Causal Idealism (forthcoming in Idealism: New Essays in Metaphysics, eds. Tyron Goldschmidt and Kenny Pearce, Oxford University Press)
Sara Bernstein According to causal idealism, causation is a product of human mental activity in a world lacking a mind-independent causal relation. The view is historically popular. Collingwood (1938) held that causal questions are essentially epistemological rather than ontological. Russell (1913) held that there was no role for causation in fundamental physics. And insofar as Hume famously took causation to be a constant conjunction of events rather than a relation, he, too, counts as a kind of idealist.1 Causal idealism has fallen out of favor in contemporary metaphysics. The majority of current causation theorists claim to be causal realists. They reject causal idealism in favor of “mixed” theories that accept a mind-independent causal relation whose truth conditions are infected by, but not entirely a creation of, human thought. In recent years, views positing contextualism, contrastivism, and pragmatism about causation have surged in prominence. Such views hold that whether c is a cause of e is subject to a particular context, contrast class, or pragmatic constraints on terminological usage. These views purport to solve a variety of problems for theories of causation, including problems with preemption, causation by omission, and the distinction between causes and background conditions. The central suspicion surrounding these types of views is that they render the causal relation subjective and mind-dependent while claiming to be realist. Metaphysical realists hold that causation is a joint-carving, mind-independent relation: something “in the world” that would exist exactly the way it does regardless of human thoughts or interventions. But views of causation that incorporate human thought and agency are taken to threaten the mind-independence and objectivity of the causal relation. This paper asks the question: why not accept causal idealism in lieu of contextualism, contrastivism, or pragmatism? Accepting idealism is commonly viewed as “giving up” on a mind-independent causal relation—a lazy way out of intractable problems with causal theories that seem solvable by partially incorporating human 1 See Beebee (2007) for an argument that this is how we should read Hume.
thought and agency. I argue that weighing contextualism, contrastivism, or pragmatism about causation against causal idealism results in at least a tie with respect to the virtues of these theories. A quick disclaimer: as a causal realist, I am invested in discovering an objective, mind-independent causal relation. Thus I will not make it my goal in this discussion to defend causal idealism against general objections or its broader critics. Rather, my goal will be to give causal idealism a fair shake by articulating it clearly and fairly, and to explain why the view should be considered a viable alternative to the mixed views I have mentioned. Roadmap: In 1, I articulate the thesis of causal idealism, and apply it to some contemporary problems for causal theories. In 2, I give an overview of a family of views that incorporate human thought and agency into the causal relation. In 3, I weigh causal idealism against the discussed mixed views, and argue that causal idealism is a viable alternative. 1. Formulating Causal Idealism Causal idealism is the view that causation is a matter of a human mental activity. For the causal idealist, a complete physical theory of the world does not include an objective, mind-independent causal relation. There are no “little arrows” uniting events or properties to be found at the bottom level of reality; there is no objective relation instanced in the vast spread of events. For the idealist, talk of causation represents a collective human projection of the causal concept. Consider “Domino A’s fall caused Domino B’s fall” and “The presence of oxygen caused the forest fire.” For the idealist, there need be no common objective feature of the world, a mind-independent instance of the causal relation, in virtue of which these two claims are true. In each case, the term “cause” picks out a mental projection on the part of an observer or observers. More formally: (Causal Idealism) c is a cause of e only if at least one observer mentally projects a causal relation between c and e.
One might hold a more Berkeleyian form of idealism according to which causation need only exist in the mind of an agent rather than be mentally projected onto the world. But for our purposes, it will not matter. The central idea, that causation is entirely a matter of human mental activity, is well captured in the above formulation. It might seem strange to formalize this kind of definition. But as we shall see, it doesn’t differ greatly from the formal definitions of other sorts of causal theories involving human agency. Idealism about causation is not to be confused with eliminativism or error theory about causation. Eliminativism formally denies the existence of any causal relation, including one that stems from human thought. Error theory holds that causal assertions ascribe causal relations to the world that don’t exist. But to the idealist, causation exists in the sense that it is mentally projected onto the world—something akin to a secondary quality.2 Idealism does not necessarily deny the existence of causation; rather, it locates the existence of causation in human thought rather than in human-independent reality. But it is “real” insofar as it occurs inside human minds. Causal claims can be the subject of true utterances under idealism. Idealism, like many theories, must deny the correspondence theory of truth, according to which truthmakers for causal claims are objective and mind-independent. But the idealist can hold that causal claims are true in virtue of a thinker or thinkers’ causal projection—not much different than the truth conditions for causal contextualists and pragmatists. Causal idealism nicely solves many problems for contemporary theories of causation. Consider a well-known problem case for transitivity: (Bomb) Jane places a bomb under Joe’s desk. Joe finds the bomb and defuses it, leading to his survival. If transitivity of causation holds, then Jane’s placing the bomb under Joe’s desk caused his survival. But an observer would not project a causal relation between Joe’s placement of the bomb and Jane’s survival. So the counterintuitive transitivity judgment is not a problem. 2 See Menzies and Price (1993) for an argument that causation is a secondary quality.
There is a similar solution to the problem of causation by omission, which afflicts theories that require a transfer of energy from a cause to an effect. Suppose that Jane’s failure to turn off the faucet caused the sink to flood. There is no transfer of energy from Jane’s failure. Idealism has a ready answer: Jane’s failure to turn off the sink caused the sink to flood because it was the subject of a human-projected causal relation. Similarly with the problem of profligate omissions, according to which counterintuitively many omissions (including, for example, Barack Obama’s failure to turn off the sink) count as a cause of an outcome due to simple counterfactual dependence. Human projection naturally divides salient from nonsalient omissions. It might seem flatfooted to solve sophisticated metaphysical problems by stipulating a token causal relation for every observer’s causal projection. Idealism is a radical and wide-ranging solution, but it is a crude one. Maslow famously held that “if all you have is a hammer, everything looks like a nail.” Idealism is a hammer. For every problem in which a key property of causation such as transitivity generates counterintuitive results, the absence or presence of human projection of causation simply overrides the troubling property. And for causal claims that do not seem to cleanly map onto the world (as in causation by omission), the idealist holds that it is the operation of the human mind that creates causation in the first place. But as I shall now suggest, using idealism is not so different than using context, contrasts, or conversational pragmatics to determine when there is causation. These views use sophisticated machinery, but end up on par with the much simpler and more radical idealism just described. 2. Contemporary causal “isms”: contextualism, contrastivism, and pragmatism Here I will summarize three current views of causation that incorporate human thought and agency: contextualism, contrastivism, and pragmatism about causation. It is important to distinguish between these views because they weigh in differently on the mind-independence of the causal relation and the truth conditions for causal claims. When referring to the views en masse, I will refer to them as “the h-isms”, where the “h” stands for human thought and agency. Given space constraints, I will not delve into the nitty-gritty details of the views. Overviews will suffice.
According to the contextualist, a single pattern of objective, in-the-world causal relations can give rise to different truth conditions for a single causal judgment. Thus the contextualist can believe in mind-independent causal relations while stipulating that the truth conditions for causal claims are contextual—a kind of compromise position between mind-dependence and mind-independence of the causal relation. Additionally, according to Schaffer, context sensitivity encompasses: “causal selection (as illustrated by whether or not the presence of oxygen is said to cause there to be a forest fire) causal inquiry (as illustrated by the different causal answers appropriate for the questions of why John kissed Mary, why John kissed Mary, and why John kissed Mary) multiple alternatives (as illustrated by the train switch with the broken, local, and express settings).” (Schaffer 2012) Consider several examples of causal claims whose truth conditions are fixed by context: (a) Jane’s failure to water the plant caused the plant to die. (b) The storm caused the building to collapse. (c) The presence of radioactivity set off the Geiger counter. In claim (a), contextual parameters operate on omissive causal claims by selecting truth conditions for which particular omission brought about an outcome. For example, Barack Obama also failed to water the plant, but “Barack Obama’s failure to water the plant caused the plant to die” is contextually false. For claims (b) and (c), contextual parameters set truth conditions by restricting the contextually appropriate causal field for each outcome. For example, meteorologists examining the effects of the storm would accept the truth of (b), whereas engineers investigating the architectural properties of the building would hold that construction errors caused the building’s collapse. Similarly, scientists measuring radioactivity would accept the truth of (c), whereas engineers of the device would take the device’s working properly to be the cause of the measurement. In each instance, context partially defines the truth conditions for causal claims. A closely related view, contrastivism, holds that the causal relation is a quaternary relation between a cause, an effect, and contextually specified contrast classes for the
cause and the effect. Schaffer’s (2005) strain of contrastivism holds that the contextsensitivity of contrast classes is due to semantic, rather than entirely pragmatic, constraints. For Schaffer, there is an underlying contrastive structure to causal language that is made explicit in this quaternary relation. There are two actual mind-independent events, c and e, paired with contrast events c* and e*. (The contrast events are unactualized.) The selection of contrast events for use in a causal claim is contextsensitive; different contrasts can result in different truth conditions. Contrastive causal claims have the structure: “c rather than c* causes e rather than e*”, where c* and e* are contrasts with the actual cause and effect c and e. Formally: (Contrastivism) c rather than c* causes e rather than e* if: (i) c and e are actual, distinct events, and (ii) c* and e* are alternative possible events, (iii) e* counterfactually depends on c*. With respect to (a), for example, the contrastivist would hold that “Jane’s failure to water the plant (rather than her watering the plant) caused the plant’s death (rather than its continued existence).” The contrastivist accepts a similar read of (b) and (c): “The storm (rather than clear weather) caused the building’s collapse (rather than its survival)” and “The presence of radioactivity (rather than its nonpresence) caused the Geiger counter to sound (rather than to remain silent).” Finally, pragmatism about causation holds that causal claims are sensitive to pragmatic factors, including practical aims of the topic in question. Price (2005) posits pragmatism about the “causal practice”, according to which the causal concept depends on the particular needs of a theoretical project.3 Like contextualism, this view is compatible with a kind of realism about causation according to which a single in-theworld causal instance can have different pragmatic interpretations. Price writes: “In my view, an adequate philosophical account of causation needs to begin with its role in the lives of agents, creatures who have the primitive experience of intervening in the world in pursuit of their ends. […] Here, I simply want to emphasise that the view is not the topic-subjective claim that talk of causation is talk about agents or agency, but rather the practice-subjective doctrine that we 3 Here I have in mind Price’s pragmatism, but see also Eagle (2007).
don’t understand the notion of causation—as philosophers, as it were—until we understand its origins in the lives and practice of agents such as ourselves.” For the pragmatist, causal claims are subject to constraints imposed by theoretical practices and endeavors. Moreover, the causal concept itself is bound up with manipulability by agents. As Menzies and Price put it: “[Our] approach makes the notion of causation an extrinsic one, to be explained by relation to our experience as agents: on our probabilistic version of the agency approach, the concept of causation is to be explained in terms of the way in which an agent's producing, manipulating or 'wiggling' one event affects the probability of another event.” (Menzies and Price 1993, p. 9) and: “…We all have direct personal experience of doing one thing and thence achieving another. We might say that the notion of causation thus arises not, as Hume has it, from our experience of mere succession; but rather from our experience of success: success in the ordinary business of achieving our ends by acting in one way rather than another.” (Ibid) Drawing on statements of the view, we can formulate pragmatism in the following way: (Pragmatism) C is a cause of e if pragmatic considerations establish that c is a cause of e. Here, both the causal concept and the causal relation are relativized to human thought and practice. The explanatory desideratum (science, law, prediction) determines the causal relation and its truth conditions. Price, pragmatism’s contemporary defender, characterizes this approach as turning ontological questions about the causal relation into questions about the human causal concept. “Causation is practice-subjective,” Price claims, “if an adequate philosophical account of causation needs to make central reference to the role of the concept in the lives and practice of creatures who use it.” (Price 2001) Price is careful to note that he does not take pragmatism to equal topicsubjectivity, or a view according to which “talk of causation is in part talk about speakers, agents, or humans.” (ibid) Now that the basic features of the three views have been laid out, it will be helpful
to clarify which elements are objective and mind-independent, and which elements are infected by human thought and intervention. Then we can better weigh them against causal idealism. To begin, consider a token instance of the causal relation expressed by the following claim: “The wind gust caused the boulder to fall over the cliff.” To a causal realist who does not subscribe to context-sensitivity, the relation instanced here is objective, mind-independent, and context-invariant. Call this particular instance of the causal relation BOULDER. And let us assume that, for the realist, the world contains countless instances of the causal relation such as BOULDER—a vast nexus of mindindependent instances of causation. And presumably, the truth conditions for “The wind gust caused the boulder to fall over the cliff” just are the conditions of BOULDER. For the realist, variations in context cannot make the sentence false given the reality of BOULDER. On to contextualism. The contextualist might agree with the description of the causal relation instanced in BOULDER as mind-independent. But she also holds that human-created context sets parameters on the truth conditions for the claim. There is no context-invariant mapping between the human causal concept and the world; however the objective causal relations are arranged, it is the context of inquiry that determines whether or not a causal claim is true. “The wind gust caused the boulder to fall over the cliff” might be true in a metereological context but false in a broader historical context. Like contextualism, contrastivism aspires to respect the mind-independence of causation while retaining contextual flexibility. According to the contrastivist, “The wind blowing (rather than not blowing) caused the boulder to fall off the cliff (rather than to remain still)” is true regardless of human-created context, as are countless other related contrastive claims. “Causation is mind independent,” Schaffer writes,” [and] “mere redescription cannot stem the causal tide.” (Schaffer 2005) Humans contextually privilege a particular contrast class, contrastivists claim, but do not thereby bring the causation into being. However, it is a controversial matter exactly how objective and mind-independent contrastive causation is. On the one hand, contrastive causation is mind-independent insofar as the quaternary relation holds between the actual and contrast events regardless
of human thought or agency. For example, “The wind blowing (rather than calm air) caused the boulder to fall over the cliff (rather than remain still)” is true regardless of human thought or intervention. And counterfactual dependencies are mind-independent. On the other hand, it is unclear that there is a difference between raising contrasts to salience and selecting the causal relata full stop. With respect to selections of contrasts, Schaffer holds the contrastivist should: “(i) deny that there is any objective basis for selection independent of the contrasts but (ii) maintain that there is an objective basis given the contrasts.” (ibid) Humans raise certain contrasts to salience, Schaffer maintains, but humans do not choose the causal relata. And yet, two of the four causal relata are included in each causal relation in virtue of a human raising them to salience. The four-place relation does not exist without a human to choose which contrasts to include for the actual-world events. Thus there is a case to be made that humans choose the causal relata simpliciter under contrastivism, and that it is not as mind-independent as it claims to be. Pragmatism, in contrast, does not lay claim to objectivity or mind-independence. Like causal idealism, pragmatist causation does not exist without the perspective of human agents. According to Price: “…[To] to treat causal concepts as perspectival, in this naturalistic spirit, is inevitably to theorise about the non-causal world to which we apply these concepts—the bare Humean world, in the midst of which we embedded creatures come to think in causal ways. I’ve already noted the Kantian character of this project. In these terms, the bare Humean world plays the role of the non-causal ‘thing in itself’.” (Price 2005) Price sees the causal concept as similar to “up” and “down”—an essentially human construct. It is worth noting that Price denies that his sort of pragmatism is anti-realist in our relevant sense. He writes: “There are two different ways to discover that something isn’t as real as we thought. There’s the kind of discovery we made with respect to phlogiston, ether, unicorns, leprechauns, and the like; and there’s the kind of discovery we made with respect to foreigners. We didn’t discover that foreigners don’t exist, but
merely that the concept foreigner is perspectival. And that’s where causation belongs, in my view, along with folk favourites such as up and down, night and day, and the rising sun itself.” (ibid) Yet it seems clear from Price’s various characterizations of the view that it is not compatible with an objective, mind-independent relation any more than the concepts “up”, “down”, or “foreigner”: if the causal concept is a matter of perspective, it doesn’t exist without human mental activity. Pragmatism is a close cousin of idealism. From these formulations, we can draw the following chart. An “x” denotes that a theory has the relevant property: Mind-independent causal relation?
Mind-independent causal relata?
Pragmatism Idealism Realism
This chart nicely illustrates the similarities and differences in ontological commitment shared by these theories. As I will argue, even the h-isms that accept a mind-independent causal relation should be considered on par with causal idealism. 3.0 H-isms versus Idealism: Two Case Studies In this section I will explore how each h-ism handles two well-known problems for theories of causation: causation by omission, and the distinction between background conditions and causes. I will then compare them to causal idealism. For our purposes, the most relevant problem involving causation by omission is selective: why does one omission rather than another count as the cause of the effect? Reconsider Jane’s failure to water the plant as the cause of the plant’s death. What makes that failure to water the plant, rather than countless others, the cause of the plant’s death? There is counterfactual dependence between Barack Obama’s failure to water the plant and the plant’s death, but it seems counterintuitive to hold him accountable for the plant’s death.
Contextualists have a ready explanation: context selects Jane’s particular omission over others (e.g. Barack Obama’s omission, Abraham Lincoln’s omission, etc.) as the relevant cause of the outcome. Even if many people and things failed to water the plant, norms of assertion governing the truth conditions for the omissive claim generally favor Jane’s failure. (We can imagine outré contexts in which Abraham Lincoln’s omission to water the plant is the cause of the plant’s death; for example, a case where Abe promised to time travel in order to water the plant but failed to do so. See Bernstein (2014) for an exploration of this issue.) Does contrastivism have an explanation? According to contrastivism, Jane’s omission to water the plant (rather than her watering the plant) causes the plant to die (rather than to live). But why is Jane responsible and not others? Schaffer (2010) holds a view according to which an omission is just a way to describe actual events upon which the relevant contrasts are based. For example, if Jane was singing karaoke rather than watering the plant, the relevant contrastive claim is “Jane’s singing karaoke (rather than watering the plant)” is the relevant causal contrast. But this doesn’t solve the problem of why Jane is the pertinent party, as Schaffer recognizes. He handles the problem of profligate omissions by arguing that normative expectations on Jane rather than Barack Obama make Jane’s actions causally germane. But imposing this normative constraint on the selection of relevant actual events thereby introduces additional human-dependence into the causal relation. Similarly, pragmatism dictates that normative or practical constraints on omissive claims determine which omission is a cause of the outcome. In a moral context, Jane’s promise to water the plant is the subject of practical interest, and thus her omission is the cause of the plant’s death. In a scientific context, the plant’s biological processes are the subject of interest, and thus the plant’s failure to deliver oxygen to its roots is the cause of the plant’s death. Generally, it would violate conversational rules to mention Jane’s omission in the scientific context or the plant’s biological processes in the moral context. All three h-isms can account for causation by omission. But in each case, humancreated contexts select which omission out of many is causally salient. Even for the theories that are compatible with a mind-independent causal relation, context, semantics, and pragmatics are doing the heavy lifting of solving the selection problem. Consider
contextualism and contrastivism, both of which posit a mind-independent causal relation. These theories could deny an objective causal relation and achieve the same results for causation by omission. This is yet more evidence that the seat of explanatory power in both theories lies in the human-injected elements rather than the causal relation that both posit. Another well-known challenge for causal theories is how to distinguish between background conditions and causes. The birth of the universe is a causal condition of Abe Lincoln’s death, but few people would call the former a cause of the latter. In a less extreme problem case, it might be hard to judge whether the engineer’s error was a cause or a mere background condition of the building’s collapse. The three h-isms are well suited to this problem as well. According to contextualism, the engineer’s error was the cause of the building’s collapse because context selects it over the background conditions. The contrast events in “The engineer’s error (rather than the building’s construction) caused the building to collapse (rather than stay standing)” are selected by context. And pragmatics dictate practice-relative interests, which restrict relevant causal candidates and cull background conditions. But here again, the objective causal structure of the world does not shoulder the explanatory load. Humans label one part of reality as a cause and another as a background condition: reality itself does not “do the work”. Reality could have been different, and yet humans would have selected the relevant contextual parameters, contrast events, or pragmatic constraints on what counts as the cause. For this reason, I will now argue, causal idealism should be considered on a par with the three h-isms. 4.0 Weighing the Views Having illustrated the views and studied how they deal with problems for theories of causation, we are now in a position to weigh causal idealism against the three h-isms outlined above. We will compare them with respect to theoretical desiderata for a theory of causation: explanation, parsimony, metaphysical fundamentality, and suitability to ground moral responsibility. Return to BOULDER. For the proponent of an objective causal relation, what explains the boulder’s falling over the cliff just is the reality of BOULDER. To the
contextualist, human-created context explains why the wind (rather than, say, the bear scraping around nearby destabilizing the ground beneath the boulder) causes the boulder to fall. The contextualist might retort that she does accept the mind-independence of causal relation; it is simply the truth conditions that are context sensitive. For example, she agrees with the objectivity and mind-independence of BOULDER, but holds that context determines when utterances involving BOULDER are true. Here I say: for the contextualist, the ontological realism about causation isn’t doing the heavy explanatory work. Context, rather than an objective causal relation, is the seat of explanatory force. If an aim of a causal theory is to give the conditions for true causal assertions, then locating the truth conditions for the causal assertions in thought rather than reality thereby relocates the theory’s explanatory power in human thought. We cannot understand why the boulder went over the cliff without appealing to a particular context. Even if the contextualist’s causal relation is mind-independent, causal explanations that require contextual parameters are irreducibly human-based. Similarly with contrastivism. Contrastivism’s flexibility lies in its ability to draw out causal truths relative to salient contrasts. But it is the human intervention of raising the contrasts to salience that achieves much of the explanatory power. Consider the following two contrasts: (1) The wind blowing (rather than not blowing) caused the boulder to fall over the cliff (rather than to remain still). (2) The wind blowing (rather than gusting) caused the boulder to fall over the cliff (rather than gently roll off the cliff.) Here the difference is illuminating due to the contextually salient contrasts, not the objective counterfactual dependencies that hold between the actual events and between contrasts. Contrastivism’s informativeness is driven by human selection, not by an objective causal relation. This is true for pragmatism as well. Unlike the other two h-isms, pragmatism doesn’t aspire to mind-independence. Pragmatism accepts that human interests drive the explanatory power of the causal relation. “The wind blowing caused the boulder to fall
over the cliff” is true in virtue of human concepts and practical application of the causal concept. Were humans to vanish, so, too, would the causal relation. I submit that idealism’s human-driven explanatory power is at least as powerful as those of the h-isms. If the explanatory power of h-ist theories lies in their humandependent elements, idealism has the same explanatory power without resorting to complicated theoretical machinery. For idealism doesn’t need context, contrasts, or pragmatic constraints. Idealism does not require the linguistic model of contextualism, the modal structure of contrastivism, or the collective perspectives or practical interests of pragmatism. Idealism requires only an observer or observers. Moreover, explanation “bottoms out” in humans in all four views: the buck stops, as it were, with human mental activity. In contrast to causal realism, according to which causation itself grounds explanation, causation itself does not ground explanation in any of our above cases. The h-ist world minus humans is explanatorily impoverished. The hist picture of explanation is at least no better than the idealist’s world in which causal explanations lie entirely with human observers. Idealism is simpler and more parsimonious than contextualism and contrastivism. Aside from the linguistic and modal machineries mentioned above, contextualism and contrastivism each claim to accept a mind-independent causal relation. But the causal relation is explanatorily inert, a shadow waiting for illumination by human thought. The theories are more bloated than they have to be. Idealism, in contrast, posits only the human-based condition for causation without the extra ontological weight. Many metaphysicians want causation be part of the “glue” of the universe—a joint carving, perfectly natural, metaphysically fundamental relation. It goes without saying that such a relation must be entirely mind-independent. Only contextualism posits such a relation, and as we have seen, it doesn’t shoulder the explanatory weight of the view. (This feature, too, makes the contextualist causal relation particularly unsuited to fundamentality, since explanation is supposed to bottom out in fundamentalia.) Contrastive causation cannot be fundamental, since possibilia are included in the causal relata. And pragmatist causation is entirely mind-dependent. While idealism does no better in this regard, it certainly does no worse than the three h-isms with respect to causation not being included in the basic glue of the universe.
There is an exception to the theoretical parity between the idealism and h-isms in the theories’ ability to ground moral responsibility. Assuming that causation is at least necessary for moral responsibility, moral chaos would ensue if moral responsibility were to be grounded in individual humans’ causal projections. Consider “Jane is morally responsible for the plant’s death.” In virtue of what is this the case? A mere idealistic projection of causation would spell trouble for a theory of moral responsibility. For we cannot hold agents accountable for actions on the basis of a mere individual projection of causation. Contextualism comes out ahead because collective human agreement is required for context-setting. Contrastivism comes out ahead insofar as many moral explanations are already contrastive in nature.4 And pragmatism comes out ahead insofar as causation can be understood from a specifically moral or legal perspective. But this disparity is not a defeater for causal idealism. All it shows is that something greater than individual human minds, namely, collective human agreement, is required for a causal concept that grounds moral responsibility. A concept can be socially constructed even if causal idealism is true. Such a project would also comport well with moral views that deny objective moral facts. One might object that causal idealism is not a theory of causation. The objection goes something like this: a theory of causation must contribute something to our understanding of the causal concept and causal relation. It must tell us when the causal relation holds and when it doesn’t. It must deliver results by providing application conditions for the causal relation. When we’re not sure if something is a case of causation, a good theory should be able to yield a result such as “Yes, it is a case of causation because there is counterfactual dependence” or “No, it’s not a case of causation because there is no lawful regularity”. But causal idealism doesn’t provide illuminating application conditions like those theories. To this I say: neither do the h-isms. The h-isms help us understand when human contexts or practices suggest that there is a causal relation. But they don’t provide substantive, non-circular definitions of the causal relation. While idealism doesn’t either, it’s no worse off in this regard. The idealisms and h-ism share a commitment to the inextricable linkage of the causal relation with the human conception of it. 4 See Schaffer (2010) for an extended discussion of contrastive claims in the law.
4.0 Conclusion Causal idealism is simpler than contextualism, pragmatism, and contextualism. It possesses equal explanatory force and greater ontological parsimony. Even though it has fallen out of favor historically, its theoretical virtues should make it a contender among mixed views of causation that incorporate human thought and agency. References Beebee, H. (2007) “Hume on Causation: The Projectivist Interpretation”. In Huw Price & Richard Corry (eds.), Causation, Physics, and the Constitution of Reality: Russell's Republic Revisited Oxford: Oxford University Press. Bernstein, S. (2014) “Omissions as Possibilities”, Philosophical Studies 167:1-23. Collingwood, R.G. (1938) “On the So-called Idea of Causation,” Proceedings of the Aristotelian Society 39: 85-112. Eagle, Antony. (2007) “Pragmatic Causation”. In Huw Price & Richard Corry (eds.), Causation, Physics, and the Constitution of Reality: Russell's Republic Revisited Oxford: Oxford University Press. Menzies, P. and Price, H. (1993) “Causation as a Secondary Quality”. The British Journal for the Philosophy of Science 44: 187-203. Price, H. (2005) “Causal Perspectivalism” in Huw Price and Richard Corry, eds., Causation, Physics and the Constitution of Reality: Russell’s Republic Revisited, Oxford: Oxford University Press. ----- (2001) “Causation in the special sciences: the case for pragmatism.” In Domenico Costantini, Maria Carla Galavotti and Patrick Suppes, eds. Stochastic Causality, CSLI Publications, 103—120. Russell, B. (1913) “On the Notion of Cause”. Proceedings of the Aristotelian Society, vol. 13: pp. 1–26. Schaffer (2012) “Causal Contextualism” in Contrastivism in Philosophy, ed. Blaauw 3563: Routledge. ------ (2010) “Contrastive Causation in the Law” Legal Theory 16: 259–297. ------ (2005) “Contrastive Causation” Philosophical Review 114.3, 327-58.