Abstract
I argue that high level causal relationships are often more fundamental than low level causal relationships. My argument is based on some general principles governing when one causal relationship will metaphysically ground another—a phenomenon I term derivative causation. These principles are in turn based partly on our intuitive judgments concerning derivative causation in a series of representative examples, and partly on some powerful theoretical considerations in their favour. I show how these principles entail that low level causation can derive from high level causation, and in particular that neural causation can derive from mental causation. I then draw out several important consequences of this result. Most immediate among these are the implications the result has for aspirations to reduce high level causation to its low level counterpart. But the result also bears on the possibility of downward causation, the relationship between counterfactuals and causation, and the idea—familiar from both the literature on the exclusion problem and the literature on proportionality constraints on causation—that causal relationships at different levels compete for their existence.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
Following Schaffer (2009), I am permissive about what sorts of things stand in grounding relationships, allowing for facts, relations, relationships, events, properties, laws, particulars and what have you to be both grounds and grounded. [I also follow Schaffer (2009), Rosen (2010) and Audi (2012) in treating ground as a relation, in contrast to those who regiment grounding talk with a sentential operator, like Correia (2010) and Fine (2012a).] If one insists on reserving the term ground for a relation that holds between facts, then the relation I have in mind might be termed metaphysical or ontological dependence (see Rydéhn, 2018). There is every reason for those who are happy to relate facts with ground to be happy to relate all manner of other things with metaphysical dependence (e.g. whole objects with their parts). In particular, it is natural for such philosophers to hold that when the fact that x occurs grounds the fact that y occurs, the event y metaphysically depends on the event x. (For those who think the causal relata are facts, like Mellor [1981] and Bennett [1988], no such retreat from ground to metaphysical dependence is necessary.) And for those sceptical that there are any generic relations of metaphysical dependence, note that the purposes I put grounding to in this paper could equally well be served by a combination of more specific dependence relations, such as the realisation relation, the determinate-determinable relation, the species-genus relation, etc.
Examples of philosophers who talk this way include List & Menzies (2009, p. 477) and Sinnott-Armstrong (2021, p. 868), who speak of high level causation given only a high level cause and an arbitrary physical effect, as well as Gibbs (2014, p. 330) and Hoffmann-Kolss (2014), both of whom treat ‘f is a high level cause of g’ and ‘f caused g is an instance of high level causation’ as interchangeable. This contrasts with some, like Kistler (2017, p. 54), who reserve the moniker high level causation for causal relationships wherein both the cause and effect are high level. Though my terminological choices align with the former group of philosophers, my argument that lower level causation is often built from higher level causation goes through on either understanding of the relevant terms.
In general there is no straightforward connection between the relative levels occupied by some entities and the presence or absence of grounding relationships between those entities. The United Kingdom is higher level than an atom in Andromeda, for instance, but the latter does not ground the former. And a collision between two atoms will be partially grounded in the behaviour of each of those atoms, but it would be a stretch to say that the collision occupies a higher level than does the behaviour of each of the two individual atoms.
I use italics to signal that the italicised phrase denotes a relationship. I will sometimes also refer to a causal relationship f caused g by way of phrases such as ‘f’s causing of g’ and ‘f causing g’.
Unless otherwise stated, every grounding claim I make should be understood as a claim about (strict, factive) partial ground (see Fine, 2012b).
I assume that the causal relata are fine-grained, so that we may distinguish between such intimately related things as my feeling thirsty and the neural event that realises it, qua causal relata. Talking, as I do, as though the causal relata are events thus involves talking as though different events can occupy the same spatiotemporal region. But this is just a way of talking—not a substantive assumption. What I say can be recovered in translation by those who deem the causal relata to be facts, property exemplifications, variable values or what have you, assuming these alternative relata are finely individuated and stand in the sorts of grounding relationships I take events to stand in.
These because claims are only diagnostic of derivative causation when we have something like this sort of building going on. That the striking of the match caused the match to light because the leak caused oxygen to enter the room is no sign of derivative causation in my sense of the term.
An anonymous referee points out that the claim that intermediate links in a causal chain ground the causal relationship between the first and last events in the chain is in tension with the idea that grounding is well-founded in the sense that every grounding chain terminates in one or more ungrounded things. For it is a plausible assumption that, for any f and g where (i) f caused g and (ii) f finishes occurring some nonzero length of time before g begins to occur, there will be some p such that f caused p and p caused g. (Note that this is different from the somewhat less plausible assumption that causation is dense in the sense that for any cause f of g there exists a p such that f caused p and p caused g. Even if we assume that causes must occur entirely prior to their effects, it might still be that no interval of time separates f from g.) If this is right, then f caused g will be grounded in p caused g, which in turn will be grounded in pʹ caused g (where p caused pʹ and pʹ is separated from g by a nonzero interval of time), and so on ad infinitum. But the above assumption is only plausible if time is dense (in the sense that there is always an instant of time between any two other instants), and the density of time is already enough on its own to generate a structurally similar threat to well-foundedness, whatever we think of derivative causation. Take, for instance, the existence of the temporal interval [t0, t100], which is grounded in the existence of the interval [t50, t100], which in turn is grounded in the interval [t75, t100], and so on ad infinitum. In light of this, I see no reason to think that my claims about derivative causation generate any novel threat to the well-foundedness of ground.
At various points in this paper I will use talk of making something happen as a more colloquial stand in for talk of one event causing or metaphysically grounding another. I do this in the present case so as to make manifest the intuitive force of the relevant claim, which would be reduced were we to substitute this colloquial manner of speaking for more technical talk of the grounding relationship running from Socrates’ death to Xanthippe’s widowing.
Why do chains of determination serve to construct causal relationships rather than grounding relationships when the chains comprise a mix of causal and grounding relationships, as in Figs. 3, 4, 5, 6, 7 and 8? One possible explanation is that ground is synchronic and causation diachronic. If this were right it would follow that the events at either end of a mixed determinational chain cannot be grounding-related since any such events are separated by a causal relationship and hence are not simultaneous. Under the same assumptions, chains composed entirely of grounding relationships could not ‘add up’ to causal relationships, hence their absence from our taxonomy, which means to catalogue only those chains that result in derivative causation.
This discussion also serves to reveal the mistake in a more general objection to derivative causation. The objection has it that if causation is counterfactual dependence (or whatever else you like), then positing derivative causation of any sort (whether due to the chaining of determination relationships or not) is idle. Even if, say, f and g are connected by a chain of causal relationships, their causal relatedness consists directly in the counterfactual dependence of g’s occurrence on f’s occurrence—no need to build up the relationship between the two events by way of intermediate relationships. But, as we have just seen, it may well be that the counterfactual dependence between f and g is itself built from other causal relationships/counterfactual dependences, in which case the appearance that the truth conditions for ‘f caused g’ bypass these intervening causal relationships is but an illusion.
I think these modified counterexamples actually reveal that, against the current of existing literature on the topic, which focuses on failures of the transitivity of determination per se, we should instead be concerned with failures of chained determination relationships to build the more encompassing relationships that they succeed in building in normal cases, as failures of transitivity are merely a symptom of this underlying phenomenon.
I allow for the Pi to be nonactual in order to deal with redundant causation. It might, for instance, have been that if Boulder B had never fallen, and so the pathway ⟨Boulder A’s fall, Boulder B’s being knocked off course, the hiker’s survival⟩ had not existed, then Boulder C would have fallen in such a way that, had Boulder A not knocked it off course, then Boulder C would have killed the hiker regardless of whether he ducked. In this case, the transitivity of determination across ⟨Boulder A’s fall, the hiker’s ducking, the hiker’s survival⟩ does not depend on the existence of ⟨Boulder A’s fall, Boulder B’s being knocked off course, the hiker’s survival⟩ alone, but on the existence of at least one of ⟨Boulder A’s fall, Boulder B’s being knocked off course, the hiker’s survival⟩ and ⟨Boulder A’s fall, Boulder C’s being knocked off course, the hiker’s survival⟩, where the latter pathway is nonactual.
In setting things up this way, I am tacitly assuming that if determination would have failed to be transitive along a pathway had some other (possibly nonactual) pathways not existed, then determination would have failed to be transitive along that pathway had some other pathways beginning and ending with the same events not existed. More general setups are available if need be.
An interesting variation on this sort of structure involves a lower0 causing a lower1 and a higher0 causing a higher1, where lower0 grounds higher0 and lower1 grounds higher1. Here Derivation tells us that lower0 caused lower1 grounds lower0 caused higher1 and that higher0 caused higher1 grounds lower0 caused higher1. This seems like the right result. Consider two machines that flash different coloured lights at one another according to the following rules: the second machine will emit a green flash if and only if the first emits a red flash, and will emit a viridian flash if and only if the first emits a scarlet flash (both machines can generate flashes in all manner of different shades of different colours). If on an occasion the first machine flashes scarlet and the second returns a viridian flash, we will have a case with the above structure, as the scarlet flash grounds the red flash which in turn causes the green flash, and the scarlet flash causes the viridian flash which in turn grounds the green flash. Here it seems right to say that the scarlet flash caused the green flash both because the red flash caused the green flash and because the scarlet flash caused the viridian flash, as predicted by Derivation. But this is different from the way in which the causal relationships that result from purely causal chaining are multiply grounded in other causal relationships, as there the causal relationships doing the grounding clearly don’t overdetermine the obtaining of the grounded relationship. Here, by contrast, it is more natural to take the obtaining of the causal relationship between the scarlet flash and the green flash to be overdetermined by the causal relationships that ground it. Given that it doesn’t bear directly on my main arguments, however, I won’t take a stand on this issue one way or the other.
All four of the above results are independent: none entails any other (though the individual cases used to illustrate the results are sometimes instances of multiple results).
Though Derivation has the consequence that higher level causation sometimes derives from lower level causation and the consequence that lower level causation sometimes derives from higher level causation, we need not fear that this will lead to grounding loops, thereby violating the irreflexivity of ground. Derivation only says that p caused q grounds f caused g when p caused q, f caused g, and one of the following conditions is satisfied: (a) f = p and q determines g, (b) g = q and f determines p, or (c) f determines p and q determines g. It follows that Derivation only says that p caused q grounds f caused g when either p or q is an intermediary on a determinational pathway running from f to g. But it is easy to see that whichever of (a), (b) or (c) is satisfied, it cannot also be the case that either f or g is an intermediary on a determinational pathway running from p to q. It follows that if Derivation says that p caused q grounds f caused g, then it does not say that f caused g grounds p caused q. Thanks to an anonymous referee for encouraging me to provide this assurance.
Thanks to an anonymous referee for persuading me to address this issue.
It is hard to think of an analogue of such a pathway in the Martha Stewart case. There, Stewart’s breaking of the law seems to be a bottleneck through which all of the causal influence of her selling those stocks on her going to prison is conveyed.
Even when causal relationships fail to confer subjective reasons they still confer objective reasons. If I believe that caffeine will cause my headache to intensify when in fact it will cause it to abate, then though my subjective reasons for/against consuming caffeine might not be beholden to the actual causal facts, my objective reasons are: I have objective reason to consume caffeine in pursuit of seeing my headache abate.
Note that it is quite possible to think that ⟨n, my feeling thirsty, my drinking⟩ does not exhibit ersatz transitivity for the reasons just given while maintaining that the existence of this high road depends on the existence of at least some low road connecting n and my drinking, such as ⟨n, d, my drinking⟩. Given such a dependence, it might be that supposing away certain possible low-road connections simply entails the absence of any high-road connection, in which case these suppositions will be inconsistent with holding fixed that ⟨n, my feeling thirsty, my drinking⟩ exists. (Over the last few paragraphs I have been effectively assuming that this is not the case.) The counterfactual we use to test for ersatz transitivity will thus have an impossible antecedent which will vacuously counterfactually imply that determination is not transitive across ⟨n, my feeling thirsty, my drinking⟩. But we have stipulated that only counterfactuals with possible antecedents can be indicative of ersatz transitivity, so situations wherein the existence of a high level determinational pathway depends on the existence of a low level one do not threaten my test for ersatz transitivity.
If indeed Kim (1974) was seduced by this tempting mistake, this would seem to be explained by his incautiously generalising from the fact that dependence between causal relationships reflects the dependence between their relata in instances like the following: the jury’s verdict caused Socrates’ death because it caused him to drink hemlock. Here an event in the dependent relationship—Socrates’ death—depends on an event in the independent relationship—his drinking hemlock. Unfortunately for Kim, the same sort of example he generalised from suffices to refute that very generalisation, as shown above.
Kim’s hasty generalisation suggests a diagnosis of our inclination to mistakenly think that the relative fundamentality of relationships always tracks the relative fundamentality of relata: it often does (as was detailed in the first three paragraphs of this section).
This argument that nʹ counterfactually depends on my feeling thirsty may look like it erroneously assumes the transitivity of the counterfactual conditional, but in fact it does not. Since it is true that if neither n nor my feeling thirsty had occurred then nʹ would not have occurred either, all I require is that the counterfactual conditional obey limited transitivity—a principle about conditionals which holds the rare distinction of not being the target of any counterexamples in the literature (Walters, 2016):
(where ‘☐ → ’ means the counterfactual conditional.) Thanks to an anonymous referee for pushing me to clarify this point.
Thanks to Brad Weslake for discussion on this point.
Derivation is neutral on the question of what the fundamental level of causation is (if indeed a single fundamental level is even in the offing). Despite having the consequence that high level causation is often more fundamental than its lower level counterpart, it does not commit us to a ‘middleist’ or ‘topist’ (see Bernstein 2020) stance on causation, according to which middle levels of causation are the most fundamental causation, or the highest levels of causation are the most fundamental causation, respectively.
This is a version of what Kroedel and Schulz dub the ‘Causal Grounding principle’ (2016, p. 1914). Their version requires that the relevant p and g be physical events, though the motivation they cite in favour of this principle is equally motivation for the principle I articulate above. In any case, Derivation also refutes their version of the principle.
Thanks to Cian Dorr for discussion on this point.
Recall that instances of mediated causation are instances of the following schema: f caused g by making p happen. Where, again, f makes g happen just in case f either causes or grounds g.
Claims like the following constitute potential counterexamples to the conjecture: my thirst caused me to consume vitamin C by causing me to drink orange juice. In the right circumstances, such a claim will strike us as true, yet it is far from clear that my drinking orange juice either caused or grounds my consuming vitamin C, and hence far from clear that this claim of mediated causation will satisfy (i)–(ii) of Derivation. I lack the space to give this example the treatment it deserves here, but suffice it to say that claims like the above will be controversial, as their truth seems to require that we are able to explain the occurrence of particular events by appeal to events that neither ground nor cause them. That my thirst caused me to consume vitamin C by causing me to drink orange juice, for instance, seems to entail that I consumed vitamin C because I drank orange juice. And this explanation will be non-causal and non-grounding given the plausible assumption that my drinking orange juice neither grounds nor causes my consuming vitamin C. Given the pedigree of the position that all explanation of particular events is either causal or grounding explanation (see e.g. Railton, 1981; Salmon, 1984; Lewis, 1986; Skow, 2014), I take it that not all will find this a persuasive counterexample to my conjecture.
References
Audi, P. (2012). Grounding: Toward a theory of the “In-Virtue-Of” relation. Journal of Philosophy, 109, 685–711.
Baker, L. R. (1993). Metaphysics and mental causation. In J. Heil & A. Mele (Eds.), Mental Causation (pp. 75–95). Clarendon Press.
Bennett, J. (1988). Events and their names. Hackett Publishers.
Bennett, K. (2017). Making things up. Oxford University Press.
Bernstein, S. (2020). Could a middle level be the most fundamental? Philosophical Studies.
Clapp, L. (2001). Disjunctive properties: Multiple realizations. Journal of Philosophy, 98, 111–136.
Correia, F. (2010). Grounding and truth-functions. Logique Et Analyse., 53, 251–279.
Dennett, D. C. (1973). Mechanism and responsibility. In T. Honderich (Ed.), Essays on freedom of action (pp. 159–184). Routledge & Kegan Paul.
Fine, K. (2012a). A Guide to Ground. In F. Correia & B. Schnieder (Eds.), Metaphysical grounding: Understanding the structure of reality (pp. 37–80). Cambridge University Press.
Fine, K. (2012b). The pure logic of ground. Review of Symbolic Logic, 5, 1–25.
Gallow, J. D. (2015). The emergence of causation. Journal of Philosophy, 62, 281–308.
Gibb, S. (2014). Mental causation. Analysis Reviews, 74, 327–338.
Goldman, A. (1970). A theory of human action. Princeton University Press.
Hall, N. (2004). Two concepts of causation. In J. Collins, N. Hall, & L. A. Paul (Eds.), Causation and counterfactuals (pp. 225–275). MIT Press.
Halpern, J., & Pearl, J. (2005). Causes and explanations: A structural-model approach. Part I: Causes. British Journal for the Philosophy of Science, 56, 843–887.
Hitchcock, C. (2001). The Intransitivity of causation revealed in equations and graphs. Journal of Philosophy, 98, 273–299.
Hoffmann-Kloss, V. (2014). Interventionism and higher-level causation. International Studies in the Philosophy of Science, 28, 49–64.
Jackson, F. (1995). Essentialism, mental properties and causation. Proceedings of the Aristotelian Society, 95, 253–268.
Jackson, F., & Pettit, P. (1988). Functionalism and broad content. Mind, 97, 381–400.
Jackson, F., & Pettit, P. (1990). Program explanation: A general perspective. Analysis, 50, 107–117.
Kim, J. (1974). Noncausal connections. Noûs, 8, 41–52.
Kim, J. (1984a). Epiphenomenal and supervenient causation. Midwest Studies in Philosophy, 9, 257–270.
Kim, J. (1984b). Supervenience and supervenient causation. The Southern Journal of Philosophy, 22, 45–56.
Kim, J. (1992). Multiple realization and the metaphysics of reduction. Philosophy and Phenomenological Research, 52, 1–26.
Kim, J. (1998). Mind in a physical world. MIT Press.
Kim, J. (2005). Physicalism, or something near enough. Princeton University Press.
Kistler, M. (2017). Higher-Level, downward, and specific causation. In: M. Paolini Paoletti, & F. Orilia (Eds.), Philosophical and scientific perspectives on downward causation, (pp. 54–75). Routledge, Milton Park.
Kroedel, T., & Schulz, M. (2016). Grounding mental causation. Synthese, 193, 1909–1923.
Lee, S. (2021). The duality of causality. Ph.D. dissertation, Department of Philosophy, New York University, New York, https://search.proquest.com/docview/2507182626.
Levine, J. (2001). Purple Haze: The Puzzle of Consciousness. Oxford University Press.
Lewis, D. (1973). Causation. The Journal of Philosophy, 70, 556–567.
Lewis, D. (1979). Counterfactual dependence and time’s arrow. Noûs, 13, 455–476.
Lewis, D. (1986). Causal Explanation. his Philosophical Papers (Vol. 2, pp. 214–240). Oxford University Press.
Lewis, D. (1994). Reduction of mind. In S. Guttenplan (Ed.), A Companion to the Philosophy of Mind (pp. 412–431). Blackwell.
List, C., & Menzies, P. (2009). Nonreductive physicalism and the limits of the exclusion principle. Journal of Philosophy, 106, 475–502.
Lyons, J. C. (2006). In defense of epiphenomenalism. Philosophical Psychology, 19, 767–794.
Menzies, P. (1988). Against causal reductionism. Mind, 97, 551–574.
Papineau, D. (2013). Causation is macroscopic but not irreducible. In S. C. Gibb & R. Ingthorsson (Eds.), Mental Causation and Ontology (pp. 126–152). Oxford University Press.
Paul, L. A., & Hall, N. (2013). Causation: A user’s guide. Oxford University Press.
Pettit, P. (2018). Three Mistakes about Doing Good (and Bad). Journal of Applied Philosophy, 31, 1–25.
Railton, P. (1981). Probability, explanation and information. Synthese, 48, 233–256.
Rosen, G. (2010). Metaphysical dependence: Grounding and reduction. In B. Hale & A. Hoffman (Eds.), Modality: Metaphysics, logic and epistemology (pp. 109–136). Oxford University Press.
Ross, D., & Spurrett, D. (2004). What to say to a skeptical metaphysician: A defense manual for cognitive and behavioral scientists. Behavioral and Brain Sciences, 27, 603–627.
Rydéhn, H. (2018). Grounding and ontological dependence. Synthese.
Salmon, N. (1984). Scientific explanation and the causal structure of the World. Princeton University Press.
Sartorio, C. (2006). On causing something to happen in a certain way without causing it to happen. Philosophical Studies, 129, 119–136.
Schaffer, J. (2009). On What grounds What. In D. Chalmers, D. Manley, & R. Wasserman (Eds.), Metametaphysics (pp. 347–383). Oxford University Press.
Schaffer, J. (2012). Grounding, transitivity, and contrastivity. In F. Correia & B. Schnieder (Eds.), Metaphysical Grounding (pp. 122–138). Cambridge University Press.
Shoemaker, S. (2001). Realization and mental causation. In: Gillett & Loewer (Eds.), Physicalism and its discontents, (pp. 74–98). Cambridge University Press, Cambridge.
Sinnott-Armstrong, W. (2021). Contrastive Mental Causation. Synthese, 198, 861–883.
Skow, B. (2014). Are there non-causal explanations (of Particular Events)? The British Journal for the Philosophy of Science, 65, 445–467.
Strevens, M. (2008). Depth: An account of scientific explanation. Harvard University Press.
Tammelleo, S. (2008). A limited defense of epiphenomenalism. South African Journal of Philosophy, 27, 40–51.
Van Gulick, R. (1993). Who’s in charge here? And Who’s Doing All the Work? In J. Heil & A. Mele (Eds.), Mental causation (pp. 233–256). Clarendon Press.
Walters, L. (2016). Possible world semantics and true-true counterfactuals. Pacific Philosophical Quarterly, 97, 322–346.
Wilson, J. (1999). How superduper does a physicalist supervenience need to be? Philosophical Quarterly, 49, 33–52.
Woodward, J. (2008). Mental causation and neural mechanisms. In J. Hohwy & J. Kallestrup (Eds.), Being reduced: New essays on reduction, explanation, and causation (pp. 218–262). Oxford University Press.
Yablo, S. (Forthcoming). Relevance without minimality. In A. Egan and D. Kindermann (eds.), Unstructured Content. Oxford University Press, Oxford
Yablo, S. (1992). Mental causation. Philosophical Review, 101, 245–280.
Acknowledgments
I am grateful to audiences at the 2019 SINe conference on the Future of Neuroethics and the 2020 University of Illinois at Urbana-Champaign conference Causation and Reduction: From Metaphysics to the Sciences for their feedback on this paper. Thanks also go to Alfred Mele, Jennifer McDonald, Reuben Stern, Nathaniel Baron-Schmitt, Zachary Goodsell, Caroline Torpe Touborg and Daniel Hoek for discussion on the material in the paper, and to Adam Lovett, Brad Weslake, David Papineau and two anonymous referees for comments on drafts. Special thanks are due to Laura Franklin-Hall for extensive comments across a number of drafts. And especially special thanks are due to Michael Strevens and Cian Dorr for time and time again helping me to enormously improve the paper through both discussion and their detailed written comments.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Lee, S. Building low level causation out of high level causation. Synthese 199, 9927–9955 (2021). https://doi.org/10.1007/s11229-021-03231-3
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11229-021-03231-3