University of St Andrews AHCR web site CSMN web site
Modality Project

The Metaphysics and Epistemology of Modality (2003-2005)


Funded by the AHRC.

TWiki pages for this project.
Modality Bibliography (available from the project TWiki pages).

The Project Team


Principal Investigator: Bob Hale (Sheffield).

Independent Auditor: Frank Jackson (ANU).

Other Project Members:
Ross Cameron (Leeds),
John Divers (leeds),
Aviv Hoffmann (Open University of Israel),
Josh Parsons (University of Otago) and
Anna Sherratt.

The Research Problem


Modal notions—notions of necessity and possibility—are implicit in our understanding of a wide class of fundamental concepts: for instance, the notion of one proposition following logically from others, the idea of what would be true if such and such were the case, the idea of a law of nature (as opposed to a mere regularity), the notion of causation (that is, of the occurrence of certain events and conditions necessitating that of others) and the idea that one kind of property or state of affairs may supervene on others (i.e. that things can never differ in the former kind of way without differing, somehow, in the latter kind of way)—each of these, explicitly or implicitly, involves modal notions of one sort or another.

This breadth of implication in ordinary thought makes it philosophically especially important to try to achieve a satisfactory analytic understanding of the content of modal claims; and the extensive attention given to modality by analytical philosophers, particularly in the latter half of the twentieth century, is testimony to this sense of their importance. Yet the key issues—concerning the nature of modal facts, and the ways they might be known—seem to remain as perplexing as ever, with almost no view of them apparently too extreme to be adopted and defended with ingenuity by leading theorists in the field. Thus the contemporary literature offers accounts of so innocent-seeming a claim as that Margaret Thatcher might never have become Prime Minister ranging from the proposal that it is made true by the existence of a flesh-and-blood counterpart of Margaret Thatcher inhabiting a concrete, natural world very much like that in which we live but spatially unrelated to it, who fails to become Prime Minister of the counterpart of Britain in that world, to the suggestions that it is made true by linguistic convention, or that it is merely an expression of our abilities of imagination. The broad aims of our proposed project will be to clarify the pressures which have generated such extreme proposals, to review what should be demanded of a satisfactory account of modality, to re-assess existing proposals in the light of that clarification, and to pave the way for-and indeed to make-progress on this difficult range of issues.

The main questions about modality naturally arrange themselves in a series of dilemmas, with the most fundamental matters looming early and the more specific issues that the theorist needs to confront presenting themselves further down. Our programme of work will concentrate (as has been normal) on the absolute (metaphysical) modalities—on the logically strongest notion of necessity and the correspondingly weakest notion of possibility respectively. (It is plausible-though a matter for investigation-that other, more specific modalities can, for the most part, be characterised in terms of these absolute notions.) The first matter for attention must be the strong current of scepticism towards the absolute modal notions exhibited in much contemporary philosophy and owing in large measure to the contributions of Quine. The most basic dilemma which any thoroughgoing discussion of modality must confront is accordingly

To sanction modal discourse, or to recommend abstention from it?

The case for abstention has to do with the unclarity of modal notions, with the difficulties in explaining them without circularity or ineptitude, with the obscurity of modal epistemology, and with the suspicion that modal notions have no important work to do—that they are, for instance, eliminable without serious loss from the vocabulary of successful empirical science. Against that needs to be set the consideration already stressed, that modal notions are implicated in very many other, fundamental ways of thinking. And more specifically, there are certain arguments of a ‘transcendental’ nature (McFetridge, Hale, Wright) which, if sustained, show that modal thought is unavoidable. Two questions to which the early stages of the project will give attention, and on which we would hope to make progress, are these:

#1: Are there compelling arguments for supposing that modal thought (embedding the absolute modal notions) is essential?

#2: Do these arguments (or are there others which) suggest that at least some modal judgements must be a priori?

If abstention from modal discourse proves to have no compelling motivation, then we acquiesce in a way of thinking which, taken at face value, seems to carry a commitment to a special category of modal states of affairs, perhaps calling for a correspondingly special category of cognitive (rational) powers. Such an acquiescence is implicit in Dummett’s well-known characterisation of the philosophical problem of necessity in particular as consisting in the questions, What is its source?, and How do we recognise it? It’s important to recognise that there is space for a position which rejects these questions without dismissing their presuppositions (that is, without replying: “There is no source of necessity, because there is no necessity” and “We don’t recognise any such thing”.) Such a position is the first option in the next large dilemma confronting the theorist:

Quietism, or Activism?

Quietism rejects Dummett’s questions but is nevertheless content to acquiesce in modal discourse. For the quietist, there is no sound analytical philosophical project which seeks to uncover the source(s) of metaphysical necessity and possibility, or to give a systematic account of how we might know of them. Such a view might flow, of course, from a mistrust of such projects in general, of the kind arguably found in the writings of the later Wittgenstein and manifest in certain contemporary philosophers such as John McDowell. Our primary concern with it, however, will be fuelled by certain points which are specific to the case of modality. An example is a dilemma formulated by Simon Blackburn: whatever considerations a theorist proposes as underlying the necessity of a proposition, the question must arise whether they themselves are conceived as obtaining of necessity or of contingency. If the former, then we have merely explained one necessity in terms of another, and have not attained the desired general account of how necessity originates. But if the latter, then we have grounded necessity in mere contingency-and that must surely mean that the explanation is inadequate. This straightforward challenge illustrates a more general concern about the mutual coherence of the constraints which might individually seem compulsory for a satisfactory ‘account’ of metaphysical necessity and possibility, a concern which arises in connection with some of the most sophisticated recent work on the issues (for example, Christopher Peacocke’s recently proposed ‘Principle-Based’ account.) A question that must be taken early in the proposed project is therefore:

#3: Are there compelling local reasons for quietism about modal notions—or at least for supposing that there are significant and topic-specific a priori limits on what can be accomplished by a philosophical account of modality?

Activism is the attempt to give a positive analytical account of the content of modal discourse and of its epistemological status. The major alternative immediately confronting activism is:

Realism, or Non-cognitivism?

It is with this division that most recent and contemporary work on modality has been preoccupied, (with the appropriateness of activism itself accordingly somewhat taken for granted.) A realist account will attempt to sustain a view of modal propositions as indeed depicting a special category of objective fact. Such accounts have to face up to a number of constraints. Prominent among them is what Peacocke has termed an ‘integration challenge’—the challenge of squaring a proposed account of the character of modal fact with the possibility of our knowledge of such facts (and indeed, with the possibility of knowledge of them by what we pre-theoretically regard as satisfactory ways of achieving it.) It’s striking that philosophers working in the field have not always clearly perceived these intertwined explanatory obligations, and that the achievements of some of the most developed programmes—for instance that of David Lewis—would seem to come somewhat lop-sided as a result. (There are parallels here between the tensions famously emphasised by Paul Benacerraf thirty years ago between straightforward construals of the semantics of mathematical discourse and a generally naturalistic conception of human knowledge and its limitations.) In general, it seems that a satisfactory realist account should be responsive to at least the following four constraints:

(a) The account should be based on a principled rejection of abstentionism and quietism, and its detail should respect the grounds for that rejection. (If, for example, abstentionism falls to a compelling argument that modalising is indispensable for certain theoretical purposes, then a realist account must respect whatever bearing on the content of modal statements is carried by that argument.)

(b) The account should be properly ‘integrated’: the conception it provides of the truth-conditions of modal statements should put no obstacles in the way of our knowledge of them, by procedures on which we actually rely.

(c) Relatedly, the account should allow for a satisfying connection between the content it ascribes to modal judgements and, in the broadest sense, our interests and purposes in making such judgements—their role in our intellectual lives.

(d) At least ideally, the account should contribute towards making conspicuous the logical relationships among modal statements (and here, of course, Lewisian possible-worlds semantics is a great success.)

Our project will attend in detail to the question

#4: How should these constraints be best formulated and refined, and what additional candidate constraints should be imposed?

Important contemporary realist proposals include

  • Lewis’s proposal that modal discourse describes other possible worlds, conceived as really existing material complexes;
  • abstractionist proposals (of the kind that have been defended in general by Kripke, van Inwagen and Lycan and in specific versions by others) that the actual world is the only material world and the other possible worlds are unactualised states of affairs (Plantinga), uninstantiated ways the actual world might have been (Stalnaker, Forrest) unactualised recombinations of actually existing entities (Cresswell, Skyrms) or consistent world-descriptions (Adams);
  • Fine’s generalised essentialism;
  • Peacocke’s principle-based conception;
  • the combinatorialism of Armstrong.

A major part of our work will be devoted to a review of the question,

#5: How do these various influential proposals fare in the light of the overall constraints on realist accounts pursued under #4?

The non-cognitivist tradition about modality includes the conventionalism of the logical positivists and the anthropocentric approach to logic and mathematics of the later Wittgenstein. The tradition encompasses accounts of significantly different kinds—varying from forms of expressivism, according to which modal judgements have no distinctive subject matter but serve merely to express attitudes to the embedded (non-modal) propositions, to accounts which allow that the role of the modal operators is indeed to contribute towards formation of a distinctive kind of assertoric content, of which however some form of anti-realist treatment is then proposed. Many of these types of accounts are subject to internal structural difficulties, some of which (for instance, the ‘Frege-Geach’ problem) are characteristically confronted by certain kinds of non-cognitivist proposal as a class, irrespective of the subject matter, while others—to do, for instance, with iterated modalities-are specific to modality itself. The project will endeavour to determine

#6: What constraints should be imposed on a satisfactory non-cognitivist conception of modality?

It is to be expected that versions of constraints (a), (c) and (d) above, to which realist accounts are subject, will continue to apply. In addition, it would seem reasonable to require that a non-cognitivist account be motivated by compelling grounds for rejecting realism and that the nature of these grounds in turn be properly respected by the detail of the account. It’s only in the light of a developed account of these constraints that a satisfactory treatment can be expected of

#7: What are the prospects for a workable non-cognitivist account of modality?

So much by way of outline of the overall architecture of the project. It remains to mention certain additional specific issues on which research is intended (in some cases as part of addressing the more general concerns just outlined.)

  • (generalisations and descendants of) the argument in lecture III of Kripke’s Naming and Necessity against physicalistic reductions of sensation (and its implicit demands on the relationship between conceivability and possibility)
  • the status and epistemology of a posteriori metaphysical necessities
  • the implication of necessity in the notion of logical consequence, and the corresponding bearing of philosophical issues about modality on the status of logic
  • the prospects for a traditionally rationalistic conception of philosophy, aimed at insight into how the world necessarily is.
We invite students and scholars interested in visiting or joining Arché to pursue research connected with any of these projects to visit our visitors pages.