Philosophy of Action from Suarez to Anscombe, edited by Constantine Sandis, features an introduction and 11 chapters, each of which focuses on the philosophy of action of one or more figures from the period, or on arguments drawn from those figures. The collection stems from the Philosophical Accounts of Action conference at Senate House, London in May 2013, and all of the chapters, along with the collection’s preface, were previously published as a special issue of Philosophical Explorations (21(1), 2018). The collection does not aim at total coverage of the period, but rather as Sandis writes, “showcases a selection of the diverse interests in action that philosophers have had in modern times, from around 1500 to the mid-twentieth century” (3).

The aims of this collection are laudable. As philosophy of action has only recently grown in prominence as a theoretical subdivision, we suspect that much historical thought on the topic is relatively underexplored. As non-historians, the prospect of learning about the kinds of issues exercising philosophers with a broad range of approaches during the titular period was very welcome. From this perspective, we hoped to find a collection of essays accessible to readers not already well-versed in the literatures on the thinkers covered. A number of the contributions provided just that (e.g. the chapters by Pink, Radcliffe and McCarty, D’Oro), although other contributions represent a difficult read for the uninitiated (e.g. the chapters by Stoneham, Zielinska, Tanney). Although the collection may not, for this reason, become essential reading (like e.g. the historical elements of the excellent A Companion to the Philosophy of Action (2010), which Sandis co-edited with Timothy O’Connor), it provides a fascinating insight into some of the action-theoretical disputes during an important period of intellectual history, and the reader is left with the firm feeling that serious philosophical interest in action-theoretical issues was widespread during the period covered by the collection.

One parenthetical gripe before turning to the chapters: the reader is requested to cite the original page numbers from the Philosophical Explorations special issue rather than those from the book itself. Doing so would be extraordinarily cumbersome and would seem to disincline one to get one’s hands on the book itself, when the PE special issue is readily accessible. For this reason we have chosen not to abide by the request (as we suspect others will too), and all page numbers refer to the volume reviewed, unless otherwise indicated. (It is not clear whether the request comes from the editor, from Routledge, or from Taylor and Francis as PE‘s publisher.)

After the introduction, the book comprises eleven chapters, split broadly into three sections (Introduction, pp. 4-6), Action in Early Modern Philosophy (Chs. 1-5), Action in Late Modern Philosophy (Chs. 6-8), and Action in Twentieth-Century Philosophy (Chs. 9-11). For reasons of space, we inevitably cannot consider each chapter in detail, but we do try to say something about each, and in some cases engage in more detail. In addition, we also highlight links between various of the issues raised by the chapters and the contemporary philosophy of action literature (with which we, as non-historians, are primarily familiar).

In Chapter 1 – “Agents, objects and their powers in Suarez and Hobbes” – Thomas Pink considers two rather different conceptions of agential powers, from Suarez and Hobbes. To simplify, Suarez thought of humans as capable of “determining outcomes for [themselves]” (25), that is, without having been determined (by their circumstances and capacities) to determine those outcomes. Whilst not determined to act (to determine outcomes for ourselves), we are in acting often moved (motivated) by the normative power of an argument or the goodness of the goal or outcome of the action (19), with reasons and goals thus acting as a final cause. Hobbes rejects such a view as “incoherent” (25), complaining that “Moved not by an efficient, is nonsense.” (18, quoting Hobbes in Hobbes and Bramhall 1656, 59). For Hobbes, all power is efficient causal power, and motivated action is action caused by aspects of the agent’s psychology – specifically, by episodes of imagining the enjoyment of an object along with a way to get it. Pink suggests that Hobbes cannot consistently hold that in rational action, one is motivated – moved – by the normative force of reasons, something Suarez builds into the heart of his account. Interestingly, as Pink notes towards the end of the paper, Hume also avoids Hobbes’ problem. But whereas Suarez avoids it by allowing a distinctively normative form of power, Hume avoids it – much more radically, thinks Pink – by denying the apparent datum that that the objects of motivation can act as “conduits of a force of justification of reason” (27), and ultimately by denying the reality of “practical reason itself” (ibid.).

Pink closes by noting that contemporary causalists about action tend to side with Hobbes over Hume, in “describe[ing] motivation and action as governed by standards of reason by which we are supposed somehow to be moved” (ibid), and in so doing, inherit the inconsistency identified for Hobbes. Contemporary causal conceptions of action done for reasons are very widely termed ‘Humean’. If Pink is correct, then perhaps they should be thought of as ‘Hobbesian’ instead. Whatever one thinks about the labelling issue, Pink’s paper raises a challenge which faces contemporary causal theorists about action just as much as it faced Hobbes: can we understand responding to the normative force of reasons in terms of having one’s behavior caused by a combination of conative and cognitive mental states? Many of the most central debates in contemporary philosophy of action in their own way stem from different views on this question: whether reasons are facts or mental states; the problem of causal deviance; whether action-explanation is efficient-causal explanation; whether teleological explanation can be naturalized; and so on.

In Chapter 2,”Human action and virtue in Descartes and Spinoza”, Noa Naaman-Zauderer argues that despite obvious and widely acknowledged differences (e.g. in their “pictures of reality and its laws [and of] human beings and their place in the world” (41)), there are nevertheless “striking similarities” between Descartes’ and Spinoza’s accounts of human action. In particular, both held a) an account of virtue in terms of activity or freedom, understood as the actual power of acting, b) a non-consequentialist approach to virtue, and c) an account of virtue qua free action as the summum bonum. We would have liked to hear more about (b), and wondered whether calling Descartes’ and Spinoza’s accounts of virtue ‘non-consequentialist’ might in fact understate how much they contrast with consequentialist theories. Descartes, for example, is shown to hold that one acts virtuously as long as one is motivated by what one judges to be best, then one has done one’s duty, “even though [the] judgment may perhaps be a very bad one” (38). Spinoza holds that virtue is “identical with activity or freedom” (40). These views allow that virtue might be completely independent of how good or bad the outcomes of one’s actions are, but ‘non-consequentialism’ per se includes views on which whether one acts virtuously might very well depend in part on what one actually does or brings about (it is not clear that Kant, for example, is committed to the claim that being motivated in a certain way suffices for doing one’s duty). So we wondered whether the similarities between Descartes and Spinoza could be drawn even more tightly.

In Chapter 3, Tom Stoneham discusses “Action, knowledge and embodiment in Berkeley and Locke”. Stoneham considers the prospects for a ‘realist’, ‘causal-volitionist’ reading of Berkeley’s philosophy of action, on which (intentionally) φ-ing is a matter of willing, and thereby causing, bodily movements of the type φ-ing (53-4). As stated, the account is inadequate because not all causings-by-volitions are genuine cases of intentional action (the argument here (53-54) is hard to follow, but the problem for such views is familiar from contemporary action-theoretical discussions of causal deviance). For help on Berkeley’s behalf, Stoneham turns to Locke, who argued for an epistemic condition on volition and intentional agency. Stoneham describes a view on which someone φ’s intentionally when their mind exerts its power in moving parts of their body, on the condition that they know that they are exerting this power in doing so (55). Could Berkeley have accepted this view? Stoneham argues that Berkeley’s famous ‘denial of blind agency’ can be read as accepting just such an epistemic condition. But the suggestion is not without interpretative hurdles, and Stoneham concedes (for reasons we will come to) that it is not straightforward to square it with Berkeley’s empiricism – although ultimately he declines to commit himself on the interpretative question.

The issues discussed in this paper tie in to contemporary debates about inter alia how to understand ‘practical knowledge’ (knowledge of what one is intentionally doing), and in particular to a debate about whether an agent’s having practical knowledge is essential to her action’s being intentional. Simplifying hugely, the contemporary literature on the nature of intentional action is split between those (following Davidson) who define intentional action as bodily behavior with certain mental causes, and those (following Anscombe), who view it as action which is the object of the agent’s practical knowledge. Stoneham can be understood as suggesting on Berkeley’s behalf a hybrid account, on which intentional action is behavior with certain mental causes, and which is known as such by the agent. In contemporary philosophy this hybrid view is rarely held, although there are particularly prominent exceptions in the work of David Velleman (esp. 1989, 2000), and Kieran Setiya (2017). Given this clear interest to non-Berkeley experts (and given that questions about action in Berkeley’s philosophy are rather generally fascinating), it is a shame that the paper does not make itself very accessible to those new to Berkeley’s philosophy in general, or his philosophy of action in particular. Important positions, including Berkeley’s general framework, are assumed without explicit introduction or explanation, meaning that an uninitiated reader may be left without a clear sense of the particular problem which action poses for Berkeley (the interested reader should start with A. D. Woozley’s seminal ‘Berkeley on Action’ (1985)).

We would also like to raise a more substantive philosophical issue. As Stoneham emphasises, Berkeley denies the possibility of ‘blind agency’, indeed (as noted) using this denial to argue that Berkeley accepted an epistemic condition on intentional action. But Berkeley’s empiricism constrains the kind of knowledge he can view as involved in intentional action. According to Stoneham’s Berkeleyan hybrid account, (e.g.) raising one’s arm intentionally requires knowing that one’s arm is going up as a result of an effective volition to raise one’s arm. From an empiricist perspective, this causal link must be known through experience – of previous instances of willing to raise one’s arm, followed by one’s arm going up. But this means that some willed arm-raisings must be unknown, viz., those which pre-date one’s knowledge of the causal link between such willings and such arm-movements. Stoneham accepts all of this, conceding that accepting his hybrid account as a reading of Berkeley would require thinking of Berkeley’s denial of blind agency as restricted to intentional agency rather than agency per se. The baby’s mind, on this view, causes its arm to go up by forming volitions, but while it is still blind to this power, the baby itself is not an agent. Causality-by-the-mind can be blind, and we can call this ‘blind agency’ if we like, but agency proper is never blind: the baby’s mind becomes an intentional agent rather than merely a cause only when it comes to know about its status as a cause. But even if this were a credible option for interpreting Berkeley, we are not sure it sufficiently relieves the tension between a knowledge condition on intentional action and empiricism. For Stoneham is clear that the knowledge condition is supposed to constrain not onlyacting on a volition, but also forming one (54; 55; 56). If this is right, then there is simply no room for babies’ minds to have volitions prior to meeting a Lockean epistemic condition, so there is no causal link between volitions and bodily movements for a Berkeleyan baby to come to know about, on its journey from (blind) cause to (knowing) agent. Whether it is open (interpretatively or philosophically) to Stoneham to reject the epistemic condition on the formation of volitions, and to hold it only for their execution, we leave, for reasons of space, as a question for further reflection.

Chapter 4 is Chris Meyns’ “Sympathetic action in the seventeenth century: human and natural”. Meyns shows how sympathy, these days a purely psychological notion, was understood much more broadly by certain thinkers in the Early Modern period. On this understanding, sympathy is a rather general kind of explanatory principle, of which explanations of human behavior is simply a species, but which also explains relationships between other (non-psychological) natural phenomena. Meyns describes different versions of this broader understanding of sympathy, due to Margaret Cavendish, Anne Conway, and Gottfried Leibniz, as well as detailing Early Modern criticisms of this use of the notion, as ‘occult’, ‘magical’, and ultimately unexplanatory (from Bacon, Descartes, Gassendi, and Malbranche; 73-74). Meyns argues that this skepticism about anything other than a restricted purely psychological conception of sympathy rests on an underlying commitment to an atomistic conception of explanation, and a rejection of a more holistic conception, on which non-psychological conceptions of sympathy do, they argue, “make sense” (77). Thus, they conclude that “[a] core disagreement about sympathetic action has to do with different approaches to explanation” in the Early Modern period (78).

As non-Early Modernists, we found the more expansive conceptions of sympathy which Meyns discusses intriguing, but wanted to know more about why we should think that there really is a single concept or phenomenon here, which is given different glosses by the different writers Meyns considers. Indeed there seems to be so much disagreement between the various (putative) phenomena called ‘sympathy’ by the various thinkers discussed (including differences in what a notion of sympathy is supposed to explain as well as how it is supposed to explain it) that we found it hard to get a grip on whether or not there is any genuine disagreement between these thinkers about sympathy. Most obviously, the psychologistic conception held in contemporary discourse, but also attributed to Smith, seems a world away from a notion of sympathy as some kind of explanatory relationship between changes in one bit of (animate) matter and changes in another. The latter put us in mind of the contemporary notion of quantum entanglement, which nobody in contemporary thought would suggest might be a different way of understanding the same phenomenon as a friend’s fellow-feeling for a friend. What makes ‘sympathy’ à la Leibniz (for example) any more of a genuine rival to the Smithian or contemporary psychological notion of sympathy than the notion of quantum entanglement might be? Even amongst the friends of a non-psychologistic conception of sympathy, the differences between the various ‘sympathies’ are stark. Cavendish thinks sympathy links bits of (animate) matter; Conway that it links substantially similar creatures; Leibniz that it links everything. Cavendish and Conway think of it as explaining the occurrence of specific happenings/actions, whereas Leibniz seems to view it transcendentally, as an underlying metaphysical precondition for action or causation. Are there reasons for thinking that the various notions of sympathy which Menys considers are related any more than etymologically? Of course answering this question might require a rather general investigation of when conceptual development gives way to conceptual speciation (so to speak), which in fairness to Meyns, goes far beyond their concerns in this interesting chapter.

The final chapter in the section on Action in Early Modern Philosophy is Elizabeth S. Radcliffe and Richard McCarty’s “Hume’s Better Argument for Motivational Skepticism” (Ch. 5). Motivational Skepticism is the view that reason cannot itself motivate action; that motivation requires desire or passion. Hume is standardly taken to argue for motivational skepticism by first claiming that reason has only two modes of operation – ‘relations of ideas’ and ‘probabilistic’ – and then arguing that neither of these could generate motives. A number of philosophers in the recent literature have claimed that this argument begs the question, ruling out motivation by reason by fiat, by adopting an overly restrictive definition of ‘reason’. Radcliffe and McCarty (R&M) argue that such objections fail to undermine Hume’s motivational skepticism, because this can also be reached from a different and less problematic starting-point, from the principle that nothing can oppose a passion or desire except an opposing impulse, and the claim that reason cannot provide opposing impulses. R&M defend the latter by arguing that impulses must be capable of being stronger or weaker in order to be capable of opposing one another, and that reason – and beliefs, as the products of reason – cannot be the source of such scalar impulses.

This chapter is one of the collection’s strongest. The appeal to scalarity is an innovative addition to debates about desire and the motivating force of reason, and has the prospect of expanding those debates beyond mere disagreement about imagined cases. However, we suspect that opponents of motivational skepticism may resist R&M’s claim that the products of reason lack the requisite scalarity. R&M consider the objection that the relevant scalarity might come from evaluative beliefs with scalar contents, ones which represent their objects as more or less good, or good to a certain degree, etc. R&M argue that such beliefs would only provide scalar impulses if we assume an implausibly strong connection between evaluative judgement and motivation, i.e. that necessarily, evaluative judgements never fail to motivate. A weaker and more plausible view about evaluative judgements, such as that evaluative judgements necessarily motivate those who are rational, would leave the way in which reason motivates arbitrary, by allowing that one’s evaluative judgements might prescribe motivation to a certain degree whilst one may in fact be motivated to some other degree. Thinking of reason as ‘arbitrary’ in this way, argue R&M, is not coherent. It is this last step where we suspect the opponent of motivational skepticism might resist. It’s not clear that it would be reason which is arbitrary when it fails to motivate to the degree it prescribes. In such a case we needn’t think of reason as telling one to be motivated less than it prescribes, which would indeed sound strange and contradictory. Rather we could hold that in such cases, extra-rational factors (e.g. weakness of the will), are interfering with or undermining the motivating force of one’s reason.

The section on Action in Late Modern Philosophy begins with a paper by Arto Laitinen, Erasmus Mayr and Constantine Sandis (LMS) on “Kant and Hegel on purposive action” (Ch. 6). LMS investigate how the contrasts between Kant’s and Hegel’s action theory result in differences between their practical philosophies more generally, concluding that “the category of action is far more central to both Kant’s and Hegel’s moral theories than standard interpretations suggest” (109). Broadly, Kant’s and Hegel’s ethical thought is claimed by LMS to diverge on key questions in philosophy of action. Firstly, Kant claims that the distinction between the internal and the external aspects of action is of key importance, with the moral law only governing the former. Hegel on the hand gives importance to “the capacity to identify with the results of one’s actions in the external reality” (105). LMS claim this difference in Kant and Hegel’s action theories is importantly related to other differences in their ethics. Firstly, it underwrites a difference in their understandings of freedom, and in particular on whether external influence must be considered heteronymous and therefore to undermine autonomy. On the one hand, Kant thinks that any kind of external influence on the will undermines freedom or autonomy. On the other hand, Hegel thinks that part of freedom is working with ones’ external influences; one aspect of freedom Hegel discusses is social freedom, i.e. freedom that we have through our relationship to social and ethical institutions. Secondly, it relates to a disagreement between Kant and Hegel on the scope of moral responsibility. Hegel, according to LMS, holds that “we are not only responsible for known characterizations of our actions and their effects but also for those which follow from their “universal” character and which we ought to be aware of, even if we are not” (ibid.). Kant holds that “what we must take responsibility for is more limited” (ibid.). The key claim made in this chapter – that differences in Kant’s and Hegel’s accounts of freedom and responsibility stem from differences in their action theory – is intriguing. In this vein, it would be interesting to compare Kant and Hegel’s disagreements to contemporary debates prefigured by the questions in this chapter – debates about the value of freedom, about cognitive conditions on responsibility, and about moral luck – and to examine whether they rest on similar or different underlying disagreements about the metaphysics of agency.

In Chapter 7, “Action, interaction and inaction: post-Kantian accounts of thinking, willing, and doing in Fichte and Schopenhauer”, Günter Zöller gives an account of how these two post-Kantian figures build on and respond to Kant in their philosophical accounts of action; in particular in relation to Kant’s understanding of the ‘activity’ involved in theoretical cognition. Zöller represents Fichte as unifying the practical and the theoretical in cognition by emphasizing ‘the primacy of the practical’, which stems from the idea of the objects of cognition themselves metaphysically depending on cognitive activity. Schopenhauer, too, finds a fundamental place in his philosophy for activity in the central role he gives to ‘the will’. However, Schopenhauer does not, like Kant, understand the will as self-legislating, autonomous, and making freedom possible, but rather as an a-rational and all-powerful force. Where we gain any genuine freedom, on Schopenhauer’s picture, is not from choosing what we will but from our knowledge of how the will works and is influenced, enabling us to negate it and remain inactive in response to its powers.

The chapter raises questions which also occur at the forefront of certain strands of contemporary work on cognition and agency. Particularly fundamental is an issue which comes up in relation to Fichte’s response to Kant, as presented by Zöller. It is not entirely clear whether Fichte’s defense of the ‘primacy of the practical’ is supposed to be a radical, substantive, and surprising claim, or merely an articulation of something uncontroversial, and agreed on by all sides, namely that thinking and cognition are non-stative. If the latter, the ‘practical’ which is primary is ‘practicality’ only in a very watered-down way, and would not give us any interesting or substantive sense in which we are (essentially) agents in relation to our theoretical cognition. Like Fichte, certain contemporary philosophers of mind in a broadly Kantian tradition often claim that theoretical cognition essentially involves some kind of agency (Burge 1996; Moran 2001; Boyle 2009, 2011), holding that recognizing this will help us understand various psychological and epistemic phenomena, including the ‘special’ nature of self-knowledge; self-ownership and self-constitution; our responsibility for our cognitive attitudes, and so on. Yet it remains a question in these debates as to how substantive the claim to mental (especially doxastic) agency is supposed to be, and so how much work is really done by the idea of cognitive agency. The jury remains out on whether (rational) cognition is essentially ‘active’ in any robust or explanatorily promising sense, but Zöller’s chapter shows Fichte to be of relevance to those interested in these and related issues.

Chapter 8 is Paul Katsafanas’ “Nietzsche’s account of self-conscious agency”. Katsafanas starts by outlining a traditional debate about the role of self-consciousness in action, between those (e.g. Locke and Kant) who hold that self-consciousness enables a distinctive kind of action, and those (e.g. Hobbes, Hume, and Schopenhauer) who claim that it is our passions which primarily determine our actions, with self-consciousness playing a merely subsidiary role. Nietzsche is often thought of as holding the latter view because he, like Hobbes et al, often highlights our ignorance of our actions and their motivations. But Katsafanas rejects this interpretation, arguing that Nietzsche is better understood as “rethinking the terms in which [this traditional] debate is cast” (130). This is because, firstly, Nietzsche’s skepticism about motivations is much more thoroughgoing than that of Hobbes et al, because he does not merely claim we are ignorant about our motivations in particular cases, but that our ordinary thinking fundamentally misunderstands the structure of human motivation. This is particularly because we don’t recognize the influence of various unconscious ‘drives’ on our motivations. This moves beyond the traditional debate because unconscious drives are not merely desires we are unaware of, but motivational forces which operate fundamentally differently from ordinary desires. Drives, for example, aim at continuous expression, shaping our conscious experience and influencing what we value, while, crucially, hiding the fact that they do this, leading us to confabulate and come up with post hoc rationa