Properties, Powers and Structures

Issues in the Metaphysics of Realism

Alexander Bird, Brian Ellis, Howard Sankey, Alexander Bird, Brian Ellis, Howard Sankey

• Publisher: Routledge
• Year: 2013
• Print ISBN: 9780415895354
• eBook ISBN: 9781136345715
• Edition: 1
• Topic: Philosophy
• Subtopic: Mathematics General

While the phrase "metaphysics of science" has been used from time to time, it has only recently begun to denote a specific research area where metaphysics meets philosophy of science—and the sciences themselves. The essays in this volume demonstrate that metaphysics of science is an innovative field of research in its own right. The principle areas covered are:

• 
  
  

• The modal metaphysics of properties: What is the essential nature of natural properties? Are all properties essentially categorical? Are they all essentially dispositions, or are some categorical and others dispositional?

• 
  
  

• Realism in mathematics and its relation to science: What does a naturalistic commitment of scientific realism tell us about our commitments to mathematical entities? Can this question be framed in something other than a Quinean philosophy?

• 
  
  

• Dispositions and their relation to causation: Can we generate an account of causation that takes dispositionality as fundamental? And if we take dispositions as fundamental (and hence not having a categorical causal basis), what is the ontological ground of dispositions?

• 
  
  

• Pandispositionalism: Could all properties be dispositional in nature?

• 
  
  

• Natural kinds: Are there natural kinds, and if so what account of their nature should we give? For example, do they have essences? Here we consider how these issues may be illuminated by considering examples from reals science, in particular biochemistry and neurobiology.

Part IV

DISPOSITIONS AND CAUSAL
POWERS

CAUSAL DISPOSITIONALISM

Stephen Mumford and
Rani Lill Anjum

1 Causation for Dispositionalists

There are various ways that a dispositionalist could go when constructing a theory of causation based on an ontology of real dispositions or what some prefer to call powers. In this paper, we will try to spell out what we take to be the most promising version of causal dispositionalism. The broad aim is to get causes from powers. Many people share this aim and, as Molnar has already said, one of the reasons to accept an ontology of powers is the work powers can do in explaining a host of other problems, causation being among them (Molnar 2003: 186). The delivery of a plausible powers-based theory of causation is, however, overdue. So far we have only hints and false starts (Harré and Madden 1975, Bhaskar 1975, Cartwright 1989, Molnar 2003: ch. 12, Martin 2008: ch. 5). None of these accounts have gone quite in the right direction, in our view, so our ambition here is to set out the first few steps.

Let us begin by assuming the reality of dispositions or powers. There is much debate on the general issue of dispositions (Armstrong et al. 1996, Mumford 1998, Ellis 2001, Molnar 2003, Bird 2007a); but if we are to make any progress on the issue of causation, we had better leave the former debate alone. We are taking it for granted that in the world there are powers that naturally dispose towards certain outcomes or manifestations. These powers exist in objects and substances and are as real and objective as any other of those things’ qualities, such as their height, shape or mass. Although such powers can issue in manifestations, they nevertheless can exist without them and may never manifest at all. Importantly, we reject the proposal that disposition ascriptions can truly be reducible to conditional statements. If that were true, there would be no need to be realist about dispositions at all. If the conditionals used only occurrent or categorical terms then dispositions would be reducible. We think, however, there is a general and systematic reason why the conditional analysis of dispositions could never work, and we will mention this shortly. This would explain why counterexamples to all the proposed conditional analyses have not been hard to come by.

So far, this has all been mandatory in order to get started and enough might already have been said to show that there ought to be some connection between real dispositions and causation. We have stated that powers can issue in certain outcomes and this of course resembles the notion of cause and effect with which we all grapple. There is still plenty of work for us to do, however, in explaining how we get causes from powers.

Before getting on to that detail, however, there is one non-mandatory step that we should signal we are taking. There is more than one general ontology that admits the reality of dispositions. Categoricalism doesn't, but both property dualism and pandispositionalism do. In property dualism, some properties are essentially dispositional but some are not. Our own sympathies are with pandispositionalism. We will be offering a view of causation based on all properties being powerful. Having some properties that are not, or at least are not essentially so, rather complicates the matter. Something would have to be said about these other, essentially powerless, properties and how they deserve their place in our ontology. We would also have to explain how the powerful and non-powerful properties related or interacted, if at all, which also seems no easy task. Instead, we go with the view that all properties are powerful and thus with pandispositionalism (see Mumford 2008).

Assuming pandispositionalism, every power's manifestation is also a power for some further manifestation. Armstrong (2005: 314) sees that this could make of causation the mere passing around of powers. He thinks that is a bad thing, while we think it gives a very good image of causation (Mum-ford 2009). To take Hume's billiard table, which he claims provides causation's ‘perfect instance’ (1740: 137)), the various balls crashing into each other are thereby passing on power. The first ball rolls across the table, its momentum being a power to move. When it strikes the object ball, it passes on that power. Momentum transfers from cue ball to object ball. And now that the object ball has that momentum, it too can pass it on to any other ball with which it collides. It need not always be the same power that is passed on, however. Causation often involves change but the change need not just be in the cause passing on the same property to its effect, as in the momentum of the billiard balls. Dropping a fragile vase on to a hard surface results in it breaking, which is as good a case of causation as any. But the vase has powers when broken that it did not have before: its pieces can now cut. And its power to hold water is lost when it becomes broken. Powers are passed on to the manifestation, but different powers from our original disposition.

There is a further question that needs to be addressed of the relation between singular and general causal truths. What is the dispositionalist's account of this? For general causal truths, we interpret them as one type disposing towards another type, for instance, such as that smoking causes cancer. It is important to read this dispositionally. Smoking only disposes towards getting cancer. We all know that some who smoke do not get cancer but the general causal claim can remain both useful and true if we read it dispositionally. If general causal truths were interpreted as universally quantified conditionals, then it would have to come out as false that smoking causes cancer since not everyone who smokes develops cancer. But if our general causal claims were restricted only to exceptionless regularities, then we would miss most, if not all, of the general causal facts. The dispositionalist solution is to say that there is an irreducibly dispositional connection involved between smoking and cancer. The connection is less than necessary, because clearly there are some smokers who do not get cancer. But the connection is more than purely contingent. It is no mere coincidence that many who smoke do get cancer. Typically, the chance of some cancers will be far greater for smokers than non-smokers. The dispositions of things would be the chance-raisers: the worldly truthmakers of the chancy truths.

For singular causal truths, we say that they concern one particular disposing towards another, such as the striking of this match disposing towards it lighting. Most typically, though not necessarily always, singular causal claims often contain a success element. If I say (past tense) that my uncle's smoking caused his cancer, I am saying not just that his smoking disposed towards cancer but that it also ‘succeeded’ in manifesting its disposition. Saying that token a caused token b entails not just the more-than-contingent less-than-necessary dispositional connection between a and b, but also that b occurred and that b was the manifestation of a’s disposition. We can say the same about most future and present tense singular causal claims. In contrast, general causal claims do not entail the further commitments but only that there is a dispositional connection between types A and B. There may well be some tokens of those types that do not manifest their disposition and it is even a possibility, if a remote one, that none of the tokens do.

Returning to the three elements of the typical singular causal truth, we can see that all are required. Only the third element might be challenged, as it could be ventured that the causal claim only amounts to a disposing towards b and b indeed occurring. But this is inadequate for dispositionalism. There might have been some other disposition c that also disposed towards b and it, rather than a, might have manifested b. Where we say that a caused b, we must mean that a specifically did at least some of the causing of b. Perhaps some events are uncaused altogether, so a true causal claim must be about more than just something being disposed towards and occurring. A causal claim is one of responsibility.

2 Contrasts

The powers ontology has its roots in Aristotelianism in which nature is active. The chief contrast is with Humean views of the world as a succession of events or facts in which we see patterns and project our future expectations on to the world. Setting aside Hume himself, the main contemporary advocate of this view is Lewis, who developed the influential view that causation consists in (the ancestral of) a counterfactual dependence between events (Lewis 1973a). Offering an account of counterfactuals within an essentially Humean, extensionalist framework is one of his great achievements (Lewis 1973b), though few find it easy to believe that there really is a plurality of other concrete worlds, more or less like ours.

The first major contrast between our account and a Humean and Lewisian one is that they offer a reductive analysis of causation, which we do not attempt. The notion of causation, in their view, can be cashed out in non-causal terms. For Hume, causation is understood just as a constant conjunction of event types, in which each cause is also spatially contiguous and temporally prior to its effect. For Lewis, causation is understood as a counterfactual dependence between events, where that amounts merely to the fact(s) that the cause and effect both occur: but in all the closest possible worlds in which the cause does not occur, the effect does not occur either. We have doubts that these, or any other reductive analysis, will be a success. Both are vulnerable to counterexamples: constant conjunctions without causation, causation without constant conjunctions, counterfactual dependences between events without causation and causation without counterfactual dependence. We will not go into the details of these counterexamples at this point, though some will be discussed later that relate to Lewis's account.

Not every concept permits analysis into others. Some concepts may be learnt directly from experience. This does not have to mean that the concept of causation is entirely simple. We think it has two parts, as we will explain shortly, but each part has to be experienced. Causation is one of our first experiences, both as a patient and agent, and it is one of our most basic, fundamental and important. Through our bodies, we act and are acted upon and have an understanding of this as soon as we have an understanding.

According to causal dispositionalism, causation involves an irreducible dispositional modality.¹ It is about one thing tending towards another, rather than necessitating it or the two being contingent accidents. If this is right, then we can never say that if some condition C occurs, no matter how large or complex C is, then an effect E will occur. We can only say that it is disposed to happen, will tend to do so, or that it is more or less likely to happen. These are ways of gesturing towards the dispositional modality but they cannot analyse it. They offer synonyms for it, and thus cannot be used in a reductive analysis. But the alternative is that non-dispositional terms would be used that failed to capture accurately the full notion of dispositionality. The putative conditional analysis of dispositions is one that fails for this reason. Something disposing towards something else can never be captured by an analysis of the form if S, in conditions C, then M. Even if a disposition is stimulated, and in the right sort of conditions to manifest, it still only tends or disposes towards its manifestation. It doesn't guarantee it. This is why we have a rich, specifically dispositional vocabulary available to us. We cannot replace it, without loss, by the idea of a conditional where, if the antecedent is true, and any associated conditions, the consequent must also be true. That has some resemblance to dispositionality, but is not close enough. We only want that if the antecedent is true, the consequent tends to be true, and this only restates dispositionality rather than analyses it.

How, then can the idea of a dispositional modality be acquired through experience? There must be two such kinds of experience, in our view. In the case where we are causal patients, we must first experience some power acting on our bodies, such as when we walk in a gale which could blow us over. But, second, we feel that we are able to resist it. The gale is for some manifestation, but it can be prevented. Where we are causal agents, it is us trying to manifest a power, such as the power to pull something. But, again, we sometimes feel that it can be resisted, such as when something else pulls in the opposite direction. These two components give us the idea of a power being for some specific manifestation to which its relation is more than contingent, but also capable of being prevented and hence to which its relation is less than necessary.

While we do not see how this dispositional modality can be explained non-circularly in non-dispositional, non-modal terms, this is not to say that it is mysterious or little understood. Indeed, we think it is the most familiar modality to any causally engaged experiencer, which virtually all humans are. And given that causation, on our account, essentially involves this, then we have to be primitivists about causation, for it contains an unanalysable element.

So what is the point of a dispositional theory of causation if it does not give us an analysis? There are still reasons to be interested in a theory that falls short of analysis. One is that, if what we have said is right, the theory tells us why an analysis is neither possible nor required, which is an important finding itself. Second, however, the dispositional theory emphasises causation's dispositional nature, which many following in Hume's wake have overlooked. The Humean tradition has concentrated on constant conjunction, and many post-Humeans have thought that the way to improve his theory was to have something in addition, such that constant conjunction has been judged a necessary but not sufficient condition for causation. The dispositional account, in contrast, tell...

Table of contents

• Front Cover
• Properties, Powers and Structures
• Routledge Studies in Metaphysics
• Title Page
• Copyright
• CONTENTS
• NOTES ON CONTRIBUTORS
• INTRODUCTION Brian Ellis, Howard Sankey, and Alexander Bird
• I SYMPOSIUM ON PROPERTIES
• II LEVELS OF INQUIRY
• III REALISM IN MATHEMATICS
• IV DISPOSITIONS AND CAUSAL POWERS
• V PAN-DISPOSITIONALISM
• VI NATURAL KINDS
• REFERENCES
• INDEX