Explanation comes in many forms.Scientific explanationstypically concern thenatural world, and typically proceed by citing causes of naturalphenomena or subsuming them under empirical laws in some illuminatingway.Mathematical explanationstypically concern the worldof quantity and abstract structure, and typically proceed by proving atheorem in some illuminating way. The present entry focuses on afamily of explanations widely thought to be in some sensedistinctivelymetaphysical in character. Consider thefollowing non-exhaustive list of putative examples, which ought to berecognizable to any metaphysician:

(1)

Socrates the philosopher, and this bust of him in our seminarroom, are both snub nosed because they share the propertybeingsnub-nosed.

(2)

Socrates’s cloak has a color because it is green.

(3)

Socrates is the very individual he is at least in part because hehas Sophroniscus as a father.

(4)

Socrates is just because his soul is well-ordered.

Statements like (1)–(4) at least purport to expressexplanations on their surface: “because” is acommon explanatory idiom, and any of a number of others could havebeen used instead (e.g., “in virtue of”, “accountsfor”, “is the reason why”, “makes it the casethat”, or simply “explains”). Moreover, statementslike (1)–(4) aremetaphysical at least in the sensethat they are commonly put forward and assessed in metaphysics, andproceed by citing distinctions of importance to metaphysics in someilluminating way. These include distinctions ofproperty-instance,determinable-determinate,essence-accident, andanalysans-analysandumthat appear to be at play in the examples above; others might includepart-whole (e.g., “The vase is fragile because of its crystallinemicrostructure”) andmember-set (e.g., “The set {Socrates} exists because Socratesexists”), among potentially others still.

Interest in metaphysical explanation has witnessed a recent spike, yetthe topic is not a recent metaphysical innovation. Indeed, it is frontand center in the very first sentence of the field’sfoundational text, when Aristotle tells us that “All humanbeings by nature desire to know” (Metaphysics I.1,980a21; see the entry onAristotle’s metaphysics). For Aristotle, this desire is not fully satisfied until oneunderstands the basic explanatory factors responsible for the broadstructure of reality (see Lear 1988: chs. 1 and 6). Set aside whetherAristotle’s claim about human nature is correct. Even if it isnot, a wealth of philosophical questions about the nature, varieties,and scope of metaphysical explanation emerge. Are these“explanatory factors” of one type or many, and what arethey? What is it about “the broad structure of reality”that they explain? What type of understanding is gained by graspingthese explanatory factors? And even if human beings by nature desiremetaphysical explanations, to what extent (if any) can this naturaldesire actually be satisfied?

Historical interest in metaphysical explanation was notAristotle’s alone. Philosophers as diverse as Plato,Nāgārjuna, Ibn Sīnā (Avicenna), William of Ockham,Spinoza, Leibniz, Bolzano, Duhem, and Husserl (see Mulligan 2004 andalsorelated entries below), among many others, explicitly grappled with questions such asthose above, although some in a more skeptical vein than others.Moreover, countless other philosophers both inside metaphysics and outclearly presuppose that the notion is in good standing when discussingthe various types of metaphysical explanation that we discuss below.Our goal in the present entry, however, is not to provide a fullaccounting of how the concept has developed over time—here, wefocus on framing the main questions and disputes that emerge when oneattempts to state what metaphysical explanation is, exactly, and howit is supposed to work (if at all).

Our first order of business is to clarify in what sense metaphysicalexplanation is a form ofnon-causal explanation (§1), and survey some putative varieties of metaphysical explanation (§2). Afterwards, we survey three central, connected choice-points abouthow to think about metaphysical explanation at the most abstractlevel. How are metaphysical explanations related to notions ofnon-causal relationships such as property-instance, part-whole,essence-accident, derivative-ground, role-realizer, and others ofcentral metaphysical importance that metaphysical explanations aresaid to derive their explanatory power from (§3)? To what extent is metaphysical explanation a unified phenomenon (§4)? And to what extent is it an objective, mind-independent phenomenon (§5)? Next, we survey a recent research program that attempts to extend oneof the four main models ofscientificexplanation—causal-mechanistic, nomological, unificationist, andpragmatic—to metaphysical explanation (§6). We conclude by surveying some questions about the scope ofmetaphysical explanation (§7).

A clarificatory comment: for stylistic reasons, we will continue toshift back-and-forth between speaking of explanations astruths of a certain kind that can be discovered and arguedover, versus speaking of them asactivities by which thosetruths are communicated or understood. That distinction, as well aswhether explanations are in fact something else (e.g.,theories in a model-theoretic sense), will rise to theforefront later, but for now, we will speak loosely.

• 1. Metaphysical explanation as a form of non-causal explanation
• 2. Some putative varieties of metaphysical explanation
• 3. Separatism vs. unionism about metaphysical explanation
• 4. Monism vs pluralism about metaphysical explanation
• 5. Realism vs antirealism about metaphysical explanation
• 6. Models of metaphysical explanation
  • 6.1 Metaphysical explanation as quasi-causal explanation
  • 6.2 Metaphysical Explanation as nomological explanation
  • 6.3 Metaphysical explanation as explanation by unification
  • 6.4 The pragmatics of metaphysical explanation
• 7. The scope of metaphysical explanation
  • 7.1 Metaphysical explanation outside of metaphysics
  • 7.2 Do all facts admit of metaphysical explanation, or none?
• Bibliography
• Academic Tools
• Other Internet Resources
• Related Entries

1. Metaphysical explanation as a form of non-causal explanation

To help us focus in on the topic at hand, contrast examples(1)–(4) above with familiar examples ofcausalexplanation—for instance, the disappearance of a puddle throughevaporation because of the warmth of the surrounding air. What makesthis an instance of explanation is a matter for debate, although agood enough approximation for our purposes is that it turns on how (inthis case causal) explanation works. As Lange puts it: a causalexplanation

derives its explanatory power by virtue of supplying relevantinformation about the explanandum’s causes or, more broadly,about the world’s network of causal relations. (Lange 2016:18)

A puddle’s disappearance does not merelycoincide withthe warmth of the surrounding air, but rather occursbecauseof it, at least in part thanks to causal interactions between the twosystems over time.

In contrast, metaphysical explanations at least appear to work in arather different manner. (1) seems to work at least in part due tosome relationship between Socrates and the bust of him on the onehand, and a fact about those two plus a further entity, a property, onthe other. Whatever that relationship is, it seems radically differentfrom ordinary causal ones, for instance in being synchronic ratherthan diachronic. The same can be said about (2), which invokes asynchronic relationship between a determinable property and itsdeterminates. Similarly, (3) seems to work at least in part due to howthe essential nature of a thing non-causally constrains what featuresit can have, while (4) seems to work at least in part due to therebeing no difference between being just and having a well-ordered soul,which would appear to preclude any causal relationships betweenthem.

It is in the minimal sense above that all should agree with the claimthat metaphysical explanation is a form of non-causal explanation (seeSchaffer 2017: 303). Nonetheless, one must take care here in at leastthree ways. First, metaphysical explanation should not beidentified with non-causal explanation. After all, there arearguably many other forms of explanation besides metaphysicalexplanation that are non-causal; we mentioned mathematical explanationabove. But this is not to say that metaphysical explanation hasno role to play in how to understand these other types ofnon-metaphysical non-causal explanations. Take, for example,Steiner’s pioneering work on mathematical explanation, whichstarts with the thought that “to explain the behavior of anentity, one deduces the behavior from the essence or nature of theentity” (Steiner 1978: 143). To the extent this works in the way(3) above works, by invoking the notion of essence, to that extentmathematical explanation involves a form of metaphysicalexplanation.

Second, to say that metaphysical explanation is a form of non-causalexplanation does not mean that it is disconnected from the causalrealm altogether. Consider another example:

(5)

Socrates is in pain because the functional role of pain isrealized by the C-fibers firing in his brain.

A functionalist about the mind might offer (5) as a part of a morecomprehensive metaphysical explanation of the nature of pain and itsinstances. And yet, (5) invokes causation in at least two places. Forone thing, theexplanans of (5)—i.e., the fact whichdoes the explaining—invokes a type of causal process. Inaddition, theexplanandum of a metaphysicalexplanation—i.e., the fact to be explained—can invokecausation, as for instance would be the case were one to offer up ametaphysical explanation of causation itself. Thus what makes ametaphysical explanation non-causal is not that the explanans orexplanandum arethemselves non-causal, but rather that therelationshipbetween them is. For another thing, (5) works atleast in part due to the relationship between the functional role of amental state and its realizer. So if this relationship itself consistsin some relationship between the causal powers of a mental state andits realizer (perhaps the former must be a proper subset of thelatter; see J. Wilson 1999, Shoemaker 2001), then to that extent, (5)derives its explanatory power from causal facts. But (5) does notderive itsolely from these, for presumably the relationshipbetween the causal powers of a mental state and its realizer is notitself a causal one (in the case of Wilson and Shoemaker, it is theis a proper subset of relation).

Third, although a useful starting point, it is not entirelyuncontroversial to say that every form of metaphysicalexplanation is a form of non-causal explanation. Take (2), which manytake to be a case ofmetaphysical grounding, the relationship by which more fundamental facts somehow “giverise to” less fundamental ones. Although most theorists classifygrounding as non-causal, A. Wilson (2018) argues that grounding isliterally a form of “metaphysical causation”, from whichit would seem to follow that (2) is a form of causal explanation (see§6.1 below). Or take Bennett (2017: 61–62), who claims thatcausation is an instance of what she calls “metaphysicalbuilding”, every instance of which is an instance ofmetaphysical explanation (at least if understood as a wholly objectivephenomenon, rather than one requiring that certain epistemicconditions hold; see§5 for more on realist vs. antirealist conceptions of metaphysicalexplanation). Nor is our starting point, useful as it is, entirelyunproblematic. Consider another example:

(6)

The natural world began to exist because of the causal activity ofa supernatural being.

Does (6) count as a metaphysical explanation? It might seem to, givena traditional conception of metaphysics as studying the “firstcauses of things” (see the entry onmetaphysics). And yet, it appears to work solely thanks to certain causal facts.Both points can be reconciled with our framing so far. Whatever“metaphysical causation” might be, surely it is notcausationof the ordinary sort that links sunlight to puddleevaporation over time, which is all that we meant to emphasize bycharacterizing metaphysical explanation as non-causal. And moreover,whether to classify all explanations offered in metaphysics asmetaphysical explanations, rather than more narrowly as just thosethat derive their explanatory power partly from non-causal factors, isa matter of terminological decision not worth engaging with at length.Clearly there issome important difference between(1)–(5) as opposed to (6) in how they work, and that differenceseems to have something to do with the non-causal ways in which theywork.That is a topic worth investigating, not whether to use“metaphysical” in the broader way or the narrower one.

2. Some putative varieties of metaphysical explanation

A classic argument for the existence ofproperties—mind-independent abstract entities, such asthose invoked in (1), that can be shared by multipleparticulars—is that they are needed in order to metaphysicallyexplain the objective resemblances we find in the world (see the entryonproperties, especially §§3–4). Both Socrates the philosopher and the bust of him in our seminar roomare snub-nosed. However (so the argument goes), it cannot be a brutefact that this is so; there must be something in virtue of which bothof these things, separated by space and time, are snub-nosed.Specifically, (so the argument concludes), the fact that these twoentities share the same property metaphysically explains why theyresemble each other in this respect. Yet even if objectiveresemblancesare, in fact, metaphysically explained by thesharing of properties by things that resemble—a big“if”, although let us suppose so for illustrativepurposes—it remains to be said whatkind ofmetaphysical explanation is to be envisaged here. Reflecting on(2)–(4) reveals at least three varieties of metaphysicalexplanation that have been the focus of recent work on the topic andin terms of which (1) could be understood.

One option is to offer (1) as agrounding explanation. Thereis considerable controversy over what grounding is, and (as we shallsee) how grounding and explanations are related and how groundingexplanations work. We can stay somewhat neutral in saying that theytypically work by citing more fundamental facts—in the case of(1), a fact about property sharing—that in some sensenon-causally “gives rise to”, “produces”, or“generates” less fundamental facts—in this case, afact of resemblance among particulars. We say “typically”,since one may count as grounding explanations those that merelyprovide information about the grounds of a given fact(presumably of some contextually relevant type; see §5.4), ratherthan by citing its ground directly, as in (1*):

(1*)

Socrates the philosopher, and this bust of him in our seminarroom, are both snub-nosed because they share some property orother.

At any rate, to treat (1) and (1*) as grounding explanations would beto group them with (2), given a standardly held view, namely thatinstances of the determinable-determinate relation are paradigmaticinstances of grounding.

Although recent interest in metaphysical explanation arose from arenewed interest in grounding over the mid to late twentieth century,grounding explanation is arguably not the only form that metaphysicalexplanation can take. This point was recognized as far back asAristotle, who in theOrganon andMetaphysicsdiscusses metaphysical explanations that work by citing facts about“the what it was to be” of some target phenomenon (see,e.g.,Posterior Analytics 83a7;Topics 141b35; andMetaphysics 1003b24, 1006a32, 1006b13), or what we would nowcall itsessence. Like grounding, there is controversy overwhat essence is; unlike grounding, much less has been said about howessence and explanation are related and how what might be calledessentialist explanations work. We can again stay somewhatneutral in saying that they typically work by citing (or perhaps justproviding contextually relevant information about) part of what it isto be some phenomenon, in terms of some other phenomenon (e.g., Fine2015 and Koslicki 2012).

Offering (1) as an essentialist explanation would be to group it with(3), which might suggest rather different truth conditions for (1)than if it were offered as a grounding explanation.Primafacie at least, (3) does not explicitly require thatSocrates’s having a certain man as a father be a morefundamental fact than him being the very individual he is. Moreover,it is widely held that the grounds of a fact are modally sufficientfor it to obtain, but need not be modally necessary for it, while theopposite is true for essence, which would place rather differentconditions on whether a claim like (2) is true as opposed to a claimlike (3). Although a number of recent accounts of grounding analyze itin terms of essence (e.g., Correia 2013, Zylstra 2018), or vice versa(e.g., deRosset 2013, Gorman 2014), others do neither (e.g., Fine2012, 2015). As a provisional matter, we shall treat groundingexplanations as distinct from (yet perhaps related to) essentialistexplanations and leave open for now whether instances of the one implyinstances of the other.

A third option is to instead offer (1) as what we will call areductive explanation, which would be to group (1) with (4)(see the entry onscientific reduction). Like grounding and essence, there is some controversy over whatreduction is, although somewhat more has been said about how reductionand explanation are related and how reductive explanations work thanin the case of essence. We can again stay somewhat neutral in sayingthat reductive explanations typically work by “analyzing”or “breaking down” some phenomenon as “nothing morethan” another phenomenon. Although a number of recent accountsof reduction analyze the notion in terms of grounding, or essence, orboth (e.g., Correia 2017; Fine 2001, 2015; Rosen 2015), others do not(e.g., Dorr 2016, King 1998). So here too, as a provisional matter weshall treat reductive explanation as distinct from (yet perhapsrelated to) both grounding and essentialist explanation and leave openwhether instances of the one imply instances of the other.

We are not suggesting that this list of varieties of metaphysicalexplanation is comprehensive. For instance, consider:

(7)

Socrates has Sophroniscus as a father because he essentially hasSophroniscus as a father.

Although Glazier (2017) calls (7) an instance of essentialistexplanation, (7) appears to us to work in a different way than (3).Unlike (3), (7) does not attempt to explainwhat it is forSocrates to be the individual he is, or evenwhat it is forSocrates to have Sophroniscus as a father. Rather, (7) concernswhy Socrates has Sophroniscus as a father, and works byciting a certain fact about thestatus of thisrelationship—namely, that it is essential toSocrates—which would presumably help to explain why Socratesmust be so. Glazier argues that explanations like (7) do not implycorresponding grounding explanations, and thus explanations like (7)are not grounding explanations. Yet there is a more general patternhere. Even if one is skeptical of the notion of essence, one mightattempt to metaphysically explain why Socrates has Sophroniscus as afather in terms of some other status of this fact; for instance, onemight appeal to the fact that it falls under a more general “lawof metaphysics”, or one might simply appeal to the necessity ofthis fact alone (see Kappes forthcoming).

We are also not suggesting that one must treat all of these (or other)forms of metaphysical explanation as genuine. Nonetheless, they areprevalent enough in both classic and contemporary metaphysics that anaccount of metaphysical explanation that aims to be comprehensiveought to either accommodate all of them or explain why they are notgenuine forms of metaphysical explanation despite initialappearances.

3. Separatism vs. unionism about metaphysical explanation

Before, we said that metaphysical explanations “derive theirexplanatory power” somehow from non-causal relationships. Morespecific proposals about the nature and varieties of metaphysicalexplanation diverge over which relationships those are, and how thatwork is carried out. Any proposal, though, must confront perhaps themost striking feature ofany form of explanation, one thatseems even more surprising in the case of metaphysical explanation.Metaphysical explanations seem to have one footin theworld—something to do with the way it is (causally ornon-causally) structured seems to play a role. Metaphysicalexplanations aremetaphysical, after all (or so one mightthink). Yet they seem to have another foot in ourthought andcommunication about the world—our desire for explanation(even if not a natural one, as Aristotle claims) is at least in part adesire tounderstand, to make the worldintelligibleto ourselves and others, tolearn the why and how of thingsand what toexpect, to satisfy ourcuriosity.Metaphysical explanations areexplanations, after all (or soone might think). Yet how could metaphysical explanation be at oncemetaphysical, but also a form of explanation?

Following Raven (2015), who applies the terminology to grounding andgrounding explanations in particular, approaches to this fundamentalquestion might be classified with regard to whether they take a“unionist” or “separatist” line. Standardly,unionism about grounding and grounding explanation is thethesis that they are one: to be grounded justis to bemetaphysically explained in a distinctive way (e.g., Dasgupta 2014;Jenkins 2013; Litland 2013, 2017; Raven 2012; Rosen 2010; Thompson2016). (A unionist might also subscribe to the stronger thesis thatthe expressions “grounded” and “explained insuch-and-such distinctive way” are, or at least ought to be,used as synonyms, but that is not essential to how we understand theview here.) Unionists thus owe some account of how the metaphysicaland explanatory elements of metaphysical explanation can be heldtogether in a coherent package, or else jettison one element or theother.Separatism about grounding and grounding explanationis the thesis that they are separate, yet grounding explanationsnonetheless “track”, or are “backed” or“underwritten” by, instances of grounding (e.g., Audi2012, Schaffer 2016, Trogdon 2013). Separatists thus owe some accountof how that link keeps the metaphysical and explanatory elements ofmetaphysical explanation from flying off in opposite directions. Onemight think that the separatist about grounding explanation hereincurs no new debt, since separatists aboutcausalexplanation must meet an analogous challenge. However, note that thequasi-causal model of metaphysical explanation at play here is farfrom the only way in which we might model metaphysical explanation(see§6), and also that to distinguish causation from causal explanation isitself a substantive commitment.

In order to extend the unionist-separatist classification tometaphysical explanationin general, though, some care mustbe taken. First (and most obviously), those who believe that there aretwo or more forms of metaphysical explanation need not be unionists orseparatists across the board. The extent to which one is a unionist ora separatist should depend on what non-causal relations one takesstatements like (1)–(4) to derive their explanatory power from,as well as on background assumptions about what kind of thing anexplanation is and what kinds of things it relates, or even whetherthese statements ascribe relations at all on the final analysis. Forinstance, we have spoken loosely of metaphysical explanations astruths of a certain kind that can be discovered and arguedover; one might take this view, and add to it the claim that they aretruths involving a relation between facts. Such a view might precludeone from being a unionist about, say, essentialist explanations like(3), if one holds that they concern relationships that may involvethings that are not facts (e.g., Socrates and his father).

Second, a further complication is that there is an open questionwhether the various non-causal relations by which metaphysicalexplanations work mustthemselves “track” stillfurther relations. For instance, one might maintain that instances ofgrounding must somehow track the relations of whole to parts,determinable to determinate, role to realizer, set to members, orothers. Those with unionist sympathies must therefore be clear aboutwhether grounding explanations are to be identified with grounding orwith some non-causal relation thatgrounding tracks, whileseparatists must be clear about whether grounding explanations work bytracking grounding, or by tracking some non-causal relation(s) thatgrounding tracks. Similar points hold for essentialist andreductive explanations, if one holds the view that there are relationsthat essence-accident and analysans-analysandum must“track”.

Finally, there are viable approaches to metaphysical explanation thatdo not fit neatly in either the unionist or separatist camp as it isstandardly understood. For instance, Maurin (2019) notes a distinctionbetween those who believe that grounding is “by itsnature” explanatory, as opposed to those who maintain that (atbest) grounding has certain other features that make it apt to backgrounding explanations. Interestingly, the former view (which countsas unionist on Maurin’s classification) does not automaticallyimply that groundingbe grounding explanation (which isrequired to count as unionist on Raven’s classification). Afterall, knowledge is by its nature true, but it does not follow thatknowledge is truth. Setting aside the less interesting question of howone ought to use the term of art “unionism”,Maurin’s proposal—when broadened to metaphysicalexplanation in general; see Thompson (2019)—shows that there isconceptual space to distinguish the metaphysical and explanatoryelements of metaphysical explanation yet without explicitly linkingthem by appeal to a relation of backing.

4. Monism vs pluralism about metaphysical explanation

In§2, we canvassed someputatively different forms of metaphysicalexplanation. But how many are therein fact? Monists aboutmetaphysical explanation think that there is just one, whereaspluralists think that there are many. There are various differentrespects in which one might be a monist or a pluralist. For example,one might think that metaphysical explanations derive theirexplanatory power from any of a number of non-causal dependencerelations (e.g., Thompson 2019), but that all metaphysicalexplanations derive their explanatory power in the same way. Thiswould be a pluralism in one sense, because various kinds of relationsproduce metaphysical explanations, but a monism in another because allmetaphysical explanations have the same form. A more monistic spin onsuch an account would hold that grounding relations producemetaphysical explanations, but that there are various groundingmechanisms (e.g., Trogdon 2018). If the source of some metaphysicalexplanations is relational but the source of others is to be found inthe essences of things, this is another version of this type of“lightweight” pluralism.

According to a more thoroughgoing pluralism, there are a number offundamentally different kinds of metaphysical explanation. Such a viewcan be motivated in a number of ways that do not at first glancepresuppose particular views about whether relations of non-causaldependence like grounding, essence, and reduction reduce to oneanother. Litland (2013) and Richardson (2020), for instance,distinguish betweenhow-explanation, which tells ushow orthe specific way in which something is thecase andwhy-explanation, which tells uswhy it isthe case at all. Bertrand (2019) further recognizes a class ofmetaphysical explanationsby constraint:“top-down” metaphysical explanations which show howparticular instances fit into a general scheme. These different kindsof explanation are plausibly elicited by the asking of differentquestions.

In reply, the monist might deny that the asking of different kinds ofquestions to elicit metaphysical explanations implies that we shouldbe pluralists. They might simply be different ways of asking for thesame kind of explanation, for example. Moreover, why-questions aregenerally taken to elicit causal explanations but can sometimes alsobe used to elicit metaphysical explanations. One might then insistthat whilst differentiating kinds of explanation requests by thequestion asked can be a useful heuristic, it is not a perfectlyreliable method for distinguishing different kinds of explanation.

5. Realism vs antirealism about metaphysical explanation

In§3, we raised a surprising tension in theorizing about metaphysicalexplanation: at least initially, it appears at once to be aboutthe world (qua metaphysical) and abouthow we think andcommunicate about the world (qua explanation).Realistsandantirealists about metaphysical explanation disagreeabout how to resolve this tension. For realists, metaphysicalexplanation is primarily (or even entirely) objective andmind-independent. That is, there is some mind-independent fact of thematter about what metaphysically explains what; what metaphysicallyexplains what depends on what the world is like, and not (or at leastnot primarily) on how we think or talk about it. Those who deny this,and think of metaphysical explanation as inescapably linked to how wethink and communicate, are antirealists.

One can be realists about metaphysical explanation to varying degrees.An extreme realist view conjoins a realism about metaphysicaldependence relations (they are themselves completely objective andmind-independent) with the view that metaphysical explanationsjust are the relevant kinds of dependence relations. Mostphilosophers do not think metaphysical explanation is quite like this.An attractive, widely held view is that metaphysical explanations haveboth a representational component and a worldly component, as iscommonly the case when people adopt a version of the separatist viewdiscussed above (§3). So long as the worldly component of metaphysical explanation(generally taken to be a non-causal dependence relation of some sort)is itself objective and mind-independent, the relevant account ofmetaphysical explanation is generally considered realist.

However, there is room to adopt a view of metaphysical explanationaccording to which constraints on how this worldly dependence relationand the relata it holds between are represented factor into whether ornot something counts as an explanation. For example, one might thinkthat an extremely complicated description of the relation betweenparticles could fail as a metaphysical explanation of the existence ofa table, because when related to subjects with our level of cognitivesophistication, it is impossible for us to understand or to expect theexistence of the table on the basis of the relevant description. Thiswould seem a less than fully realist account of metaphysicalexplanation, because whilst the explanation includes a worldlycomponent, its counting as a metaphysical explanation or otherwisedepends on psychological factors such as the cognitive sophisticationof its likely recipients.

There are various ways to resist the above account without giving upon the idea that acts of metaphysical explanation might sometimes failon the basis of psychological factors. One might insist that ametaphysical explanation is to be distinguished from anactof metaphysical explanation, and that all that is required for thelatter is a true representation of worldly non-causal dependence.Metaphysical explanations abound to various degrees of complexity(etc.) and their existing and counting as metaphysical explanations isnot something constrained by pragmatic, epistemic, or psychologicalfactors. How well they fare in acts of explaining is another matter,one that falls under the question of whether or not they aresuccessful explanations. A different way to resist is toinsist that all that is required for realism is that one component ofthe explanation be objective and mind-independent, because whilst theterm “metaphysical explanation” refers to a representationof a worldly relation, whether something counts as a metaphysicalexplanation or not fully depends on the worldly component. (One mightmotivate this sort of view by claiming that, e.g., grounding is theessential component of any metaphysical explanation, but explanationsare representational entities and so strictly speaking it is a mistaketo describe the relation as an explanation, even though it alone doesall the explanatory work, so to speak.)

In contrast, antirealists about metaphysical explanation at the mostextreme end of the spectrum do not see any role for worldly dependencerelations (see, e.g., Dasgupta 2017 and Taylor forthcoming fordiscussion). They might adopt the two-component view discussed aboveand maintain that both the relation involved in metaphysicalexplanation and how it is represented are conceptual matters (e.g.,because they are antirealists about the relevant non-casual dependencerelations), or they might think that metaphysical explanation has onlyone component. That component might be a conceptual relation or anon-causal dependence relation understood in an antirealist way, or itmight be something more representational. For example, one mightfollow Shaheen (2017) in taking the use of causal metaphors to be acomponent of our talk about metaphysical explanation, while goingfurther than he does by denying that the metaphorical cause-like senseof “because” used in metaphysics has objective,mind-independent truth conditions. Or one might follow Thompson(2019), who thinks of metaphysical explanations as answers to certainkinds of questions. Both views are consistent with thinking thatnon-causal dependence relations play a role in metaphysicalexplanation (by being what certain kinds of because-claims are about,or helping to make it the case that a particular answer to a questionis correct) but the relations are not themselves a component of theexplanation.

Norton and Miller (2019) present what they describe as apsychologistic theory of metaphysical explanation; metaphysicalexplanations have truth conditions that appeal to the dispositions ofagents evaluating the explanation, as well as necessitation relationsbetween the explanans and explanandum. Since these necessitationrelations are objective this is not an entirely antirealist account ofmetaphysical explanation, but from the way in which mental states varybetween individuals it follows that metaphysical explanations are onlytrue or false relative to a particular context of assessment. It isthus to be considered a broadly antirealist account.

6. Models of metaphysical explanation

A common strategy when trying to get to a grip of some phenomenon isto model our understanding of its nature on some pre-existing andwell-established understanding of the nature of a (supposedlyrelevantly similar) phenomenon. In this section we survey someinstances of this strategy by investigating attempts to modelmetaphysical explanation as a kind ofcausal (§6.1),nomological (§6.2),unificationist (§6.3), orpragmatist explanation (§6.4).

6.1 Metaphysical explanation as quasi-causal explanation

That metaphysical explanation islike causal explanation is aview with quite a few proponents today (though see Taylor 2018 forsome problem cases for this view). Mostly those who think it is, alsothink that all or most metaphysical explanations track or aregrounding relations (see§3 above). According to these philosophers, more precisely, metaphysicalexplanation is like causal explanation because grounding is likecausation (Fine 2012; Schaffer 2016, 2017; A. Wilson 2018). In whatsense grounding is like causation varies among proponents of thisview. According to Schaffer, grounding and causation, although similar(because both species of the kinddirected determinationrelation), are distinct. According to Wilson, grounding and whathe calls “nomic causation” are both kinds of causation(although of a rather different sort than ordinary causation; see§2). On both accounts, metaphysical and causal explanation are alikebecause grounding and causation resemble each other. One might howeverthink that metaphysical explanation is like causal explanation simplybecause the relations those explanations trackplay the samerole vis-a-vis explanation. Fine (2012: 40) may be said to giveexpression to something like this view when he states that:“Ground, if you like, stands to philosophy as cause stands toscience”. Holding this view is compatible with the relationsthose explanations pick out being in all other respects different fromone another.

If grounding is, or is like, causation, what follows concerning ourunderstanding of the nature of metaphysicalexplanation? Thatdepends. One suggestion is that, if grounding is (like) causation, howwe understand the nature of causation (and grounding) somehow directlytranslates into an account of the nature of causal (and metaphysical)explanation. Some accounts of the nature of causation seem a betterfit than others for this type of “translation”. If, forinstance, causation is understood along counterfactual orinterventionist lines, causation and causal explanation (and, is theidea, by analogy, grounding and metaphysical explanation) seemespecially closely related (see the entry oncausation: and manipulability and the entry oncounterfactual theories of causation). On this type of account, a causal explanation amounts to laying baresystematic patterns of counterfactual dependence. In interventionistlanguage:

[T]he common element in many forms of explanation, both causal andnon-causal, is that they must answer what-if-things-had-been-differentquestions. (Woodward 2003: 221)

Schaffer (2016) proposes we model grounding “in the image ofcausation” along counterfactual lines (see A. Wilson 2018, whooffers a similar model). More precisely, he suggests we understandgrounding via so-calledstructural equation models. These aremodels which can be characterized using a structure (introduced byHalpern 2000) featuring a set of exogenous (independent) variables, aset of endogenous (dependent) variables, a set of (incompatible)values representing contrasts (such that there is a function mappingevery variable to an at least two-membered set of values), a set ofstructural equations (telling us how each endogenous variable is to beevaluated on the basis of the values of other variables), and someassignment of values to each exogenous variable. Schaffer thinks thisformalism is excellent for modeling directed dependency relationsgenerally, and so is equally fitting for modeling both causal andmetaphysical dependencies. All that needs adjusting when we move fromcausation to grounding is some of our background understanding ofsurrounding notions.

Schaffer is a separatist (see§3). Causal and metaphysical explanation are in other words distinct fromthe relations (causation and grounding, respectively) they track, andit is important not to think that everything that is true about theexplanation side of things must be true of the relation side of things(andvice versa). Schafferdoes draw one rathersubstantive conclusion from the analogy between causation andgrounding, however: that there is nodistinctive notion ofmetaphysical (as opposed to causal) explanation. There isjust explanation (simpliciter) and explanation can be backed bycausation, grounding or whatever other relations are apt to backexplanation (which, on his view, is the same as being apt to bemodeled using structural equations). Similarly, A. Wilson (2018)thinks that a “major theoretical benefit” of regardinggrounding as metaphysical causation is that this accounts, in astraightforward way, for the explanatory nature of grounding claims:grounding explanations are a type of causal explanation, and they workby identifying metaphysical causes. The result is a unified theory ofexplanation, at least of explanations that work by tracking someworldly relation (see§4), according to which all explanation is causal explanation.

Against the account, a couple of things can and have been said. First,just as there are important analogies between grounding and causation(of the ordinary, run-of-the mill sort:§1), there aredisanalogies as well. This is accepted by bothfriends and foes of the view. The question then is whether thosedifferences are serious enough to undermine the work the analogy doesin the account (proponents of the account, not surprisinglydon’t think so; for some who do: see, e.g., Bernstein 2016,Janssen 2018, and Roski 2021; see Ylikoski 2013 for a related view).Another argument (due to Janssen 2018) against understandinggrounding, and hence metaphysical explanation, in terms of structuralequation models points to the representational status of modelsgenerally. Deciding whether or not a causal model includes the rightvariables (both in number and kind) to be able to represent the causalstructure of interest sometimes involves reference to what isconsidered the theory that best captures the actual features ofparticular (local) causal processes or mechanisms. These are typicallytheories for which we havea posteriori evidence. Yet in thecase of grounding, these sorts of “evidence” formodel-aptness are not available. In fact, it is not even enough forthe grounding theorist to include allpossible scenarios, itseems the grounding theorist must countenancecounterpossibles as well. That this is so, is admitted anddefended in A. Wilson (2018). For those who think this isunacceptable, however, this is a serious problem with the account.

Another way of exploiting the analogy between grounding and causationis by modeling metaphysical explanation on so-calledcausal-mechanistic explanation via modeling grounding on causation,mechanistically understood (see the entry onmechanism in science). Something along these lines is suggested in Trogdon (2018). On thisview:

some metaphysical explanations are representations of groundingrelations as being instances of grounding mechanisms. (Trogdon 2018:1296)

The explanandum of a grounding-mechanical explanation is the obtainingof a grounded fact, and just as we can causally explain a phenomenonby providing a model of the inner workings of the mechanism thatproduces it (see Machamer et al. 2000), we can metaphysically explainwhy a grounded fact obtains by providing a model of how groundingfacts and facts they ground are connected. We explainwhy agrounded fact obtains by revealing the metaphysical determinationrelations—examples, provided by Trogdon, include set formation,the determinate-determinable relation, functional realization,constitution and mereological realization—that hold between theconstituents of the relevant facts. In other words: we explainwhy a fact obtains via accounting forhow it is(mechanistically) grounded.

Canall metaphysical explanations be understood inquasi-causal terms? Those who want to argue that they can have theirwork cut out for them. They first need to explain how explanationslike our (3) and (4) can be modeled in this way. Perhaps moresurprisingly, some argue that this sort of account might not even workas an account of all types of metaphysical explanation involvinggrounding. It is generally thought that a conjunctive fact is groundedin its conjuncts. Yet, unless we think of conjunction-introduction asa (mechanistic) determination relation—something Trogdoncautions strongly against (and others: see, e.g., Wilsch2015)—this is grounding in the absence of a groundingmechanism.

If it follows that conceptual or logical metaphysical explanationswork in a different way than other types of (grounding) metaphysicalexplanation, then the unification gained by modeling metaphysical ascausal explanation is arguably contradicted by thedisunification of metaphysical explanations of thegrounding-kind. If you think this is a problem, there are a couple ofdifferent options available: you may, first, claim that, althoughcases of “bare” grounding are bare in the sense that theydo not involve anygrounding-mechanisms, they are not bare inthe sense that they involveno mechanism whatsoever. They do,only the mechanisms they involve are notgroundingmechanisms, but ratherconceptual orlogicalmechanisms: mechanisms involving the conceptual and/or logicalanalogue of the grounding determination relations. Another option isto try to redraw the boundaries for what counts as ametaphysical explanation and argue that cases of baregrounding do not backmetaphysical but rather other typesof—conceptual or logical—explanation (e.g., Kovacs 2017;McSweeney 2020). This frees the proponent of the mechanistic view fromhaving to introduce different kinds of mechanisms but does not reallysolve the problem of disunification. Alternatively, the proponent ofthe mechanistic view could argue that, while there is metaphysicalexplanation in cases of bare grounding, those are cases wheregrounding doesn’t play an essential role in the explanation(grounding doesn’t back metaphysical explanation). The resultwould be an account of metaphysical explanation able to provide uswith a unified account only of such explanations whichessentially involve grounding. In view of the fact that manybelieve that conceptual and logical explanation, for example, need tobe separately accounted for anyway, this should perhaps beacceptable.

Thus, even if there is this interesting sense in which we can unifyquasi-casual metaphysical explanations and causal scientificexplanations, it is not clear whether all metaphysical explanation areof this form, since at least on first glance there seem to beexceptions—sometimes major exceptions—that fail to becovered by this model. Whether this is considered a problem or notwill depend on whether or not one thinks there is a special value inmonism with respect to one’s account of metaphysicalexplanation, and to what extent one takes putative examples that donot fit the mold are genuine.

6.2 Metaphysical Explanation as nomological explanation

Some have argued that metaphysical explanation ought to be understoodas a species of the kind deductive-nomological explanation. On thisview (see Hempel & Oppenheim 1948 and Hempel 1965) (scientific)explanations are sound deductive arguments, featuring the followingingredients (seescientific explanation, §2):

1. An initial condition: An event of typeFobtains.
2. A law statement: For all events, if an event of typeF obtains, then an event of typeG obtains.
3. A conclusion: An event of typeG obtains.

This way of understanding the nature of (scientific) explanation doesnot mention causation and is compatible with a Humean understanding oflaws of nature, for example one that takes laws of nature to be truthsthat belong to any deductive system that best balances simplicity andfit with the facts. In this sense, understanding explanation indeductive-nomological terms can easily be combined with an antirealismabout such explanations, if the notions of best balance, simplicity,and/or fit with the facts are understood partly in terms of how wethink or communicate (see§5). Or the account can be understood as realist in nature, if, forexample, one’s preferred account of the nature of laws of natureis in turn realist (those proposed by Armstrong 1983, Dretske 1977,and Tooley 1977 being one example).

Most of those who have argued that metaphysical explanation isdeductive-nomological have tended to defend a comparatively“heavyweight” account of the (metaphysical) laws thatmetaphysical explanations involve, and so have opted for a realistseparatist version of the account. Laws, on this view, are understoodas something more or other than the Humean regularities they entail,and in at least some cases involve or reduce to facts about anunderlying “construction” relation or essence (e.g.,Glazier 2016; Kment 2014; Wilsch 2015). For more on the nature of“laws of metaphysics”, see, e.g., Sider (2011: sect. 12.4)and Rosen (2006; 2017). For a comparatively speaking less heavyweightaccount of the nature of the laws that feature in metaphysicalexplanation—one according to which such laws are nothing but“counterfactual-supporting general principles” andtherefore “inclusive” in the sense of being neithernecessarily fundamental nor necessarily exceptionless—seeSchaffer (2017). See also Woodward and Hitchcock (2003).

Adopting a more heavyweight account of the nature of metaphysicallaws, and hence of metaphysical explanation, may come with ametaphysical cost—for instance, if one takes them to beontologically basic entities, or if one takes the distinction betweenlaws and non-laws to be a primitive one. According to proponents ofsuch an account, this is a price worth paying. One important reasonwhy the traditional deductive-nomological account of (scientific)explanation has few proponents today was that the symmetry ofdeductive entailment in certain cases seemed to interfere with ourintuitions about the asymmetry of explanation (e.g., Lehrer 1974, whogets the example from Bromberger’s (1966) famous flagpole/shadowexample, see also the entryscientific explanation §2.5). If the laws that explanation invokes involve or even reduce tosomething which has direction (from more to less fundamental) alreadybuilt in, this problem is entirely avoided. Of course, if the accountis to be able to also cover metaphysical explanation in the sense ofessentialist explanation (see§2 above), the fact that the laws have direction already built in couldbe considered a flaw rather than an advantage.

Finally, if metaphysical explanation is deductive-nomological (or“covering law”), it seems that there is no room forexplanations of rare or one-off phenomena in terms of the mechanismsthat ground (or cause) them, even in cases where, had a similarmechanism occurred, it need not have given rise to a similarphenomenon (e.g., Taylor 2016). According to Trogdon (2018; see alsoAnscombe 1971), it seems likely that some explanationsareone-off in this sense, from which it would seem to follow that, evenif some or most metaphysical explanations turn out to be deductivenomological, this cannot be true of all of them. For an argumentagainst this conclusion—on the grounds that laws of metaphysicscan be indeterministic and that (like indeterministic laws of nature)they can hence help explain rare events—see Wasserman(2017).

6.3 Metaphysical explanation as explanation by unification

Unificationism is a view of scientific explanation that, although itbuilds on some of the insights of the classic DN-model, extends andrefines that model by focusing, not on individual arguments, but onthose argument patterns which best unify the relevant phenomena. Itscore idea is that to explain is to derive as many conclusions aspossible from as few premises as possible, using as few and asstringent argument patterns as possible. The fewer patterns used, thehigher the stringency. The greater the range of conclusions derived,the greater the unification. The account (inspired by Friedman 1974)was first developed by Kitcher (1981, 1989). For a good overview, seethe entry onscientific explanation (§5).

In metaphysics, an account of explanation along something like theselines can (as noted by Roski 2019) be found already in the work ofBolzano. Recent attempts to adapt Kitcher’s model to suitmetaphysical and not just scientific explanation include Kovacs (2020)and Baron & Norton (2021) (see Bertrand 2019 for an account that,although not couched in terms of unification, has a lot in common withtheir views).

A unificationist account of metaphysical explanation is an accountaccording to which such explanations aim to unify the beliefsimplicated in metaphysical theorizing. On this view, roughly speaking,a particular derivation is a metaphysical explanation just in case itis an instance of an argument pattern that unifies metaphysicalbeliefs. Several things can be said in favor of this view. If you areno fan of grounding, the fact that it takes metaphysical explanationto be something completely independent of whatever determinative(grounding) relations there happen to be (or not), will count as a bigpoint in its favor. Another positive is the fact that it is an accountof metaphysical explanation that makes unusually good sense of thelink between explanation and understanding. Unificationism correspondsto an intuitive way of thinking about understanding: we understandbetter when we can derive large amounts of information from a fairlysmall number of premises and a small number of argument patterns. Thisis because the ability to do so suggests an ability to identifyconnections between things; to fit things together. Yet the resultingaccount is not (just) psychological. As Friedman points out,explanatoriness remains in an important sense objective if “whatis scientifically comprehensible is constant for a relatively largeclass of people” (1974: 7–8).

For those who think of metaphysical explanation as an entirelyobjective and worldly affair the account will however remain highlyunattractive. Indeed, even those who think of metaphysical explanationas partly representational—the so-called separatists—willmost likely not find it very appealing. Note, though, that some thinkthat metaphysical explanation can beeither unificationist orunderstood in terms of grounding (Bertrand 2019 is one example). Ifso, although metaphysical explanation will become a less unifiedphenomenon, the fact that it can make sense of both what Bertrand(inspired by Salmon 2006) calls “top-down” and“bottom-up” metaphysical explanation, should perhaps countas a point in its favor.

Some further objections to the account—including, among otherthings, the question of what should count asthe bestsystematization; if understanding metaphysical explanation inunificationist terms means having to give up on the “impurelogic” of such explanations, and; if given unificationism,logical truths will lack metaphysical explanation—are discussed(and ultimately rejected) in Kovacs (2020). A classic objection to theunificationist account of causal explanation, finally, is that it,just like its close cousin the DN-account (see§6.2), fails to make sense of the asymmetry of such explanations, at leastnot without assuming causal realism. A similar objection may well beavailable to those who combine unificationism and anti-realism aboutmetaphysical explanation.

6.4 The pragmatics of metaphysical explanation

Pragmatic accounts of metaphysical explanation hold that whethersomething counts as an explanation or not is context-dependent. Thisgoes deeper than thinking that explanatoryinformation iscontext-dependent; that there are pragmatic constraints on how much,how little, or what kind of explanatory information to request or givein a particular conversation is fairly widely accepted. It also goesdeeper than noting that which relation the term “explains”picks out differs from context of utterance to context of utterance(since, perhaps, one takes the term “understands” to becontext-sensitive in this way, and one takes the conferral ofunderstanding to be necessary to explanation). The thesis is ratherthe far more controversial one that metaphysical explanation isitself a context-sensitive phenomenon in the sense that thefactors relevant to evaluating the truth of a statement ofmetaphysical explanation, once the context of utterance has settledthat statement's truth-conditions, involve facts about actual orpossible (types of) explanation givers or receivers, such as factsabout their interests, conceptual proficiencies, background knowledge,and the like.

In the literature on causal explanation, the most famous proponent ofpragmatic explanation is van Fraassen (1980) who argues thatexplanations are answers to why-questions. A fairly straightforwardanalogue for metaphysical explanation would be to hold thatmetaphysical explanations are answers to what-makes-it-the-case-thatquestions. There are then various different ways in which aspects ofcontext become salient. First, the context might focus the question.When we ask “What makes it the case that the ball isred?”, we will expect different answers depending on thecontrast class we had in mind (e.g., red rather than some other color,or red rather than merely appearing to be red). Second, the contextalso determines the kinds of answers that are acceptable, which mightdiffer based on (for example) the background knowledge of theexplanation-seeker and explanation-receiver. An artist will typicallyhave a far more finely-grained conception of what particular shademakes a certain Rothko painting red than your average schoolchildwould, and so a particular answer might count as explanatory for theartist but not for a child (who may not even possess the relevantcolor concept).

There has, as of yet, been comparatively little discussion ofpragmatic accounts of metaphysical explanation in the literature. Oneexception is Thompson (2019), who develops an account along the linesdescribed above, where metaphysical explanations are pairs ofcontext-sensitive questions and answers. Because the role of contextis ineliminable, there is no sense to be made of a notion ofmetaphysical explanation free from pragmatic constraints. However,Thompson accounts for the correctness of particular explanations (inpart) by appeal to worldly dependence relations, and so thatmetaphysical explanation is pragmatic does not imply that it is whollyunconstrained by worldly dependence. A different kind of pragmatictheory of metaphysical explanation is defended by Miller and Norton(2017). Their “psychologistic” account of metaphysicalexplanation locates the truth conditions for metaphysicallyexplanatory claims in the psychological states of the agentsevaluating the relevant “because” claims, alongsideworldly necessitation relations. The relevant psychological states aredispositions towards particular beliefs, which are expected to varybetween individuals. Claims about what metaphysically explains whatare thus only true relative to a specific (type of) context withinwhich the claim's truth is being assessed.

7. The scope of metaphysical explanation

7.1 Metaphysical explanation outside of metaphysics

Thus far, we have discussed thenature of metaphysicalexplanation, but have kept silent on what itsextent mightbe. At first glance, facts of all sorts inside metaphysics and outseem to at least in principle admit of metaphysical explanation. Inthis section, we discuss recent applications in ethics, philosophy ofmind, and philosophy of social science respectively.

First, ethics. In Plato’sEuthyphro, the eponymousEuthyphro is asked by Socrates to consider the following question: Isthat which is pious loved by the gods because it is pious, or is itpious because it is loved by the gods? (Euthyphro 10a). Thisis a very early example of a request for a metaphysical explanation,for the connection between piety and being loved by the gods imaginedin the second horn does not seem to be a causal connection, but aconstitutive one (although what Plato means by “because”in the first horn, and the overall structure of Plato’sargument, remains controversial). Requests for explanation abound inmoral philosophy: one of the key aims of first-order moral philosophyis toexplain why the right actions are right and the wrongones are wrong (Fogal & Risberg 2020), and of metaethics is tosystematize those explanations. Fine (2012: 37) distinguishes betweenthree different explanatory relations: metaphysical, normative, andnatural, apparently suggesting that normative explanations andmetaphysical explanations are different kinds of explanation. However,explanations in the moral realm (e.g., that an action is bad becauseit causes pain, or that something has intrinsic value just in case itis valuable in virtue of its intrinsic properties) are oftenconsidered paradigm cases of metaphysical explanations. Berker (2018),who is a unionist, argues at length that it is metaphysical grounding(and thus metaphysical explanation) that is at issue in moralphilosophy.

Second, philosophy of mind. Although there has been limited systematicdiscussion of metaphysical explanation in the area, considerationsabout metaphysical explanation inform some of the most fundamentalquestions in the discipline, such as how it could be the case thatconsciousness exists. It is fairly common for the grounding literatureto mention by way of elucidation of the notion of grounding that wemight think of mental facts as grounded in the physical facts. Thisexample is supposed to be compelling because of the attractiveness ofthe idea that physical facts determine and metaphysically explain themental facts, and grounding is aprima facie appealing notionwith which to characterize the relationship between the mental and thephysical precisely because of its explanatory nature. This idea ismade more precise in the development ofgrounding physicalism(see, e.g., Ney 2016; O’Conaill 2018; Rabin forthcoming-a,forthcoming-b; Schaffer 2021), though this new version of physicalismhas not been met with universal approval (see, e.g., Melnyk 2016; J.Wilson 2018).

For physicalists in the philosophy of mind who maintain that themental is grounded in, supervenient on, or identical with thephysical, there is a pressing need to explain what makes it the casethat some particular phenomenal feel should be associated with a givenphysical state (see, e.g., Chalmers 1996; Levine 2001). Thisexplanatory gap is an epistemic one, and so there are two ways wemight proceed. The first is to insist that unless the explanatory gaphas been closed, we do not have a metaphysical explanation of themental in terms of the physical. The second is to allow that we mighthave a metaphysical explanation without a corresponding epistemic one.Schaffer (2017) argues that grounding, and the metaphysical laws thatprovide the framework for grounding, close the explanatory gap betweenthe mental and the physical (and many more besides). Rabin (2019)disagrees, claiming that the gap between consciousness and physicalfundamentalia has a special status, persisting even if we grant thatthe former is metaphysically explained by the latter.

Third, philosophy of social science. Theorizing about the nature ofsocial reality at least partly involves metaphysically explainingsocial reality. Debates between (methodological) holists andindividualists, for example, concern whether social activities andevents are to be explained by appeal to macro-level (social)phenomena, or by the thoughts and actions of the individuals that,most agree, make up or ground those social phenomena (see, e.g.,Popper 1945; Watkins 1952). The metaphysically explanatory nature ofsocial theorizing has not always been appreciated. This is brought outin the distinction between explanatory and ontological individualism(Lukes 1968). Explanatory individualism is the thesis that socialreality is best explained in terms of individuals, their thoughts andinteractions. Ontological individualism, on the other hand, holds thatsocial reality consists in or is exhaustively determined byindividuals, their thoughts and interactions. If, however, socialreality is or is grounded in individuals and their interactions,social reality is metaphysically explained by those individuals andtheir interactions (Epstein 2015). Both these theses are theses aboutexplanation.

Metaphysical explanation is also relevant when considering theexistence and nature of social kinds. According to Haslanger (2012),for example, being a woman amounts to occupying a certain (subordinateand hence oppressed) position in a social hierarchy. This type of viewgives rise to Euthyphro-like queries of the kind: isx awoman because she occupies a certain position in an oppressivestructure, or does she occupy this position because she is a woman?(see Ásta 2010). According to some philosophers involved inthis debate, how you answer this question will decide if your accountis realist (Barnes 2017) or deflationist (Díaz-León2018). Haslanger herself has argued that the account is bestunderstood as something in between. Against the“deflationists”, she is an extensionalist about meaning,from which it follows that it would make no sense to say that we haveto sort out what the word means before determining its extension.Against the realists, Haslanger sometimes talks as if she rejects theidea that reality is hierarchically structured and holds instead thatthe natural world is neither explanatorily prior to nor ontologicallymore fundamental than the social world (pace, e.g., Schaffer2017).

7.2 Do all facts admit of metaphysical explanation, or none?

Abstracting from specific putative applications of metaphysicalexplanation outside of metaphysics proper, can one say in more generalterms what the extent of metaphysical explanation is? On one extreme,proponents of the famous “principle of sufficient reason” contend that everything (oreverything of a certain sort, e.g., contingent facts) has anexplanation of some form or other, and Dasgupta (2016) has prominentlydeveloped a version of the principle regardingmetaphysicalexplanation, specificallygrounding explanation on his view.Notably, Dasgupta distinguishes between two ways that metaphysicalexplanation might be limited in extent. Call asubstantivefact one that is “apt for being grounded”, in the sensethat

the question of what grounds it can legitimately be raised and admitsof a sensible answer (an answer that either states its ground or elsestates that it has none),

and call a factautonomous just in case it is notsubstantive, and thus is one for which “the question of why itobtains does not legitimately arise” (2016: 383).Dasgupta’s version of the PSR is then the claim that“every substantive fact has an autonomous ground”(2016: 84).

Broadening Dasgupta’s distinction, there are two questions aboutthe scope of metaphysical explanation to consider. First, which (typesof) facts have metaphysical explanations—all, none, or somethingin the middle? Second, of those that lack metaphysical explanations,which (types of) facts are not apt to be metaphysically explained atall? Dasgupta suggests that facts about essence are autonomous facts(2016: 385). Other philosophers have challenged the idea that factsabout essences are ungrounded, and so have opened the door to factsregarding essences having metaphysical explanations (see Tillman 2016,Glazier 2017, Raven 2021).

On the other extreme, one might maintain that no fact is apt to bemetaphysically explained. One type of worry is that notions thought bysome to be constitutive of metaphysical explanation, such as groundingand essence, are somehow defective (see the entry onmetaphysical grounding §8 and the entry onessential vs. accidental properties §5). Another type of worry is that metaphysical explanations would competewith scientific, mathematical, and/or other forms of explanation in amanner that render metaphysical explanations objectionablysuperfluous. Jeff Engelhardt (2016) develops this line of thought andargues that many philosophers are implicitly committed to anobjectionable sort of overdetermination involving metaphysicalexplanation. The sort of overdetermination he has in mind regardscases where something seems to have a sufficient causal explanation aswell as a sufficient metaphysical explanation. For example, for anygiven house, that house exists both because it was caused to exist(causal explanation), as well as because it is grounded in its parts(metaphysical explanation). Perhaps the best solution to the problemwhich we can come up with, Engelhardt suggests, is that in cases suchas this one the cause in question (C) causes the grounds ofthe house to obtain, butC does not cause the house to exist.In this case, whileC does explain why the house exists, itdoes notcausally explain why the house exists. Rather,C causally explains why the parts of the house are configuredin such-and-such a manner, and those parts being configured in thatmanner in turn grounds the existence of the house.

Bibliography

• Aizawa, Kenneth and Carl Gillett (eds.), 2016,ScientificComposition and Metaphysical Ground, London: Palgrave MacmillanUK. doi:10.1057/978-1-137-56216-6
• Anscombe, G. E. M., 1971,Causality and Determination,Cambridge: Cambridge University Press.
• Aristotle,The Complete Works of Aristotle, 2 volumes,Jonathan Barnes (ed.), Princeton, NJ: Princeton University Press,1984.
• Armstrong, David M., 1983,What is a Law of Nature?,Cambridge: Cambridge University Press.doi:10.1017/CBO9781316499030
• Ásta, 2010, “Siding with Euthyphro:Responde-Dependence and Conferred Properties”,EuropeanJournal of Philosophy, 18(1), 108–125 (published underthe name “Ásta Kristjana Sveinsdóttir”)
• Audi, Paul, 2012, “Grounding: Toward a Theory of theIn-Virtue-Of Relation”,Journal of Philosophy,109(12): 685–711. doi:10.5840/jphil20121091232
• Barnes, Elizabeth, 2017, “Realism and SocialStructure”,Philosophical Studies, 174(10):2417–2433. doi:10.1007/s11098-016-0743-y
• Baron, Sam and James Norton, 2021, “MetaphysicalExplanation: The Kitcher Picture”,Erkenntnis, 86(1):187–207. doi:10.1007/s10670-018-00101-2
• Bennett, Karen, 2017,Making Things Up, Oxford: OxfordUniversity Press. doi:10.1093/oso/9780199682683.001.0001
• Berker, Selim, 2018, “The Unity of Grounding”,Mind, 127(507): 729–777. doi:10.1093/mind/fzw069
• Bernstein, Sara, 2016, “Grounding Is Not Causation”,Philosophical Perspectives, 30(1): 21–38.doi:10.1111/phpe.12074
• Bertrand, Michael, 2019, “Metaphysical Explanation byConstraint”,Erkenntnis, 84(6): 1325–1340.doi:10.1007/s10670-018-0009-5
• Bromberger, Sylvain, 1966, “Why-Questions”, inMind and Cosmos: Essays in Contemporary Science andPhilosophy, Robert Garland Colodny (ed.), Pittsburgh, PA:University of Pittsburgh Press, pp. 86–111.
• Chalmers, David J., 1996,The Conscious Mind: In Search of aFundamental Theory, Oxford: Oxford University Press.
• Correia, Fabrice, 2013, “Metaphysical Grounds andEssence”, in Hoeltje, Schnieder, and Steinberg 2013:271–296.
• –––, 2017, “Real Definitions”,Philosophical Issues, 27(1): 52–73.doi:10.1111/phis.12091
• Correia, Fabrice and Benjamin Schnieder (eds.), 2012,Metaphysical Grounding: Understanding the Structure ofReality, Cambridge: Cambridge University Press.doi:10.1017/CBO9781139149136
• Dasgupta, Shamik, 2014, “The Possibility ofPhysicalism”,Journal of Philosophy, 111(9):557–592. doi:10.5840/jphil20141119/1037
• –––, 2016, “MetaphysicalRationalism”,Noûs, 50(2): 379–418.doi:10.1111/nous.12082
• –––, 2017, “ConstitutiveExplanation”,Philosophical Issues, 27(1),74–97
• deRosset, Louis, 2013, “Grounding Explanations”,Philosophers’ Imprint, 13: art. 7. [deRosset 2013 available online]
• Díaz-León, Esa, 2018, “On Haslanger’sMeta-Metaphysics: Social Structures and MetaphysicalDeflationism”,Disputatio, 10(50): 201–216.doi:10.2478/disp-2018-0013
• Dorr, Cian, 2016, “To Be F Is To Be G”,Philosophical Perspectives, 30(1): 39–134.doi:10.1111/phpe.12079
• Dretske, Fred, 1977, “Laws of Nature”,Philosophyof Science, 44(2): 248–268.
• Engelhardt, Jeff, 2016, “The Problem of SecondaryEffects”,The Southern Journal of Philosophy, 54(2):247–266. doi:10.1111/sjp.12168
• Epstein, Brian, 2015,The Ant Trap: Rebuilding the Foundationsof the Social Sciences, Oxford: Oxford University Press.doi:10.1093/acprof:oso/9780199381104.001.0001
• Fine, Kit, 2001, “The Question of Realism”,Philosophers’ Imprint, 1: art. 2. [Fine 2001 available online]
• –––, 2012, “Guide to Ground”, inCorreia and Schnieder 2012: 37–80.doi:10.1017/CBO9781139149136.002
• –––, 2015, “Unified Foundations forEssence and Ground”,Journal of the American PhilosophicalAssociation, 1(2): 296–311. doi:10.1017/apa.2014.26
• Fogal, Daniel and Olle Risberg, 2020, “The Metaphysics ofMoral Explanations”, inOxford Studies in Metaethics, Volume15, Russ Shafer-Landau (ed.), Oxford: Oxford University Press,pp. 170–194.
• Friedman, Michael, 1974, “Explanation and ScientificUnderstanding”,The Journal of Philosophy, 71(1):5–19. doi:10.2307/2024924
• Glazier, Martin, 2016, “Laws and the Completeness of theFundamental”, inReality Making, Mark Jago (ed.),Oxford: Oxford University Press, pp. 11–37.doi:10.1093/acprof:oso/9780198755722.003.0002
• –––, 2017, “EssentialistExplanation”,Philosophical Studies, 174(11):2871–2889. doi:10.1007/s11098-016-0815-z
• Gorman, Michael 2014, “Essentiality asFoundationality”, inNeo-Aristotelian Perspectives inMetaphysics, Daniel D. Novotny and LukášNovák, New York: Routledge, pp. 119–137.
• Halpern, Joseph Y., 2000, “Axiomatizing CausalReasoning”,Journal of Artificial IntelligenceResearch, 12: 317–337. doi:10.1613/jair.648
• Haslanger, Sally, 2012,Resisting Reality: Social Constructionand Social Critique, Oxford: Oxford University Press.doi:10.1093/acprof:oso/9780199892631.001.0001
• Hempel, Carl, 1965, “Aspects of ScientificExplanation”, in hisAspects of Scientific Explanation andOther Essays in the Philosophy of Science, New York: Free Press,pp. 331–496.
• Hempel, Carl G. and Paul Oppenheim, 1948, “Studies in theLogic of Explanation”,Philosophy of Science, 15(2):135–175. doi:10.1086/286983
• Hoeltje, Miguel, Benjamin Schnieder, and Alex Steinberg (eds.),2013,Varieties of Dependence: Ontological Dependence, Grounding,Supervenience, Response-Dependence, (Basic PhilosophicalConcepts), Munich: Philosophia.
• Janssen, Lina, 2018, “When Are Structural Equations ModelsApt? Causation versus Grounding”, inExplanation BeyondCausation: Philosophical Perspectives in Non-Causal Explanation,Alexander Reutlinger and Juha Saatsi, Oxford: Oxford University Press,pp. 250–266.
• Jenkins, C. S. I., 2013, “Explanation andFundamentality”, in Hoeltje, Schnieder, and Steinberg 2013:211–242.
• Kappes, Yannic, forthcoming, “Explanation by Status asEmpty-Base Explanation”,Synthese, first online: 9November 2020. doi:10.1007/s11229-020-02901-y
• King, Jeffrey C., 1998, “What Is a PhilosophicalAnalysis?”,Philosophical Studies, 90(2):155–179. doi:10.1023/A:1004254128428
• Kitcher, Philip, 1981, “Explanatory Unification”,Philosophy of Science, 48(4): 507–531.doi:10.1086/289019
• –––, 1989, “Explanatory Unification andthe Causal Structure of the World”, inScientificExplanation, Philip Kitcher and Wesley C. Salmon (eds),Minneapolis, MN: University of Minneapolis Press, pp.410–505.
• Kment, Boris, 2014,Modality and Explanatory Reasoning,Oxford: Oxford University Press.doi:10.1093/acprof:oso/9780199604685.001.0001
• Koslicki, Kathrin, 2012, “Varieties of OntologicalDependence”, in Correia and Schnieder 2012: 186–213.doi:10.1017/CBO9781139149136.008
• Kovacs, David Mark, 2017, “Grounding and the Argument fromExplanatoriness”,Philosophical Studies, 174(12):2927–2952. doi:10.1007/s11098-016-0818-9
• –––, 2020, “Metaphysically ExplanatoryUnification”,Philosophical Studies, 177(6):1659–1683. doi:10.1007/s11098-019-01279-z
• Lange, Marc, 2016,Because Without Cause: Non-CausalExplanations in Science and Mathematics, Oxford: OxfordUniversity Press. doi:10.1093/acprof:oso/9780190269487.001.0001
• Lear, Jonathan, 1988,Aristotle: The Desire toUnderstand, Cambridge: Cambridge University Press.
• Lehrer, Keith, 1974,Knowledge, Oxford: ClarendonPress.
• Levine, Joseph, 2001,Purple Haze: The Puzzle ofConsciousness, Oxford: Oxford University Press.doi:10.1093/0195132351.001.0001
• Litland, Jon Erling, 2013, “On Some Counterexamples to theTransitivity of Grounding”,Essays in Philosophy,14(1): 19–32. doi:10.7710/1526-0569.1453
• –––, 2017, “Grounding Ground”, inOxford Studies in Metaphysics, Volume 10, Karen Bennett andDean W. Zimmerman (eds), Oxford: Oxford University Press, pp.279–316. doi:10.1093/acprof:oso/9780198791973.003.0012
• Lukes, Steven, 1968, “Methodological IndividualismReconsidered”,The British Journal of Sociology, 19(2):119–129. doi:10.2307/588689
• Machamer, Peter, Lindley Darden, and Carl F. Craver, 2000,“Thinking about Mechanisms”,Philosophy ofScience, 67(1): 1–25. doi:10.1086/392759
• Maurin, Anna-Sofia, 2019, “Grounding and MetaphysicalExplanation: It’s Complicated”,PhilosophicalStudies, 176(6): 1573–1594.doi:10.1007/s11098-018-1080-0
• McSweeney, Michaela M., 2020, “Debunking Logical Ground:Distinguishing Metaphysics from Semantics”,Journal of theAmerican Philosophical Association, 6(2): 156–170.doi:10.1017/apa.2019.40
• Melnyk, Andrew, 2016, “Grounding and the Formulation ofPhysicalism”, in Aizawa and Gillett: 2016: 249–270.
• Miller, Kristie and James Norton, 2017, “Grounding:It’s (Probably) All in the Head”,PhilosophicalStudies, 174(12): 3059–3081.doi:10.1007/s11098-016-0846-5
• Mulligan, Kevin, 2004, “Essence and Modality: TheQuintessence of Husserl’s Theory”, inSemantik undOntologie: Beiträge zur Philosophischen Forschung, MarkSiebel and Mark Textor (eds), Frankfurt: Ontos Verlag, pp.387–419.
• Ney, Alyssa, 2016, “Grounding in the Philosophy of Mind: ADefense”, in Aizawa and Gillett 2016: 271–300.
• Norton, James and Kristie Miller, 2019, “A PsychologisticTheory of Metaphysical Explanation”,Synthese, 196(7):2777–2802. doi:10.1007/s11229-017-1566-x
• O’Conaill, Donnchadh, 2018, “Grounding, Physicalismand Necessity”,Inquiry, 61(7): 713–730.doi:10.1080/0020174X.2017.1385524
• Popper, Karl, 1945,The Open Society and Its Enemies,Princeton, NJ: Princeton University Press.
• Rabin, Gabriel Oak, 2019, “Grounding the Gaps or Bumping theRug? On Explanatory Gaps and Metaphysical Methodology”,Journal of Consciousness Studies, 26(5–6):191–203.
• –––, forthcoming-a, “FundamentalityPhysicalism”,Inquiry, first online: 2 January 2020.doi:10.1080/0020174X.2019.1688177
• ––– (ed.), forthcoming-b,Grounding andConsciousness, Oxford: Oxford University Press.
• Raven, Michael J., 2012, “In Defence of Ground”,Australasian Journal of Philosophy, 90(4): 687–701.doi:10.1080/00048402.2011.616900
• –––, 2015, “Ground”,PhilosophyCompass, 10(5): 322–333. doi:10.1111/phc3.12220
• –––, 2021, “Explaining Essences”,Philosophical Studies, 178(4): 1043–1064.doi:10.1007/s11098-020-01483-2
• Richardson, Kevin, 2020, “Grounding Pluralism: Why andHow”,Erkenntnis, 85(6): 1399–1415.doi:10.1007/s10670-018-0083-8
• Rosen, Gideon, 2006, “The Limits of Contingency”, inIdentity and Modality, Fraser MacBride (ed.), Oxford: OxfordUniversity Press, pp. 13–39.
• –––, 2010, “Metaphysical Dependence:Grounding and Reduction”, inModality: Metaphysics, Logic,and Epistemology, Bob Hale and Aviv Hoffman (eds), Oxford: OxfordUniversity Press, pp. 109–136.
• –––, 2015, “Real Definition”,Analytic Philosophy, 56(3): 189–209.doi:10.1111/phib.12067
• –––, 2017, “Ground by Law”,Philosophical Issues, 27(1): 279–301.doi:10.1111/phis.12105
• Roski, Stefan, 2019, “Bolzano and Kim on Grounding andUnification”,Synthese, 196(7): 2971–2999.doi:10.1007/s11229-017-1593-7
• –––, 2021, “Metaphysical Explanations andthe Counterfactual Theory of Explanation”,PhilosophicalStudies, 178(6): 1971–1991.doi:10.1007/s11098-020-01518-8
• Salmon, Wesley, 2006,Four Decades of ScientificExplanation, Pittsburgh, PA: University of Pittsburgh Press.
• Schaffer, Jonathan, 2016, “Grounding in the Image ofCausation”,Philosophical Studies, 173(1):49–100. doi:10.1007/s11098-014-0438-1
• –––, 2017, “Laws for MetaphysicalExplanation”,Philosophical Issues, 27(1):302–321. doi:10.1111/phis.12111
• –––, 2021, “Ground Functionalism”,inOxford Studies in Philosophy of Mind, Volume 1, UriahKriegel (ed.), Oxford: Oxford University Press, 171–207.
• Shaheen, Jonathan L., 2017, “The Causal Metaphor Account ofMetaphysical Explanation”,Philosophical Studies,174(3): 553–578. doi:10.1007/s11098-016-0696-1
• Shoemaker, Sydney, 2001, “Realization and MentalCausation”, inPhysicalism and Its Discontents, CarlGillett and Barry Loewer (eds.), Cambridge: Cambridge UniversityPress, 74–98. doi:10.1017/CBO9780511570797.005
• Sider, Theodore, 2011,Writing the Book of the World,Oxford: Oxford University Press.doi:10.1093/acprof:oso/9780199697908.001.0001
• Steiner, Mark, 1978, “Mathematical Explanation”,Philosophical Studies, 34(2): 135–151.doi:10.1007/BF00354494
• Taylor, Elanor, 2016, “Explanation and the ExplanatoryGap”,Acta Analytica, 31(1): 77–88.doi:10.1007/s12136-015-0260-1
• –––, 2018, “Against ExplanatoryRealism”,Philosophical Studies, 175(1): 197–219.doi:10.1007/s11098-017-0862-0
• –––, forthcoming, “Backing WithoutRealism”,Erkenntnis, first online: 8 April 2020.doi:10.1007/s10670-020-00249-w
• Thompson, Naomi, 2016, “Grounding and MetaphysicalExplanation”,Proceedings of the Aristotelian Society,116(3): 395–402. doi:10.1093/arisoc/aow012
• –––, 2019, “Questions and Answers:Metaphysical Explanation and the Structure of Reality”,Journal of the American Philosophical Association, 5(1):98–116. doi:10.1017/apa.2018.51
• Tillman, Chris, 2016, “Essence Facts and Explanation”,Thought: A Journal of Philosophy, 5(3): 190–195.doi:10.1002/tht3.210
• Tooley, Michael, 1977, “The Nature of Laws”,Canadian Journal of Philosophy, 7(4): 667–698.doi:10.1080/00455091.1977.10716190
• Trogdon, Kelly, 2013, “An Introduction to Grounding”,in Hoeltje, Schnieder, and Steinberg 2013: 97–122.
• –––, 2018, “Grounding-MechanicalExplanation”,Philosophical Studies, 175(6):1289–1309. doi:10.1007/s11098-017-0911-8
• van Fraassen, Bas, 1980,The Scientific Image, Oxford:Clarendon Press. doi:10.1093/0198244274.001.0001
• Wasserman, Ryan, 2017, “Vagueness and the Laws ofMetaphysics”,Philosophy and Phenomenological Research,95(1): 66–89. doi:10.1111/phpr.12269
• Watkins, J. W. N., 1952, “Ideal Types and HistoricalExplanation”,The British Journal for the Philosophy ofScience, 3(9): 22–43. doi:10.1093/bjps/III.9.22
• Wilsch, Tobias, 2015, “The Nomological Account ofGround”,Philosophical Studies, 172(12):3293–3312. doi:10.1007/s11098-015-0470-9
• Wilson, Alastair, 2018, “Metaphysical Causation”,Noûs, 52(4): 723–751. doi:10.1111/nous.12190
• Wilson, Jessica, 1999, “How Superduper Does a PhysicalistSupervenience Need to Be?”,The PhilosophicalQuarterly, 49(194): 33–52. doi:10.1111/1467-9213.00127
• –––, 2018, “Grounding-Based Formulationsof Physicalism”,Topoi, 37(3): 495–512.doi:10.1007/s11245-016-9435-7
• Woodward, James, 2003,Making Things Happen: A Theory ofCausal Explanation, New York: Oxford University Press.doi:10.1093/0195155270.001.0001
• Woodward, James and Christopher Hitchcock, 2003,“Explanatory Generalizations, Part I: A CounterfactualAccount”,Noûs, 37(1): 1–24.doi:10.1111/1468-0068.00426
• Ylikoski, Petri, 2013, “Causal and Constitutive ExplanationCompared”,Erkenntnis, 78(S2): 277–297.doi:10.1007/s10670-013-9513-9
• Zylstra, Justin, 2018, “Essence with Ground”,Analytic Philosophy, 59(2): 193–207.doi:10.1111/phib.12130

Academic Tools

	How to cite this entry.
	Preview the PDF version of this entry at theFriends of the SEP Society.
	Look up topics and thinkers related to this entry at the Internet Philosophy Ontology Project (InPhO).
	Enhanced bibliography for this entryatPhilPapers, with links to its database.

Other Internet Resources

[Please contact the authors with suggestions.]

Related Entries

analysis |Aristotle, General Topics: metaphysics |Bolzano, Bernard: logic |causation: and manipulability |causation: counterfactual theories of |determinates and determinables |Duhem, Pierre |essential vs. accidental properties |grounding, metaphysical |Husserl, Edmund |Ibn Sina [Avicenna]: metaphysics |Leibniz, Gottfried Wilhelm |mathematical: explanation |mechanism in science |metaphysics |Nāgārjuna |Ockham [Occam], William |Plato: middle period metaphysics and epistemology |principle of sufficient reason |properties |reduction, scientific |scientific explanation |set theory |Spinoza, Baruch: modal metaphysics