Date of Degree

9-2017

Document Type

Dissertation

Degree Name

Ph.D.

Program

Philosophy

Advisor

Graham Priest

Committee Members

Shamik Dasgupta

Barbara Montero

David Papineau

Jonathan Schaffer

Subject Categories

Philosophy

Keywords

Ontological Dependence, Causality, Fundamentality, Metaphysics

Abstract

Unity is enchanting. That is, it is an enchanting idea that although reality seems to be fragmented in various ways, this apparent fragmentation belies a fundamental unity. This dissertation is an attempt at theoretically capturing such unity via the theoretical unification of causation and grounding. And this unification supports another: the unification of reality’s causal structure and reality’s hierarchical structure.

To propose a unification of causation and grounding, I employ a common conception of grounding on which grounding is understood as a dependence relation. On this conception, if the Xs ground the Ys, then the existence of the Ys depends upon the existence of the Xs. Thus, grounded entities ontologically depend upon their grounds. According to the standard view on the matter, this kind of dependence is distinct from causal dependence, and thus grounding and causation are distinct dependence relations. So, on this view, dependence comes in distinct kinds: causal and ontological. Dependence thus contrasts with a relation like identity, for example, which does not admit of variegated kinds: identity, unlike dependence, is uniform in all instances, since identity does not admit of distinct kinds. The standard view thus suggests that dependence is disunified. For this view suggests that there is a robust non-uniformity between instances of dependence: a non-uniformity between the instances of dependence that are considered cases of causation and the instances of dependence that are considered cases of grounding.

In critical response, I explore a revisionary view on which the distinction between causation and grounding is a false distinction. On “grounding-causation-identity or “GCI” for short, causation and grounding are numerically identical relations: what it is for causation to hold is no different than what it is for grounding to hold. For dependence does not admit of distinct kinds, such as causal dependence and ontological dependence: just as the nature of identity does not vary across instances, for identity does not admit of distinct kinds, the nature of dependence does not vary across instances, for it also does not admit of distinct kinds. GCI thus suggests, against the standard view, that dependence is fundamentally unified. For the nature of dependence is uniform in all instances.

Further, as mentioned above, this unification of causation and grounding supports a unification of causal structure and hierarchical structure. That is, the unity of ontological structure is a ramification of the unity of dependence. This is because, as it’s commonly thought, ontological structure is “built” from dependence relations: on the one hand, reality’s causal structure is built from causal relations, and on the other, reality’s hierarchical structure is built from grounding relations. Since ontological structure is built from dependence relations, the distinction between causal structure and hierarchical structure hinges on the distinction between causal dependence and ontological dependence. Thus, because the standard view holds that causal dependence is distinct from ontological dependence, this view implies that reality’s causal structure is distinct from reality’s hierarchal structure. And so this view implies that ontological structure is fundamentally disunified. In contrast, since GCI denies the distinction between causal dependence and ontological dependence, GCI thus implies that the distinction between causal structure and hierarchical structure is also a false distinction: just as dependence is of one basic nature that does not vary, ontological structure is of one basic nature that does not vary. In this sense, GCI implies that on ontological structure is fundamentally unified.

The discussion throughout the dissertation consists in a multi-stage comparative analysis in which GCI is compared with the standard view as well as other alternative views: views which either portray the connection between causation and grounding differently than GCI or deny that there is any such connection. Though GCI’s rival views disagree about how or if causation and grounding are connected, these rival views are in agreement that causation and grounding are numerically distinct. In this respect, these rival views are each a specific variety of what I call “grounding-causation-non-identity” or “GCN” for short: the basic view that the relations are numerically non-identical.

The standard view on the matter, discussed above, I call “common-genus-GCN.” For on this view, although the relations are numerically distinct, they are tightly connected in the respect that they are species of a common genus: the genus of dependence. On another variety of GCN, although the relations are numerically distinct, they are tightly connected in the respect that one of these relations is a species of the other: either grounding is a species of causation or causation is a species of grounding. So, one of the relations subsumes the other in the way that a genus subsumes its species. Hence, I call this view “subsumption-GCN.” Lastly, on “strict-GCN,” the relations are numerically distinct, and there is no tight ontological connection between them: causation and grounding form a gerrymandered pair of relations, not a pair which is genuinely integrated.

Thus, listing the views to be discussed, the discussion consists in a comparative analysis of:

  • GCI: The view that there is no ontological distinction between causation and grounding, thus they are numerically identical.
  • Common-genus-GCN: The view that causation and grounding are numerically distinct, but they are tightly connected in the respect that they are species of the same genus.
  • Subsumption-GCN: The view that causation and grounding are numerically distinct, but they are tightly connected in the respect that one of these relations is a species of the other.
  • Strict-GCN: The view that causation and grounding are numerically distinct relations, and there is no tight ontological connection between them.

I compare these views by appeal to five main points of comparison. Thus, I present five central arguments in the discussion. And for each of these arguments, I defend a claim concerning which view is most plausible, or which views are more plausible, with respect to one of these points of comparison. These points of comparison are:

  • How well the views explain the likeness between causation and grounding.
  • How conceptually parsimonious the views are.
  • How well suited the views are to the connection between causal explanation and metaphysical explanation.
  • The ability of the views to provide an account of an anomalous dependence relation which crosses both time and ontological levels.
  • The extent to which the views are supported or undermined by disputes about the purported differences between causation and grounding.

The goal of the discussion is to establish that GCI is at least as plausible as the best rival view or views, and so GCI is to be taken just as seriously. Thus the discussion presents considerations which favor GCI over rival views, considerations which favor GCI and rival views equally, and considerations which threaten GCI. And as I propose, these considerations balance out in such a way that GCI is just as serious of a view as the alternatives.

The dissertation is arranged as follows. Chapter 1 provides the background of the discussion via an outline of grounding and a host of related issues. Further, chapter 1 motivates the discussion by explaining the similarities between causation and grounding. As it is explained, since the relations bear such remarkable and systematic similarities, it is worth asking why they bear these similarities. For it would seem reasonable to think that the relations are connected in some way which explains these similarities.

In chapter 2, the first four central arguments are discussed. The “argument from likeness” (quite obviously) focuses upon the first point of comparison: how well the views explain the likeness between causation and grounding. I draw two conclusions from this point of comparison. Firstly, I conclude that strict-GCN is the least plausible view in question. As this argument goes, it speaks in favor of a view if the view can explain the similarities between the relations and it speaks against a view if it cannot. Thus, because strict-GCN is the only view of the four which cannot explain the similarities, it is the least plausible. Because of this, strict-GCN is left out the discussion from this point on. Concerning the second conclusion which I draw from the first point of comparison, I argue as follows. GCI, common-genus-GCN and subsumption-GCN offer equally plausible explanations of the similarities between causation and grounding. Therefore, these views are equally plausible with respect to this point of comparison.

Chapter 2 then presents the “argument from parsimony,” which (quite obviously) focuses upon the second point of comparison: how conceptually parsimonious the views are, where a theory’s conceptual parsimony, or lack thereof, is a matter of how many primitive notions it employs. As before, I draw two conclusions from this point of comparison. Firstly, I argue that because common-genus-GCN is less conceptually parsimonious than GCI and subsumption-GCN, common-genus-GCN is the least plausible with respect to this point of comparison. Secondly, I argue that this point of comparison favors GCI and subsumption-GCN equally: since neither of these views employ fewer primitive notions than the other, they are conceptually parsimonious to the same extent.

Chapter 2 then present the “argument from explanation” which focuses on the third point of comparison: how well suited the views are to the connection between causal explanation and metaphysical explanation. As it is commonly thought, causation corresponds to causal explanation and grounding correspond to metaphysical explanation. As I suggest, if the distinction between causal explanation and metaphysical explanation holds, or if it clear what this distinction amounts to, this speaks in favor of GCN and against GCI. For GCN is better suited to the distinction between these kinds of explanation. This is because if one adopts GCN (in any variety), then one can treat the distinction between causal and metaphysical explanation such that it reflects the distinction between causation and grounding. And in this case, if one adopts GCI, then one must treat causal explanation and metaphysical explanation such that they fail to reflect the identity of causation and grounding.

Conversely, if the distinction between causal explanation and metaphysical explanation does not hold, or if it is not clear what this distinction amounts to, this speaks in favor of GCI and against GCN (in all varieties). For GCI is better suited to the lack of distinction between these kinds of explanation, or it not being clear what this distinction amounts to. This is because if one adopts GCI, then one can treat the lack of distinction between causal and metaphysical explanation such that it reflects the identity between causation and grounding. And in this case, if one adopts GCN, then one must treat causal explanation and metaphysical explanation such that they fail to reflect the non-identity between causation and grounding. As I argue, it is not clear what the distinction between causal explanation and metaphysical explanation amounts to, and so the distinction can be plausibly denied. Thus, GCI is better suited to the connection between causal explanation and metaphysical explanation, and so this point of comparison favors GCI over GCN.

Lastly, chapter 2 presents the “argument from double-crossers,” which focuses upon the fourth point of comparison: the ability of the views to provide an account of an anomalous dependence relation which crosses both time and ontological levels. I call such dependence relations “double-crosser” relations. Because double-crossers cross both time and ontological levels, they qualify as both causal relations and grounding relations. As I argue, because double-crossers have this dual status, they pose a serious challenge to all varieties of GCN. For all varieties of GCN rely on the idea that one of the features which individuates grounding is the ordering that grounding imposes. And this means that all varieties of GCN require that causal relations cannot impose the kind of ordering which is characteristic of grounding. However, because double-crossers qualify as causal relation and grounding relations, double-crossers are causal, though they impose the ordering that is characteristic of grounding. So all varieties of GCN fail to provide a tenable account of double-crossers. In contrast, double-crossers pose no problem for GCI. Since GCI denies the distinction between causation and grounding, GCI denies that there is any feature of grounding which causal relations cannot have (in fact, GCI relies on the idea that causal relations can possess any feature of grounding, and vice versa). Thus, with respect to this point of comparison, I conclude that GCI is the most plausible view in question.

Chapters 3 and 4 both discuss the fifth point of comparison: the extent to which the views are supported or undermined by disputes about the purported differences between causation and grounding. Considerations about these purported differences are crucial to the discussion. For if one thinks that the relations do indeed bear differences, then one must affirm that the distinction between causation and grounding holds. And if one affirms that the distinction holds, then one must reject GCI. For GCI relies on denying that the distinction holds. Conversely, if all the purported differences between the relations can be denied or called into substantial doubt, then the distinction between the relations can be plausibly denied. And therefore, GCI need not be rejected. Thus, the goal of the “argument from instability”—the argument which concerns the fourth point of comparison—is exactly that: showing that all the purported differences between causation and grounding can be denied or called into substantial doubt. In my terminology, because all the purported differences between the relations are argumentatively “unstable,” the distinction between the relations is argumentatively unstable. For if all the differences between the relations can be rejected or doubted, then there is no argumentatively stable foundation for affirming that the distinction holds.

Since there is a wide variety of purported differences between the relations, and since many of the disputes about these differences are quite complex, the discussion of these differences spans two chapters: chapters 3 and 4. And so, the argument from instability spans chapters 3 and 4. The first part of the argument, discussed in chapter 3, addresses a cluster of argumentatively-integrated differences between causation and grounding. These differences are argumentatively-integrated in the sense that, if it’s reasonable to think that one of the differences in the cluster does not hold, this provides support for thinking that some of the other differences in the cluster do not hold. Thus my criticisms of these differences intertwine. The second part of the argument from instability, discussed in chapter 4, addresses argumentatively-unintegrated differences between the relations: differences which require their own independent criticisms.

In chapter 5, I offer a general conclusion of the discussion. The discussion shows that there is a huge variety of issues to take into account, and many of these issues are quite complex and controversial. As I suggest, these considerations balance out in such a way that GCI is no less plausible than the alternative views (i.e. common-genus-GCN and subsumption-GCN). Thus, GCI is to be taken just as seriously. I then close the discussion by elaborating some interesting ramifications of GCI.

Included in

Philosophy Commons

Share

COinS
 
 

To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.