter, motion, and energy, chemistry with the constitution of stuffs out of elements, biology with the nature of living beings, history with ‘the crimes, follies and misfortunes of mankind’ (Gibbon), and so forth.

The empirical sciences aim not only to discover truths but also to explain the phenomena they study. The natural sciences produce theories (typically with predictive powers) to explain the facts and laws they discover. The moral sciences too aim to explain the phenomena they study – although not to the same extent by way of theory and general laws; and their predictive powers, if any, are more limited. Mathematics and logic strive to produce theorems by means of proofs, and are not subject to confirmation or falsification by experience.

If philosophy is a cognitive discipline, then the truths it attains need to be characterized. Are they a priori or empirical? To answer this question, we should cast around for established philosophical truths – examine the fund of philosophical knowledge achieved over two and half thousand years. But two disturbing features immediately spring to the eye:

First, if one asks a physicist or biologist, a historian or a mathematician what knowledge has been achieved in his subject, he can take one to a large library, and point out myriad books which detail the cognitive achievements of his subject. But if one asks a philosopher for even a single book that will summarize the elements of philosophical knowledge – as one might ask a chemist for a handbook of chemistry – he will have nothing to present. There is no general, agreed body of philosophical knowledge – although there are libraries full of philosophical writings from antiquity to the present day, which are in constant use.

Secondly, each cognitive discipline has its own object of study. But if we examine the history of modern philosophy, it appears to be a subject in search of a subject matter. In the modern era, great philosophers recurrently attempted to isolate a distinctive subject matter for philosophy, and a proper method for achieving the knowledge, which, they held, had evaded their predecessors. Descartes thought that the task of ‘first philosophy’ was to disclose the foundations of all human knowledge, and to erect a certain and secure structure of knowledge on indubitable truths. The key to achieving this was his new method. Only thus could philosophy participate in the quest for knowledge. Hume supposed that the subject matter of philosophy was the human mind, and the task of philosophy to explain how it functions. Philosophy must do for psychology what Newton had done for physics, and must introduce the experimental method of reasoning into the study of the mind. Kant held that philosophy must determine the a priori categories of thought and the a priori principles of conceptualized experience. It must, above all, explain how synthetic a priori knowledge is possible. Only then will metaphysics be set upon the true path of a science. Russell thought that the subject matter of philosophy consists of the most general truths about the universe, and its task to discover them and catalogue their logical forms. Only by adopting the scientific method in philosophy can genuine progress be achieved. And so on.

The striking feature of these programmatic objectives is that none survived for long. Each collapsed, for one or another of four reasons:

1. (i) A vital assumption proved unsustainable. So, for example, it was a Cartesian error to suppose that genuine knowledge must be indubitable; or resistant to hyperbolic doubt. It was erroneous to suppose, as Russell did, that propositions of logic are all generalizations that describe the most general features of the universe. On the contrary, propositions of logic need not be general (e.g. ‘Either it is raining or it is not raining’ is, contra Russell, a proposition of logic), and they describe nothing at all (e.g. the latter tautology tells one nothing about the weather).

2. (ii) The subject matter, correctly understood, was taken over by an empirical science. So, for example, the experimental study of the exercise of human cognitive faculties that Hume allocated to philosophy was taken over by experimental psychology.

3. (iii) The goal proved to be chimerical: the conception of knowledge as resting on indubitable foundations is wrong. Hence the goal of displaying the structure of human knowledge as a hierarchy based on subjective experience is illusory. The Kantian goal of explaining how synthetic a priori truths are possible foundered over the misconception of such truths as propositions to which nature must conform, rather than as expressions of norms of representation.

4. (iv) The method proved broken-backed: Cartesian method is not a reliable way of discovering truths, Kant's Copernican revolution is misconceived, and Russell's scientific method in philosophy is a chimera.

This should give us pause. How can it be that after two and a half thousand years of endeavour philosophy has still not reached the status of a science, has no agreed subject matter, and has no fund of philosophical knowledge? How is the poverty of philosophy, construed as a cognitive discipline, to be explained?

2. Philosophy as the midwife of the sciences

Many questions that were opened by philosophers were subsequently handed over to scientists, for example questions concerning the constitution of things, the infinity or finitude of the universe, the nature of the stars, the origin of life, the innateness of ideas. Physics, although it continued to be known as natural philosophy down to the nineteenth century, became independent of philosophy in the seventeenth. Psychology broke free of philosophy at the end of the nineteenth century, and mathematical logic is doing so today. This midwifery has been invoked (by Russell and Austin, for example) to explain the poverty of the results of philosophy – namely that once questions are sufficiently sharply formulated to be answerable, they are handed over to an independent science, which then contributes to the extension of human knowledge.Footnote ³

This is misleading, for four reasons:

First, although independence was achieved by such sciences, new areas of philosophical investigation were thereby generated, e.g. philosophy of physics or philosophy of the psychological sciences. But it would be misguided to suppose that questions in the philosophies of the special sciences remain philosophical only because they are insufficiently clearly understood to be handled by a new meta-science.

Secondly, although these sciences achieved independence, it would be mistaken to suppose that they achieved freedom from conceptual confusion. The conceptual confusions of the sciences, in physics, psychology, neuroscience, economics (not always recognized as such by scientists) are grist for philosophical mills – not philosophical problems for experimental investigation. (Of course, scientists may grind the grist too – we are not concerned with trade union disputes, but with distinguishing different forms of intellectual enquiry.)

Thirdly, the birth of an independent science does not free philosophy from a host of questions which have always been on the philosophical agenda associated with the subject matter of that special science. Despite the fact that investigations of matter in motion had achieved a degree of clarity that made it possible for them to be handled by an independent science of physics, such questions as: What distinguishes substances from properties? How are substances related to events and which is ontologically prior? – such problems were not allocated to physics. Similarly, even when questions about what material things are made of, what the ultimate chemical elements are and what kinds of chemical combinations they enter into, were sufficiently clearly understood to be handed over to chemistry, other questions, such as how things (substances) are related to the stuff of which they consist, remained exactly where they had always been – on the agenda of philosophy. And the autonomy of psychology has not removed from the domain of philosophy the fundamental questions in philosophical psychology, such as ‘What is the mind?’ or ‘How is the mind related to the body?’

Fourthly, the suggested explanation is implausible when we turn to practical philosophy (in Kant's sense of the term) – to ethics, political and legal philosophy. Moral philosophy has not and will not give birth to a science of morality, and so called ethicists are not moral scientists. The emergence of political science in the nineteenth century was not a result of philosophical midwifery, and legal philosophy is not going to be displaced by a science of law. Moral, legal and political philosophy do not give birth to new sciences, but contribute to the emergence of new moral, legal and political distinctions, principles and constitutional arrangements.

So, the poverty of philosophy qua cognitive discipline cannot be explained as a consequence of the fact that once knowledge is achievable the subject becomes a science.

3. ‘Philosophy has only just come of age’

There is another move here, that might, in honour of its recent advocates, be called the Wykeham Chair gambit. Thirty years ago, Professor Michael Dummett, Wykeham Professor of Logic at the University of Oxford declared that ‘philosophy has only just very recently struggled out of its early stage into maturity: the turning point was the work of Frege, but the widespread realization of the significance of that work has had to wait for half a century after his death…’Footnote ⁴ Recently, Professor Timothy Williamson, Dummett's successor but one in the Wykeham Chair of Logic at the University of Oxford, declared that we have only now (in 2005) arrived at ‘the end of the beginning’ of philosophy.Footnote ⁵ Well, one can blow the Last Trumpet once, but not once a generation.

Less parochially – the suggestion that philosophy has not achieved the results of a science because the subject is so difficult that only NOW has it been discovered how it may do so, has been advanced by numerous great philosophers who were not holders of the Wykeham Chair of Logic at the University of Oxford. They all enjoyed the brief illusion that they had, at long last, found the real key to unlock the riches promised by philosophy, to achieve real philosophical knowledge and to set philosophy at last upon the true path of a science. Descartes thought that his new method of analysis and systematic doubt would enable anyone to establish the indubitable foundations of knowledge, and to derive all possible knowledge in absolutely sure and certain steps. Locke thought that with his new Way of Ideas, he would be able to determine for the first time the scope and limits of human knowledge. Hume proposed ‘a compleat system of the sciences, built on a foundation almost entirely new, and the only one upon which they can stand with any security’. Kant, comparing the method of his critical philosophy to the Copernican revolution, supposed that by following the principle that ‘objects must conform to our [a priori] knowledge’, rather than our a priori knowledge conforming to objects, he would at last be able explain how synthetic a priori knowledge is possible, and to place metaphysics ‘upon the true path of a science’. Russell too recognised the scandal that ‘Philosophy, from the earliest times, has made greater claims, and achieved fewer results, than any other branch of learning.’ He boldly declared ‘that the time has now arrived when this unsatisfactory state of affairs can be brought to an end.’

The promise that after two thousand years of irresponsible adolescence, philosophy will at last produce a flood of truths and well-founded theories – tomorrow, has been made, and proven empty, far too often to carry conviction. Moreover, such declarations of the incompetence of one's predecessors does scant justice to the endeavours of some of the greatest geniuses of mankind. And it renders it well nigh unintelligible that we still read, and should still read, the works of Plato and Aristotle, Aquinas and Scotus, Descartes and Kant. (Scientists do not need to read the works of Galen or Paracelsus, Tycho Brahe or Kepler.) Finally, it is implausible to suppose that twenty-five centuries of endeavour by some of the greatest minds of our culture should have failed to come up with some solid philosophical knowledge because the problems of philosophy are so much more difficult than problems in the sciences. Is the philosophical problem of what a substance is, and how substances are related to the stuffs of which they are made so much more difficult than the question of what are the elements of which all things are made? Is the philosophical problem of what knowledge is so much more difficult than the question of the descent of man? Is the relation of mind to body so much more complex than the Krebs cycle? Is that why we can discern so little achievement in this ‘sector in the quest for knowledge’ (as Dummett once put it)? Surely the difficulty of philosophical questions is not to be compared to that of scientific questions in degree, but in kind.

When bombarded throughout the ages with incompatible claims about the subject and unfulfilled promises of how this is going to be set right, the correct move is to challenge the fundamental assumption that is taken for granted by all participants in the debate, namely the assumption that philosophy is a cognitive discipline.Footnote ⁶

4. Philosophy as a quest for understanding rather than knowledge

Philosophy is not a contribution to human knowledge, but to human understanding. It is neither an empirical science nor an a priori one, since it is no science. The difficulty of philosophy does not consist in the difficulty of discovering new, let alone arcane, truths about the world; nor yet in producing proofs concerning its existenceFootnote ⁷, the existence of recherché ‘entities’ like universalsFootnote ⁸, or of common or garden ‘entities’ like events.Footnote ⁹It is a quest for understanding, not for knowledge.

As a slogan, this is correct. Like all slogans, it needs clarification and qualification. First, some clarification:

If one claims that philosophy is a contribution to human understanding, one must explain what the object of understanding is, and how achieving understanding in philosophy differs from adding to one's fund of knowledge.

It has been suggested that philosophy seeks not knowledge of new facts but an understanding of old facts; or that its role is that of organizing the knowledge we already possess. These suggestions are partly right and partly wrong.

Scientists seek to understand why the phenomena they investigate are as they are and behave as they behave. They do so by way of empirical explanation, which may take various forms, e.g. hypothetico-deductive, inference to the best explanation, or explanation by reference to intervening mechanisms. All these are subject to empirical confirmation or refutation. To that extent it is misleading to suggest that philosophy seeks not for knowledge of new facts but for an understanding of familiar facts – as if science did not satisfy that need. Philosophy cannot explain phenomena in that sense at all. So whatever its quest for understanding is, it is not akin to the understanding achieved by the empirical sciences.

Nevertheless, philosophy can contribute in a unique and distinctive way to understanding in the natural sciences and mathematics. It can clarify their conceptual features, and restrain their tendency to transgress the bounds of sense. It is a Tribunal of Reason, before which scientists and mathematicians may be arraigned for their transgressions. Indeed, the sciences (and to a lesser degree mathematics), in our times, are the primary source of misguided metaphysics – which it is the task of philosophy to curb, not to encourage. Disabusing a Hilbert of the character of Cantor's paradise contributes to the deeper understanding of arithmetic in general and of the calculus of transfinite arithmetic in particular.Footnote ¹⁰ Explaining that alternative geometries are not alternative theories of space but alternative grammars for the description of spatial relationships contributes to a better understanding of the enterprise of geometry. Making it clear that parts of the brain are not possible subjects of cognitive predicates contributes to a better understanding of the manner in which neuroscience can explain the neural foundations of our cognitive powers. Nevertheless, it would be misleading to suggest in general that the understanding that philosophy seeks is parasitical on the sciences in this way. For the illumination philosophy can thus contribute characterizes primarily the philosophies of the special sciences.

Similarly, there is some truth to the claim that philosophy does not add to the sum of our knowledge of the world (or of mathematics), but rather organizes what we already know. Certainly distinctions that philosophers have progressively drawn since the days of Aristotle have contributed to clarity regarding the sciences. It is thanks to philosophy that we distinguish the empirical sciences from logic and mathematics, the natural from the moral sciences, deductive from inductive reasoning, a priori from empirical probability, nomothetic from idiographic explanation, causal from hermeneutic explanation, and so on. These distinctions are crucial for a proper understanding of the manifold scientific (as well as non-scientific) enterprises of trying to gain knowledge and understanding of the world we live in, of ourselves within it, and of the mathematical apparatus we have invented to quantify it. Nevertheless, it would be misleading to characterize the task of philosophy as being to organize, or to put order into the knowledge we already possess. Insofar as philosophy has to organize material, what it has, above all, to organize are forms of description (or norms of representation) by means of which we present what we know and what we strive to know. I shall elaborate this below.

The kind of understanding philosophy pursues is distinctive. It can be described in various more or less misleading ways:

In the metaphysical mode: philosophy strives for an understanding of the a priori natures of things and of internal relations between things (but there are no ‘metaphysical facts’ to be discovered, and internal relations are creatures of reason, not of nature).

In the conceptual mode: philosophy strives for an overview of the structure of (parts of) our conceptual scheme and of logico-grammatical relations between its elements (but that does not make concepts the special subject matter of philosophy).

In the linguistic mode: philosophy strives for an overview of segments of our language that in one way or another, give rise to conceptual problems (but philosophy is not in general about language).

Correctly understood, these are descriptions of one and the same enterprise. Of course, investigating the use of a word need not be a logico-grammatical investigation into the concept it expresses. It may be a non-philosophical, purely linguistic, investigation into etymology, phonetics, syntax, morphology, and so forth. But a philosophical investigation into the use of a word is an investigation into the concept expressed, for it is an investigation, geared to philosophical purposes, into the presuppositions, implications, compatibilities and incompatibilities linked with the use of the word in sentences. For the most part, philosophers will abstract from irrelevant local differences between languages. A philosophical investigation into the use of ‘know’, for purposes of epistemology for example, will yield much the same results as a philosophical investigation into the uses of ‘wissen’ and ‘kennen’, the manifest differences often being irrelevant to the investigation.Footnote ¹¹ For the investigation, whether conducted in English or in German, is an investigation into those features of usage that determine the common concept of knowledge.Footnote ¹²

The a priori nature of things is fixed by the sense-determining rules for the use of expressions signifying things.Footnote ¹³ To suppose that things, their properties and relations have an a priori nature in any other sense is to fall victim to illusion. For it is to take for ‘objective (language independent) necessities’ what are actually no more than the shadows cast on the world by grammar. To describe the nature of substance, for example, is to characterize the categorial concept of substance, just as describing the nature of events is to characterize the concept of an event and to describe the nature of the mind is to characterize the concept of mind. But there is no way to characterize a concept other than by describing the relevant features of the uses of expressions that express that concept or belong to the category of concepts it subsumes. So to describe the nature of substance just is to spell out, and order, the salient sense-determining rules for the use of that subclass of concrete count nouns that signify substances, and their similarities to and differences from other kinds of nouns. This may be done in the formal mode or (more commonly) in the material mode. To state the nature of events just is to describe (directly or indirectly) the constitutive features of event-designating expressions, and to compare and contrast them with the use of other general types of expression, such as material object names. And to describe the nature of the mind is to describe and order the relevant features of the use of the expression ‘the mind’ and its cognates, and of psychological predicates ascribable only to creatures that can be said to have minds.

Philosophy has no subject matter in the sense in which the empirical sciences do. It deals with philosophical questions, which are different in kind from questions in the empirical sciences and in mathematics. What philosophical questions are is best displayed by an array of uncontroversial and incontrovertible examples. These will be very various: ‘What is …’-questions (e.g. what is the mind, knowledge, truth); ‘What is the difference’-questions (e.g. what is the difference between knowledge and belief? or between a reason and a motive?'); ‘How possible’-questions (such as ‘How is it possible to measure time, given that the present has no extent, the past no longer exists and the future does not yet exist?’, ‘How is it possible for Achilles to overtake the tortoise, given that he has to traverse an infinite number of spaces in a finite time?’; ‘Why necessary’-questions? (Such as ‘Why must 2 and 2 make 4? or ‘Why can't something be both red all over and green all over?’) and ‘Do so-and-so's exist-questions?’ (Such as ‘Do universals exist?’,'Do objective values exist?'). But the form of questions is little guide as to whether they are philosophical. ‘What is matter?’ can be a philosophical question in an appropriate context, but it can be a scientific one in another context. ‘What is a dodo?’ is no philosophical question, but ‘What is belief?’ is. ‘Do dragons exist?’ is not a philosophical question, but ‘Do universals exist?’ is. ‘Why can't I go back to Africa?’ is not a philosophical question, but ‘Why can't I go back to the reign of Queen Elizabeth I?’ is. Philosophical questions cannot be circumscribed by their form. Nor can they be circumscribed by their content, since they can, in principle, be concerned with any subject matter at all – any subject matter that gives rise to conceptual confusions and unclarities. These questions cannot be resolved by the empirical sciences, since they are not empirical questions. They are all questions that are, directly or indirectly, solved, resolved or dissolved by conceptual investigation. One might therefore say, as above, that, in one sense, philosophy has no subject matter; but one might also say that, in another sense, philosophy has everything as its subject matter.

5. Philosophy and conceptual investigation

Philosophy is a conceptual investigation. This assertion can easily be misunderstood. Does it mean that philosophy has a subject matter after all – namely concepts? That would be misleading. Being a conceptual investigation does not mean being solely about concepts. The traditional questions of whether an omnipotent, omniscient and benevolent God who created the universe exists, whether we have an immortal soul, whether we are free, are philosophical. They are about whether God (thus conceived) exists, whether human beings have immortal souls and whether we are free agents. But they are answered by conceptual investigations, not by observation and experiment.Footnote ¹⁴ These investigations involve scrutiny of the concepts of GodFootnote ¹⁵, the soul, and voluntariness. Similarly, the questions of whether machines can think or whether the brain can think, are philosophical. Neither can be answered by experimental science. To deny that they are about machines, brains, and what it is to think, would be misleading. But to suggest that they are not, in a very distinctive sense, about the concept of thinking and its intelligible applicability or inapplicability to machines and brains would be grossly to misrepresent the investigation. For such questions are concerned with what does or does not make sense. And the way to examine whether something does or does not make sense, for example whether it makes sense to say that computers think or that the prefrontal cortices think, requires methodical investigation of the use of the verb ‘to think’ and its ramifying logico-grammatical connections and presuppositions. It would be mistaken to suppose that if a question is about a concept it is not also be about what falls under the concept – as if Hart's Concept of Law were not also about the law, or Ryle's Concept of Mind were not also about the mind. In truth, ‘about’ is no jack with which to lift the vehicle of philosophy.

The conceptual investigations that characterize philosophy are a priori. It is the characterization of our current concepts and the description of their relations within the conceptual field to which they belong that can contribute to the resolution of philosophical problems. The features of our concepts that are marshalled for philosophical purposes are specified by conceptual truths. Conceptual truths – for example: that events occur at a time, but do not exist at a time; that they may need space but do not occupy space; that they lack spatial dimensions; that they may have phases; that they can move, not as objects move, but in the sense that their successive phases occur at different places; and so on and so forth – are not empirical, but a priori. They describe aspects of the nature of their subject; they characterize the concept at hand; and they are manifest in the use of words.

That philosophy is an a priori investigation does not mean that it is an a priori science. Mathematics is a priori. But it is not a science after the manner of the natural sciences. It does not discover new facts about the realm of numbers and spatial relations as physics or chemistry discover new facts about the realm of nature. The mathematician is an inventor, not a discoverer.Footnote ¹⁶ What he invents are new forms of mathematical description. For mathematics is the grammar of number and space. Its business is concept-formation by means of proof. A proof grafts a new conceptual articulation onto the body of mathematics. The concepts thus formed have their ultimate (though not necessarily their proximate) rationale in providing rules for the transformation of empirical statements involving magnitudes, quantities, and so forth. Philosophy, by contrast, does not consist of a body of theorems at all. Nor is it the task of philosophy to form novel concepts by means of deductive proofs. It does not produce new rules for the transformation of descriptions of empirical phenomena. Its task is concept-elucidation for the purpose of resolving philosophical problems. That philosophy can be done in an armchair does not show that it is an a priori science, any more than the fact that it can be done peripatetically shows that it is an a posteriori one. It is not a science of any kind, not even in the Pickwickian sense in which mathematics might be said to be. But the fact that philosophy is not an a priori science does not mean that it is not an a priori investigation. The distinction between a priori investigations and empirical ones is categorial. Hence it is as deep as any categorial gulf. No philosophical question can be answered by scientific enquiry, and no scientific discovery can be made by philosophical investigation. Philosophy can reveal the incoherence, not the falsity, of a scientific claim.

6. Philosophy and linguistic investigation

Philosophy is a conceptual investigation by means of which philosophical questions are answered, or shown to be confused or incoherent. In order to answer or dissolve philosophical questions, the relevant concepts have to be examined, the presuppositions of their employment brought into view, their logico-grammatical relationships spelled out, the conceptual field within which they are embedded characterized, the human needs they fulfil specified, and the behavioural and cultural contexts in which they are at home described. But concepts are no more than abstractions from the uses of symbols, and concept-possession is no more than mastery of the use of concept-expressing symbols. So a conceptual investigation is inevitably and unsurprisingly also an investigation into the uses of words, phrases and sentences.

Linguistic investigations pertinent to philosophical enterprises are, however, very different from those of linguists. Language is the subject matter of linguistics. It is not the subject matter of philosophy. Of course, philosophy of language concerns itself with the conceptual network formed by such concepts as word, sentence, meaning, understanding, truth, reference, predication, description, quantification, and so forth. Philosophy of language is indeed about the nature of language – also about the concept of language and about aspects of the use of the word ‘language’; and so forth. But philosophy in general is not. Philosophy's general concern with language is twofold. First, confusions and unclarities of one kind or another about the uses of words, phrases and sentences is one source, a major source, of philosophical puzzlement and confusion. Secondly, describing the uses of words is one method, a major method, for answering or dissolving philosophical questions, for removing philosophical puzzlement and eradicating conceptual incoherence. Moreover, the aspects of the uses of words, and indeed the very words and phrases, that interest philosophers are, by and large, very different from those that interest linguists. It is of little interest to a linguist to investigate whether, and in what sense, one can say that events move, or whether it makes any sense to speak of visual sensations, or whether the term ‘person’ is a substance-noun. The linguistic investigations that are pertinent to philosophy in general are precisely those that shed light on philosophical problems, which are not usually of concern to linguists. Furthermore, by contrast with linguistics, no theories are involved in, and no new linguistic information is relevant to, the philosophical description of the uses of words – merely reminders of the familiar, and realization of the obvious. How can this be?

Philosophy is concerned with questions that require, for their resolution or dissolution, the clarification of concepts and conceptual networks. But, apart from the philosophies of the special sciences, most of the concepts that need to be thus clarified are ordinary ones, familiar to any mature speaker of the language, expressed by such words as ‘know’, ‘believe’, ‘doubt’, ‘certainty’, ‘mind’, ‘body’, ‘thought’, ‘understanding’, ‘true’ and ‘false’, ‘good’ and ‘evil’, ‘beautiful’ and ‘ugly’. These concepts are constituted by the sense-determining rules for the use of the words that express them. These are rules that we follow in our daily discourse. They determine the meanings of the words we use. So we are perfectly familiar with them – otherwise we would not understand what we say or know what we mean.

It is important not to conceive of such rules in too formal a manner – we are not dealing with the rules of a calculus, nor yet with regimented grammar or lexicography, let alone with rules inaccessible to consciousness ‘buried deep within the mind/brain’ (as Chomsky and his followers put it). Rather, we are concerned with the familiar rules of a human practice which all normal human beings master. Their mastery of the practice is exhibited in their uses of words in sentences, in the contextualized explanations which they give, or would accept, of what they mean and of what the words thus used mean. Sense-determining rules for the use of words can be given in various forms. They are not necessarily expressed by a meta-linguistic assertion. They may be expressed by such utterances as ‘Vixens are female foxes’, which is used as a definition. The meaning-determining rules that are the business of philosophy are commonly expressed by a priori propositions that look like descriptions but are normative in function.Footnote ¹⁷ So, for example, ‘Understanding is an ability, not a mental state or process’ is tantamount to the grammatical explanation that to say that someone understands something is not to say what mental state he is in or what process is taking place in his mind, but to indicate something he can do. Similarly, the statement that red is a colour, employed as an explanation of meaning, amounts to specifying the rule that anything that can be said to be red can be said to have a colour, just as the explanation ‘That ☞ ☐ is white’ supplies a rule for the use of the word ‘white’, namely that anything that is that ☞ ☐ colour can be said to be white. They may be articulated by explicative utterances such as ‘A proposition is true if things are as the proposition describes them as being’ – which is an explanation of a salient aspect of the use of the truth predicate. Exclusionary rules may be expressed by modal propositions about what cannot be the case. Despite looking like descriptions of de re necessities, these are tantamount to asserting that there is no such thing as … For example, ‘Nothing can be both red all over and green all over’ is tantamount to ‘There is no such thing as being both red all over and green all over’. And that in turn is equivalent to saying that it is senseless to predicate these two predicates of the same object at the same time, i.e. that this conjunction is a form of words that is excluded from the language. These, and many other forms of sentence, even though they may not appear to be expressions of rules, are in fact employed normatively. Their typical (although not uniform) role is to provide standards of correctness for the use of an expression and licenses for specific inferences.

The rules for the use of words that are of philosophical relevance cannot be unknown to speakers. For one cannot guide oneself by reference to unknown rules, and one cannot use unknown rules as standards of correctness. It would be absurd to suppose that we must wait upon future discoveries by linguists, logicians or philosophers in order to find out what we mean by the words we use and by the sentences we utter. We are not parrots, that emit words without understanding. We speak and know what we thereby say. To learn to speak is to learn to act, and our acts of speech are, for the most part, done knowingly and for a purpose. We can say what we mean and, other things being equal, what we mean and what the words we utter mean in the context of utterance coincide. So the logico-grammatical observations that are to be mustered in order to resolve philosophical problems must be news from nowhere. Indeed, one might say, with only a little exaggeration, that in philosophy, ‘If it's news, it's wrong’. It is no news that events occur, happen and take place, but do not exist; that they have no coloured surfaces but may emit a smell or make a noise; that there are colourful events, but no coloured ones; that they have phases, but no spatial parts; and so on.

If a salient method of philosophical clarification consists in no more than reminding people of the way in which they use words, then it may seem mysterious that the problems of philosophy are not solved with the greatest of ease. If every intelligent speaker of the language is perfectly familiar with the sense-determining rules for the use of the words he uses, and if these rules are a key to resolving philosophical problems, then it may seem that any intelligent speaker ought to be able to resolve such problems ad libitum. But it is not so. Why not?

Every competent speaker of the language has, by definition, mastered the use of the ordinary (non-technical) expressions of his language. Every English speaker knows, for example, how to use the words ‘nearly’ and ‘almost’. But few are able, off the cuff, to identify the differences in their use, namely: that they behave differently under negation. Nevertheless, every speaker will notice that the sentence ‘Although there were a hundred students already seated, the lecture room wasn't almost full’ is ungrammatical. The criteria for knowing what an expression means consist of correct use (and recognition of incorrect use), intelligent responses to use, and giving correct explanations of the meaning of the expression in utterances in given contexts. But mastery of use does not imply mastery of comparative use. To have mastered the uses of ‘nearly’ and ‘almost’ one does not have to have reflected on their similarities and differences. Nor does mastery of the technique of use of an expression mean that one can readily describe the complex relationships between it and the uses of related expressions in the web of words that one takes for granted in one's normal linguistic activities. A fortiori, it does not imply that one can order the expressions and types of expression whose use one has mastered so that light will be shed upon conceptual problems. But it is precisely these skills that are necessary for resolving philosophical problems.

The differences between ‘nearly’ and ‘almost’ are of no philosophical interest. The differences and relationships between the uses of the expressions ‘the mind’ and ‘the body’, and ‘my mind’ and ‘my body’, are of the greatest philosophical moment. Everyone knows how to use phrases in which the word ‘mind’ occurs – for example, to make up one's mind, to be in two minds whether to do something, to have a mind of one's own, to call something to mind, to have a thought at the back of one's mind, to have an enquiring mind, and so forth. But when confronted with the question of what the mind is, of what it is to have a mind, we are typically at a loss. For mastery of use does not require mastery of a synopsis of use. We all speak of our own and of other people's bodies. We are proud of our graceful body, complain about our aching body, are pleased with our healthy body, dislike having a sweaty and dirty body – and so forth. But when confronted with the question of what it is to have a body, how the body one has is related to the body one is, what it is that has both a body and a mind, we stumble and lose our grip on these familiar expressions. For mastery of their use does not require an overview of use. But that is precisely what is needed for the solution and dissolution of philosophical problems.

Philosophical understanding consists in possessing an overview of a conceptual network that one can bring to bear upon philosophical problems in such a manner that they dissolve, or are answered by a description of the relationships between parts of the network. To put the same point slightly differently, as both Wittgenstein and Ryle did, it consists in the mastery of the logical geography of concepts in a given domain. If one can describe the conceptual landscape, then one can (a) select from, and (b) order, the familiar grammatical rules for the uses of expressions, and (c) present a comparative morphology of uses, in a surveyable representation that will shed light upon the philosophical question, puzzlement or confusion at hand. The ordering of what we know is an ordering of the rules for the uses of expressions with which we are perfectly familiar. The comparative morphology consists, for example, in comparing the familiar use of the problematic expression with that of expressions with which it is commonly wrongly conflated, in order to highlight differences. It is noteworthy that in philosophy we already have all the information we need to solve our problems. No new information is required – only reminders of the familiar. If we do not solve our problems, it is not due to lack of information, but to lack of insight. The difficulty, the immense difficulty, is to bring into view the right aspects of usage – right for the purposes at hand; and to make the right comparisons that will bring out overlooked differences and unexpected similarities; and then to order all these in the right way – the way that will illuminate the problem, and resolve or dissolve it.

It is as if we were confronted by a pointillist painting from close up. We can see all the coloured dots, but cannot stand back to see the pattern. With the greatest effort, we can move our heads a little, and discern (and often only think we discern) a small fragment of the picture. Only the greatest geniuses, such as Plato and Aristotle, or Kant and Wittgenstein, have the ability to stand back and to see – unclearly – a significant part of the pattern, which they then describe. That is one reason why we need to study the history of philosophy.

7. Philosophical understanding: elaboration and qualification

That philosophy is a quest for understanding, rather than for knowledge, needs elaboration and qualification. It is correct to say that philosophy cannot discover new empirical truths about the world around us and can offer no theories about it on the model of the theories of the sciences. It is also correct to say that philosophy cannot discover metaphysical truths about the world – for there are none to discover (as we have seen, what masquerade as metaphysical truths are at best no more than norms of representation in deceptive guise). However, is it really true to say that in doing philosophy we never come to know things we did not already know? After all, it is not true that everyone who has mastered the use of event-designators knows how the movement of objects differs from the movement of events. Nor is it true that everyone knows that to have a body or to have a mind is not to possess anything – even though one can sell one's body and lose one's mind.

Nevertheless, in so far as philosophy provides knowledge in this sense, the form of knowledge, unlike that achieved by the natural sciences, is not that of observation, detection or experimental discovery, but of realization. And the object of knowledge is not an empirical truth, but a normative feature of our linguistic practices – of our form of representation (and hence too, an aspect of our concepts, and an internal property or relation of things). But philosophy does not teach us any new logico-grammatical nexus; we learn no new rules of inference; by contrast with the enlargement of mathematical knowledge by means of new proofs, the conceptual structure we operate remains exactly the same as before. That is why the achievement is best characterized as a contribution to understanding rather than to knowledge. For we achieve a deeper understanding of our conceptual scheme, a better grasp of its reticulations and of the comparative morphology of its elements, that enables us to avoid the confusions to which we are prone. We realize that, of course, this is the way we use these words, that, of course, this is how the uses of these apparently similar words (e.g. ‘to have a bodkin’ and ‘to have a body’) differ, and the uses of those apparently different words (e.g. ‘to sell one's body’ and ‘to sell sexual services’) are similar. The consequences of such realization can be dramatic – light dawns over the conceptual landscape. One sees the road through the woods. And one sees why one took the wrong turning and ended in a morass. That is why the characteristic reaction to an advance in scientific knowledge is ‘Goodness me, who would have thought of that!’, whereas the characteristic response to a philosophical insight is ‘Of course, I should have thought of that!’

It is true to say that philosophy does not explain phenomena as the sciences do. By contrast with theories in the empirical sciences, there is nothing hypothetical about the conceptual clarifications and elucidations of philosophy. The empirical sciences may postulate entities in order to explain observed phenomena, and go on to validate such conjectures. Philosophy, by contrast, cannot legitimately postulate entities, such as simple natures, noumena, or universals, in order to explain the a priori natures of things, or the structure of our conceptual scheme, or our uses of language. Nor is there room in philosophy for deducing the existence of such entities, on the model of inferences to the best explanation in the sciences.Footnote ¹⁸ Nevertheless, there is much that philosophy can and does explain. It explains, by description, how the various elements in the web of concepts are woven together. It explains why forms of words that at first blush appear to make sense do not, or why forms of words that appear to fulfil a given role actually fulfil an utterly different one. It explains the sources of conceptual puzzlement and confusion. And it explains how to eradicate such confusions. These explanations are logico-grammatical or conceptual.

Does philosophy not result in conceptual truths – and is that not a cognitive achievement? That would be misleading. Many of the conceptual truths in question, for example: that we know of the existence of objects in the world around us by the use of our eyes and ears – are news from nowhere. No one would have the effrontery to claim that among the cognitive achievements of philosophy is the discovery that our knowledge of other people's states of mind is warranted by what they do and say. Philosophical achievement does not consist in presenting such logico-grammatical trivialities, but in showing that the apparently powerful reasons for denying that we can know such things on the basis of such grounds are spurious. Other conceptual truths have less of an air of triviality, for example, that memory is knowledge retained, and need not be of the past; or that the beneficial for artefacts is preventive but for animals also augmentive; or that the imagination is an ability to think up possibilities. Such truths pinpoint adjacent nodes in the web of concepts. We realize that they are true when our attention is drawn to these normative connections between concepts, but they would not otherwise have occurred to most of us. Yet others are even further removed from the obvious – for example, that the limits of thought are the limits of the possible expression of thought, so that the limits of intelligible (true or false) ascription of thinking to a non-language using animal is fixed by the animal's behavioural repertoire; or that arithmetic is not a science of the relationships between numbers, but a system of interwoven rules for the transformation of empirical propositions about magnitudes, quantities, etc.; or that the conception of God as an omnipotent, omniscient and benevolent creator is incoherent. Here too we are dealing with realization – but what we have in view is not merely adjacent nodes in the web of our concepts, but a large and ramifying network. To take it in, to grasp the complex conceptual relationships that are thus articulated requires one to discern a pattern that cannot readily be detected, but rather comes into view only when the right logico-grammatical features are deployed in an appropriate manner, when the right analogies are arrayed and the illuminating disanalogies marshalled. Of course, these conceptual truths are not statements of fact. They are descriptions of normative connections within the web of concepts that constitute our form of representation. They are said to be true. Indeed, they are often said to be necessary truths. That, of course, is correct – but misleading. Their truth is akin to that of the proposition that the king in chess moves one square at a time. What we realize when a philosophical insight dawns on us is a feature of our form of representation. We attain an understanding of the way in which our familiar modes of description of things hang together.

A final important qualification and elaboration: the picture that I have presented is tailored to theoretical philosophy, i.e. to general analytic philosophy (‘descriptive metaphysics’ as Strawson misleadingly called it), to epistemology, philosophy of language, philosophy of logic, and philosophical psychology. But although the same kind of intellectual activity is appropriate in practical philosophy (moral, political and legal philosophy), other factors come into play. It is not the business of theoretical philosophy to introduce a better, logically or conceptually more perfect, language – if indeed there is any such thing. Its business is to describe our existing conceptual scheme, not to improve it, to disentangle the knots we tie in it, not to weave a new web. For the problems that plague us are rooted in the language we have, and they can be solved or resolved only by its systematic logico-grammatical description. The only concepts it can fruitfully introduce are new, technical, classificatory concepts within philosophy itself, such as concepts of inductive and deductive reasoning, of a priori and a posteriori judgements, of species and genus, of determinates and determinables – the purpose of which is to facilitate logical geography. But in practical philosophy there is room for the introduction of novel first-order concepts and for the remoulding of existing concepts. Concepts of rights (both moral and legal), of sovereignty, of the nation state, of international law, etc., have been introduced by philosophers and then moulded by fruitful dialogue over centuries between lawyers and legal and political philosophers. Similarly, concepts of liberty, justice, and democracy that were refined and elaborated by philosophical argument, have informed political debate and stimulated political and constitutional reform. Here, in the domain of the rules under which we live, and the rule-governed organization of societies in which we live, the development of the most illuminating, useful, and practical concepts to describe and prescribe normative relationships has been an integral part of philosophical reflection.

One might wonder what explains this difference between theoretical and practical philosophy. It is, I think, a corollary of the fact that at the heart of practical philosophy lie our evolving conceptions of the values which we should pursue, the norms to which we should conform, and the virtues to which we should aspire. The concepts of concern to theoretical philosophy are employed primarily in the description and explanation of what is (or is not) actually the case. But the central concepts that engage our attention in practical philosophy articulate our conception of the ideal – of what we ought to be and what we ought to do.

It is, therefore, hardly surprising that practical philosophy has a further task that has no parallel in theoretical philosophy. Since the time of Socrates, philosophers have undertaken the task of rational reflection upon the ways in which human beings should live their lives and organize their societies, of distinguishing the different values in human life and relationships, and of clarifying forms of justification and evaluation. Although conceptual clarification plays a role in such reflections, it is only part of the task. What remains is reasoned debate about the variety of values, their role in human lives, the ways of ordering them, of the incommensurability of values, of what is right and what is obligatory, of the nature of conflicts of duty and of the place of the virtues in human life. Similarly, it falls to political and legal philosophy not only to clarify (and sometimes refashion) salient concepts in political and legal discourse, but also to reflect on the justifications of various forms of legal and political institutions and to recommend legal and constitutional arrangements suitable for rational beings living under the rule of law.

8. Can there be progress in philosophy?

If, in the sense explained, philosophy is not a cognitive discipline, can there be said to be progress in philosophy? Progress characterizes the sciences. But how can there be progress in a subject that has no subject matter in the manner of the sciences, and that adds nothing to human knowledge save for the realization of the ways in which various elements in our conceptual scheme hang together? Is lack of progress in philosophy not born out by the fact that problems that were discussed by Plato and Aristotle are still being discussed today?

There is no progress on the model of the sciences. In the sciences, knowledge is cumulative, and hierarchies of theories are constructed. In the natural sciences, advances in instrumentation make possible new factual discoveries, which lead to new questions, and that in turn leads to new theories that explain the phenomena. Advances in scientific theory and in instrumentation in turn generate advances in technology. Philosophy, however, is not hierarchical. It has no foundations. It erects no theoretical structures on the insights and conceptual clarifications it achieves. There is no instrumentation to aid observation and empirical discovery – but, of course, there is neither observation nor discovery. There is no technological spin-off from theories, since there are no theories that are validated in experience. Nor are there theorems that are proven and then applied to experience. No men are sent to the moon on the back of philosophical elucidations nor is anyone guided through the seas by the charts of logical geographers.

Nevertheless, there are three senses in which there can be progress in philosophy: discriminatory, analytic and therapeutic.

First, clearer distinctions are drawn between forms of reasoning, types of proposition, and kinds of concepts (discriminatory progress). We distinguish between deductive, inductive and other forms of reasoning, and thereby are able better to handle conceptual problems that arise out of different kinds of argument. We distinguish between the question of how a truth is learnt and what are its grounds, and so are able to separate questions that were once conflated. We distinguish between determinates and determinables, and between determinate-determinable relations and species-genus relationships. And so on. Progress, in this sense, often appears to be less than it really is. For such distinctions are rapidly taken for granted, and we forget that the articulate differentiation of inference patterns, proposition types and kinds of concepts are hard-won insights obtained from philosophical reflection. So the progress that has been made is sometimes not recognised for what it is.

Secondly, there is progress in the characterization and clarification of problem-generating concepts (analytic progress). There has been advance in the philosophical understanding of such concepts as truth, existence, probability, mind, person, goodness, rights, obligations, i.e. improvements in the descriptions of the conceptual network surrounding these pivotal, but problematic, concepts. And there has been advance in seeing what was awry with a variety of explanations advanced by past thinkers.

Thirdly, there have been advances in dissolving certain kinds of conceptual confusion (therapeutic progress). No longer need we puzzle ourselves over the question of whether our knowledge of necessary truths is innate or acquired, or whether the nature of substances is knowable or not, or whether the self is given in experience, presupposed by experience, or is transcendent. No longer need we strive to justify inductive reasoning, to prove the existence of the world, of universals, or of events. Here too progress is often not discerned, since sometimes the refuted arguments and the futile endeavours vanish from sight, and tempting pathways to illusion and confusion are permanently closed off and forgotten.Footnote ¹⁹

Nevertheless, precisely because philosophy is a contribution to understanding and not to knowledge, these forms of progress may be less than they appear. For they are distinctly precarious, for two reasons.

First, a conceptual field may be partially illuminated for a generation or two, only to be cast into shadow again. For cultural innovations, technical or theoretical, occur (e.g. the invention of the computer, or of function-theoretic logic) and novel scientific theories are introduced (e.g. quantum mechanics, relativity theory), which cast long shadows over conceptual articulations previously clarified. That may require old ground to be traversed afresh from a new angle, as when the concept of mind needed to be clarified yet again in response to the temptation to conceive of the mind on a computational model, or the concept of natural language reconsidered in the light of the invention of the predicate calculus.

Second, if there can be progress of a kind that is not akin to progress in the sciences, so too there can be regress of a kind that does not occur in the sciences. Precisely because philosophy has no foundations, because it is not hierarchical, because it produces neither theories validated in experience nor theorems proven and then applied to experience, because it is not the basis for technology of any kind, distinctions may be lost from sight, methods of clarification may fall into disuse, and the skills they require may vanish. Distinctions that were clearly drawn may become muddied through a novel conundrum that is mishandled – as the insight that all a priori knowledge is of necessary truth became muddied by the confused idea that knowledge that the standard metre is a metre long is both a priori and contingent.Footnote ²⁰ Old confusions may prove irresistible to a new generation (e.g. the attraction of talk concerning the self, mystification about consciousness, the allure of metaphysics, conceived as a science of objective necessities). For conceptual confusions are comparable to diseases – diseases of the intellect. They may be cured for one generation, but the virus may undergo mutation and reappear in even more virulent forms.

Precisely because philosophy is not a quest for knowledge but for understanding, what it achieves can no more be transmitted from generation to generation than virtue. Philosophical education can show the way to philosophical clarity, just as parents can endeavour to inculcate virtue in their children. But the temptations, both old and new, of illusion, mystification, arid scholasticism, scientism, and bogus precision fostered by logical technology may prove too great, and philosophical insight and overview may wane. Each generation has to achieve philosophical understanding for itself, and the insights and clarifications of previous generations have to be gained afresh.Footnote ²¹