image source https://www.flickr.com/photos/127130111@N06/29736944616/in/photostream/  https://creativecommons.org/licenses/by/2.0/

Project Overview

We will organise our researches into two concurrent streams, each conducted collaboratively by a dedicated core project team working together on a well-defined set of topics to a set schedule with clear milestones and goals, and supported by an international network of philosophers and scientists with research interests in the field who will come together regularly to give feedback on the local research and to present their own ideas in progress.

There is a striking match in the broad dialectical architecture of the problems of Self-knowledge and the A Priori —something that motivates the co-ordinated approach to the two sets of issues that we are proposing. For both streams, our project naturally divides into three successive large tasks, to which we will devote three successive phases of research, of approximately nine, twelve and nine months respectively. The final months of the overall programme will be devoted to a collective critical overview of drafts of our findings and to writing up.

We give below an account of the philosophical research agenda of the two streams and of the principal hypotheses to be interrogated. Note that in each case these hypotheses have the status of conjectures that we anticipate vindicating.

Research Agenda: Self-Knowledge

Phase 1: The problem of characterisation.
It is essential to pose the philosophical problem of self-knowledge correctly and in full generality. There is a long-standing tendency in the literature towards casual use of phrases such as ‘privileged access’ and ‘first person authority’. However it is clear that there are at least three different apparent facets involved in our ordinary idea of the special relationship (we here, for want of a better, retain the caption, ‘privileged access’) between selves and their mental states:

Immediacy—the idea that we each have a kind of direct awareness of our own mental states that is denied to others;
Authority— the idea that our opinions about our own mental states are usually correct and are to be deferred to; and
Salience— the idea that, in the normal run of cases, our own mental states are evident to us.

It is also clear that these notions, though intuitive, are imprecise and need elaboration, qualification and maybe further division before the project of attempting a philosophical account of them can be well framed.

The task for the first phase of the self-knowledge stream will be, what has never been done, to review and accurately describe this prima facie trifecta of privileged access in detail: to explore the limits of the three aspects and the contrasts between the ways they are realised for different types of mental state, to re-think the shortcomings of the classic Cartesian view of them, and to consider whether other aspects should be added to the list. This will provide the stage setting needed for:

Phase 2: Scepticism and Deflationism.
There has been a recent and influential tendency towards sceptical critiques of the very idea of privileged access. They include Carruthers’ empirical-scientifically-based arguments for the Rylean idea that we know of our own attitudes in much the same way that we know of others’; Schwitzgebel’s arguments for the unreliability of introspection; Snowdon’s arguments that neither authority nor salience really applies even in the case of sensation; and, perhaps most influential of all, Williamson’s epistemological arguments against ‘luminosity’. We will scrutinise these sceptical arguments in detail.

Hypothesis 1: While there are some important lessons to learn from the sceptics about the limitations of privileged access, and about variations in what it consists in for different types of mental state, the idea survives that selves do have something worth describing as ‘privileged access’ throughout the range of mental states to which it intuitively applies.

Deflationism contrasts with scepticism in that it acknowledges privileged access but contends that nothing cognitively remarkable is entrained, in particular nothing in tension with naturalism. One kind of deflationism holds that the real phenomena at which ‘privileged access’ gestures are linguistic. Two examples are the expressivism developed and defended by, for example, Bar-On and by Finkelstein, and the Wittgensteinian idea that the notion of privileged access as something cognitively special reflects a misunderstanding of certain ‘grammatical’ features of the language-game of self- and other-ascription of mental states. A quite different constitutivist version of deflationism, variously developed in the writings of, for instance, Davidson, Bilgrami, and Wright, contends that it is in the metaphysical nature of a subject’s mental states to coincide broadly with what she herself takes them to be. A third proposal which some have viewed as deflationary in spirit is the so-called transparency thesis that originates in Evans and has been developed by, for instance, Boyle and Byrne: the core idea is illustrated by the suggestion that to find out whether, for example, one believes that P, one has only to investigate whether P; no ‘looking within’ is needed. We will explore and critique these views.

Hypothesis 2: None of these deflationary proposals can provide a satisfactory account of all aspects of privileged access.

Phase 3: Indispensability and Constructive proposals; Stocktaking.
The last phase of our planned research will juxtapose exploration of the question whether self-knowledge, exhibiting the features of privileged access, is essential to rational deliberation and decision-making with a critical review of the very few extant constructive, naturalistic models of how self-knowledge is achieved.

Naturalistic proposals to be reviewed include offerings by Armstrong, Lycan, Fumerton, Dretske and Tye. They range from the postulation of internal scanning mechanisms, to the attempt to work out non-Cartesian conceptions of ‘inner sense’, to applications of the idea of ‘displaced perception’ (as when one sees that a fuel tank is empty by observing the position of the needle on the fuel gauge.)

Hypothesis 3: There is in prospect no satisfactory naturalistic model of how beliefs about one’s own mental states can be knowledgeably attained that simultaneously conserves the trifecta of privileged access, is of sufficient generality to cover the intuitive scope of privileged access, and is free of serious internal difficulties.

There is relatively little in the literature by way of explicit argument for the indispensability of privileged access in rational deliberation, though important prospectively helpful lines of thought are to be found in writings of Burge, Moran and Peacocke; this is accordingly a place where our researches will need to prove themselves especially innovative.

Hypothesis 4: Beliefs about ourselves that exemplify each of the trifecta of privileged access are indeed an essential component in rational agency.

It is our expectation that clarifying the grounds for the indispensability of self-knowledge in our rational lives will provide a better understanding of what it is for, and that this will prove a crucial first step in in a better understanding of its nature.

The reader will note that if each of hypotheses 1-4 is redeemed, they will collectively enforce a conception of self-knowledge that is both an ineliminable component of rational agency and holds out no prospect of naturalistic explanation or deflation.

Research Agenda: The A Priori

Our work is again naturally corralled into three phases corresponding closely to those outlined for Self-knowledge.

Phase 1: The problem of characterisation.
The classic characterisation of a priori knowledge (and mutatis mutandis, of a priori justification) is that it is knowledge achieved in a manner ‘independent of experience’. The familiarly very resilient problem has been to say, without over- or under-extending the intuitively intended category, what for these purposes should count as ‘experience’, and what forms of independence of experience are to the purpose. If, for example, ‘independent of experience’ is taken to require only independence of sense experience, then knowledge of the gist of a daydream is going to count as a priori; a bad result if we wish to characterise a mode of knowledge of special relevance to basic logic and mathematics. For the same reason we will not want ‘independent of experience’ to characterise any and every type of innate knowledge—for instance, of depth grammar—that we might happen to have. On the other hand, we will not want to outlaw the kind of dependence on sense experience involved in keeping track of the detail of a written mathematical proof, or taking in a complex geometrical diagram.

In the first phase of the research, we will explore this oscillation in detail, reviewing all the more notable recent struggles to formulate an intuitively satisfactory, stable version of the classic characterisation. Our anticipated findings are:

Hypothesis 5: There is actually no intuitively satisfactory development of the classic characterisation.
Hypothesis 6: A promising alternative is to allow that experiences may play a role, even an evidential role, in a process leading to priori knowledge and to focus instead on the shortfall between what is strictly justified by such experiences, when they are involved, and the stronger claims—claims of necessity, relative indefeasibility, or certainty—that we take to be justified by the process in question.

Phase 2: Scepticism and Deflationism
A persistent trend towards scepticism about the a priori has been closely connected with the problem of characterisation. Both (the first two thirds of) Quine’s ‘Two Dogmas (though officially directed against the analytic-synthetic distinction) and Hawthorne’s more recent attack, for instance, contend that no account of the notion is to be had that is sufficiently clear to bestow on it any interesting epistemological role.

Hypothesis 7: These forms of scepticism about the a priori can be effectively addressed by an improved characterisation along the lines gestured at above.

A different template for scepticism is manifest in critical work by Kitcher and, more recently, by Williamson, both of whom credit the notion of the a priori with sufficient clarity to underwrite argument, respectively, that it has no application and that it essentially over-generalises to include a wide range of cases that supporters of the notion would want to regard as a posteriori. Kitcher’s argument has been generally rejected (Casullo and others) on the ground that it depends on the unmotivated assumption that beliefs that are warranted a priori should be empirically indefeasible. This needs further investigation. Williamson’s argument has to date provoked little published counterargument. It is the focus of work in progress by Melis and Wright, arguing that it rests on a subtle oversight that can be forestalled by clarification of the kinds of roles that empirical lemmas may and may not legitimately play in a priori confirmation. We propose:

Hypothesis 8: There is no extant, compelling form of scepticism of this kind about the a priori.

One kind of deflationism about a priori knowledge is manifest in the tradition, originating with Mill and continued by the Logical Positivists, of attempting to restrict its scope to a few explicit definitional truths, together with their ascertainable consequences. The problems of this kind of account are very familiar. The more interesting descendant of this proposal embodied in Boghossian’s notion of epistemic analyticity is no longer deflationary and will be considered in phase 3. A potentially more competitive deflationism is presented by non-cognitivist accounts of the processes involved in a priori judgement; for example the idea recurrent in Wittgenstein’s Remarks on the Foundations of Mathematics that our acceptance of proof always involves a cognitively optional element, akin to a kind of decision, occasioning some form of modification of the content of what is proved. The resilience of this idea has not been generally appreciated, and it will be an important part of our project to work out its strongest development. We nevertheless anticipate finding for the following conjecture:

Hypothesis 9: Non-cognitivism will prove to founder on the fact that our abilities of basic (non-inferential) a priori judgement are productive—they apply to an unlimited array of previously unconsidered cases—in much the same way that our ability to parse novel utterances in our native tongue is productive, and that the non-cognitivist lacks the resources to provide a satisfactory account of our non-collusive agreement in novel basic a priori judgements.

Phase 3: Indispensability and Constructive proposals; Stocktaking.
The last phase of our planned research will, following the model of phase 3 of the Self-Knowledge stream, juxtapose exploration of the question whether a priori judgement is an ineliminable part of rational enquiry with a critical review of the principal positive proposals about how recognition of truth a priori might proceed.

In recent literature, the supported positive proposals essentially boil down to two: the idea that knowledgeable a priori judgement is facilitated by reflection on content, typified by Boghossian’s rehabilitation of the notion of epistemic analyticity (which notably relies on an assumption that we have substantial self-knowledge of content, and so is a point at which our two research streams make direct contact), and the idea, defended by BonJour and others, that basic a priori knowledge is delivered by a faculty of rational intuition, or insight. Work in progress by the principal applicant suggests that there is point to both proposals (and that there is corresponding work to do for something akin to Kant’s distinction between the analytic and the synthetic a priori). However, we expect to support

Hypothesis 10: There is no prospect of any genuinely explanatory, general account of how the recognition of analyticity on the basis of understanding of meaning may be supposed to work, nor of the modus operandi of a faculty of rational intuition.

Even among philosophers hospitable to the notion there has been, again, relatively little explicit argument for the indispensability of a priori judgement. Two exceptions are the important systematic defence of the role of rational insight in recent work of BonJour, and an unrebutted argument by the principal applicant, published almost three decades ago, that Quine’s global empiricism can provide for no coherent scientific methodology. We will endeavour to develop these lines of thought further and to uncover other original lines of support for the following:

Hypothesis 11: A thinker who eschewed all judgement made a priori could rationally undertake no more than a small fraction of the projects of enquiry that are engaged in by science, and indeed by ordinary thought.

The reader will note that if each of hypotheses 5-11 is redeemed, they will collectively enforce a conception of a priori knowledge that is both an ineliminable component of rational enquiry and holds out no prospect of naturalistic explanation or deflation.


Milestone Seminars
To ensure that progress is made at an appropriate pace, and that outputs are delivered to the agreed timescale, we will convene six-monthly Milestone Seminars throughout the lifetime of the project, at which the core team will take joint stock of all aspects of progress. Conclusions from these meetings will inform regular progess reports to our Academic Auditors and to the John Templeton Foundation.

Academic Audit
A process of annual Academic Audit will be key to the success of the project. The Auditors will visit Stirling once each year. Prior to their visit they will receive work in progress from the core team and self-assessments. During their visits they will interview project members, and hear research presentations. Their feedback will be used to guide future research and as an additional resource in preparing progress reports to the John Templeton Foundation.

Professors Alex Byrne (MIT) and James Pryor (NYU) have agreed to serve as Academic Auditors for Knowledge Beyond Natural Science.