After a long period during which metaphilosophy was shunned as philosophers chose to focus instead on first-order philosophical problems, reflections on the method of philosophy are once again occupying centre stage (Williamson 2007; Chalmers et al. 2009; Overgaard et al. 2013; D’Oro and Overgaard 2017). Collingwood was the author of two explicitly metaphilosophical treatises, An Essay on Philosophical Method (1933) and An Essay on Metaphysics (1940), and regarded the question “what is philosophy?” to be part of philosophy: “Reflection on it is part of itself” (EPM : 1). There has therefore never been a better time to revisit Collingwood’s conception of the role and character of philosophical analysis and assess where he stands on a spectrum of views that range, as in the title of a recent collection, from the armchair to the laboratory (Haug 2013). Clearly Collingwood rejected both these extremes: he was neither the kind of metaphysician who conceived of philosophy as a form of armchair science seeking after ontological truths through reflection alone, nor did he think that philosophers could abandon the reflective method in favour of the experimental one.
But where exactly did he stand? Would he have been sympathetic to the Quinean view that philosophy is continuous with natural science? Or would he have taken the more moderate Lockean view, recently revived by Jackson (1998), that philosophy is an underlabourer to science whose task is to establish what aspects of the manifest image are compatible with the scientific image? While the editors of this collection are united in thinking that Collingwood had a much more robust sense of the role of philosophical analysis (and it may be better to declare one’s views from the outset, rather than smuggle them in as a fact of the matter), the literature is quite divided on this issue and it is not always easy for the reader to navigate their way round it. Some think that Collingwood was indeed in the business of liquidating philosophy, even if the preferred special science in which he thought it should be dissolved was not physics but history. On this view Collingwood conceived of philosophy not as a normative but as a purely descriptive undertaking, whose task is to describe the belief systems of different people at different times and places. Metaphysics, on this view, is superseded not by physics but by cultural anthropology. This is still an influential view of Collingwood’s philosophy with a long pedigree (Donagan 1962; Rotenstreich 1972; Toulmin 1972), and one that is shared by Williams (2006) (reprinted in this volume). Others have thought that Collingwood had a more robust sense of the distinction between philosophy and the special sciences and that he saw philosophy as distinct from them: rather than being continuous with any of the special sciences philosophy is in charge of excavating the presuppositions on which they rest. On this view there are two levels of investigation. The first-order level is that of the special sciences, which are the laboratories of knowledge; the second-order level is that of metaphysics, whose task is to make explicit the presuppositions that are entailed by the questions that are characteristically asked (and answered) in the sciences. While the metaphysician does not dictate to the historian or the physicist what their method should be, metaphysical analysis does make explicit to them what their method is, what is their distinctive explanandum, and how the subject matter of a special science differs from that of other forms of inquiry. The task of excavating these presuppositions does not therefore consist in describing what people believe (as if philosophy were a form of cultural anthropology) but what presuppositions are entailed by the characteristic questions asked in a given form of inquiry. In other words, on this reading, the task of metaphysical analysis is to uncover the presuppositions that are constitutive of a given form of knowledge and are mandatory for its practitioners. The consequence of this reading is that, far from being science’s underlabourer (the Lockean view) or being continuous with science (the Quinean view), metaphysics is an autonomous form of inquiry whose subject matter are the presuppositions which are constitutive of the knowledge claims arrived at in the special sciences. To be clear, this reading of Collingwood’s metaphysics of absolute presuppositions does not mean that he was in the business of defending the Cartesian view of metaphysics as an ontologically first science, lying at the roots of the tree of knowledge whose trunk is physics, and from whose branches (the special sciences) hang the fruits of knowledge which humanity enjoys (health and technological advances). Metaphysics, on our reading, is a first science only in the order of logical priority because it uncovers those presuppositions which the practitioners of the special sciences must make, as a matter of logical necessity, in order to engage in the production of knowledge. To read Collingwood in this way is to acknowledge a distinction between the modus operandi of the special sciences and that of philosophy: the former are involved in the production of knowledge; the latter enables one to understand the conditions on which knowledge rests, rather than produce a different kind of (metaphysical ) knowledge.
The divide between these interpretative lines is often traceable to whether or not one thinks that An Essay on Metaphysics , which is normally read as advocating the dissolution of philosophy into history, is continuous with the earlier An Essay on Philosophical Method or whether it marks a historicist departure from it. A discussion of this issue is beyond the scope of this introduction but for the discontinuity thesis, which often goes hand in hand with the claim that Collingwood sought to liquidate metaphysics into history, the reader might consider Donagan (1962), Rotenstreich (1972), and Toulmin (1972). For the continuity thesis, which tends to go hand in hand with the claim that Collingwood did not relinquish a commitment to philosophy as an autonomous discipline with its own method and domain of inquiry, the reader might look at Connelly (1990) and D’Oro (2002). What is at stake between those who see discontinuity between An Essay on Philosophical Method and An Essay on Metaphysics and those who do not is the status of the a priori in Collingwood’s later philosophy, and whether the notion of the a priori that is captured by the claim that knowledge rests on presuppositions of a certain kind (absolute presuppositions) is sufficiently robust to uphold the autonomy of philosophy or whether on the other hand it is weak enough to support the claim that philosophy is either a form of cultural anthropology or perhaps continuous with it.
An Essay on Metaphysics defends the view that all knowledge rests on some presupposition or other and that certain presuppositions, which Collingwood calls “absolute”, are constitutive of a particular kind of knowledge in the sense that the form of knowledge which they support would not be possible without them. The principle of the uniformity of nature, for example, is implicit in any attempt to extrapolate general laws from particular observations and is thus presupposed by any attempt to advance empirical knowledge. The principle of the uniformity of nature is therefore absolutely presupposed by the making of inductive generalizations and constitutive of the kind of knowledge that natural science yields.
The view that the presuppositions on which knowledge rests are constitutive of the form of knowledge which they enable already marks an important departure from a strong (Kantian) notion of the a priori, where “a priori” means valid at all times and places. But how weak or robust is Collingwood’s constitutive notion of the a priori? On the one hand one could argue that since metaphysics, as Collingwood conceives it, begins from knowledge as “we” have it, and since the “we” changes with the location of the subject in space and time, the presuppositions that metaphysics uncovers are historically parochial rather than universal and valid at all times and places. On the other hand, one could argue that although the notion of the a priori captured by the view that absolute presuppositions are constitutive of forms of inquiry is weaker than the Kantian notion of the a priori, it is not reducible to a mere (time-relative) historical a priori since it makes a claim that is stronger than “this is what ‘we’ in a given spatio-temporal slice, believe to be the case”: the principle of the uniformity of nature, for example, is not merely something that we (here and now) believe to be true. It is rather a principle that anyone engaged in empirical enquiry is committed to, whether they are aware of it or not. Since the principle of the uniformity of nature is entailed by the advancing of inductive generalizations, anyone, at any time, who extrapolates an empirical law from particular observations can do so only by presupposing the principle. Be that as it may, as Williams (in the chapter reprinted here) says, there is often an ambiguity in the way in which philosophers who abandon foundationalism use the word “we” and it is this ambiguity that interpreters trying to make sense of the notion of the a priori at work in An Essay on Metaphysics must grapple with and try to resolve.
This interpretative divide has substantive consequences for how one understands the sense in which Collingwood belongs to the historicist tradition. For the term “historicist” h...