Research‎ > ‎

Getting Causes From Powers

Stephen Mumford and Rani Lill Anjum

Oxford University Press 2011. Order here (UK) or here (US).

Causation holds the whole universe together and it is relevant to every single subject area. If there were no causes, nothing we did could be of any consequence. But as a species we still have failed fully to grasp what causation is. Does anything else have quite this combination of being so ubiquitous and vital to us yet also so little understood?

The authors construct a new theory of causation based on real dispositions or powers. It will be a surprise to many of those who already work on causation yet it should strike a newcomer as good common sense. A number of philosophical orthodoxies are challenged. For centuries it has been assumed that correlation is evidence of causation, that the cause always precedes the effect in time, that causes necessitate their effects, and that inductive inference is flawed. The authors show in turn how each of these and other orthodoxies can be challenged.

Reviews and comments

E J Lowe: Mumford and Anjum on causal necessitarianism, Analysis 2012

Julia Weckend: Book Review of Getting Causes from Powers, Ratio 2014

Luke Glynn: Review of Getting Causes from Powers, Mind, 121, 2012

Jennifer McKitrick: Book Review of Getting Causes from Powers, Analysis, 2013

Anjan Chakravartty: Book Review of Getting Causes from Powers, British Journal for the Philosophy of Science, 2013

Ruth Groff: After Hume: Realism about Powers in Contemporary Analytic Metaphysics, Journal of Critical Realism, 13(4)

Steve Esser: Short review of Getting Causes from Powers, Guide to Reality

Esteban Cespedes: Review of Getting Causes from Powers, Critica 2012

José Sebastián Briceño: Review of Getting Causes from Powers, The Review of Metaphysics 2012

Boris Hennig: Book Review of Getting Causes from Powers, Philosophical Quarterly 2015

47 ideas from Getting Causes from Powers, Philosophy Ideas Database

Colin Batchelor and Janna Hastings: 'Waves and fields in bio-ontologies'

Jennifer McKitrick, Anna Marmodoro, Stephen Mumford, Rani Lill Anjum: 'Causes as Powers - book symposium on Getting Causes from Powers', Metascience 2013

Maria José Garcia-Encinas: Interview with Stephen and Rani about Getting Causes from Powers in The Reasoner 6 (10), October issue, 2012

Summary of the book

Chapter 1: Passing Powers Around

This chapter sets out the basics of the dispositional theory of causation.  An ontology of dispositions is to be assumed, rather than defended in detail, and a pandispositionalism is offered in which properties are taken to be, as Shoemaker once argued, cluster of powers. Causation is then understood in terms of powers manifesting themselves. But as a manifestation will always be in a property, the manifestation of a power will amount to the acquisition of a new power. Rather than being a problem for pandispositionalism, as Armstrong has suggested, this actually gives us a very attractive way of understanding causation, namely, as the passing around of powers.

Chapter 2: Modelling Causes as Vectors

Neuron diagrams is the standard for representing causal situations. But their inadequacies allow us to overlook certain key features of causation. Causes are typically complex or polygenous, as Mill’s notion of total cause allowed. Lewis rejected Mill’s notion, choosing to focus on what it was to be a cause rather than the cause. We will argue that this was a mistake but first we need an alternative way of modelling causation. We offer a novel representative model. Powers are plotted as vectors on an n-dimensional quality space. Like powers, vectors have a direction and an intensity. The model will also allow us to make use of a further notion of vector addition to explain composition of causes.

Chapter 3: Against Necessity
This chapter presents an argument against causal necessitarianism, the view that causes necessitates their effects. Something could always be added to the total cause, an interferer, that can prevent the effect. Even where a cause did succeed in producing its effect, it cannot have done so by necessitating it. A threshold account is presented to explain how there can be causal production without necessitation.

Chapter 4: Composition, Non-linearity and Emergence
This chapter takes a closer look at the composition of causes. Vector addition suffices as a model of composition only for linear interactions, i.e. where composition is simple addition. But there are many cases where causes do not add. One chocolate bar disposes towards pleasure, for instance, while ten eaten together dispose towards pain. Non-linear composition can be explained in the following ways. First, there can be composition by non-linear functions. Addition is the very simplest function, but there is no reason why powers cannot compose in other ways. Second, a form of emergentism can be accepted, which means that when two powers meet then can give rise to some novel phenomenon. Third, the notion of a threshold illustrates that, at a certain point, conspicuous change can be triggered.

Chapter 5: Reciprocity and Simultaneity
The chapter considers two traditional issues in causation in the light of the powers theory. A principle of reciprocity is attractive: during causation, it is not just the effect that involves change but there is also a reciprocal and opposite change in the cause. The powers view explains this in terms of mutual manifestation partners. Ice cools the drink, but also itself melts, making it unclear what is cause and what is effect. On the powers view, reciprocal partners come together to produce a single process at the end of which both partners are changed. This has implications for simultaneity. Causes will be simultaneous with their effects, though this does not mean that causation is instantaneous.

Chapter 6: Explanation, Absences and Counterfactuals
This chapter deals with the question of whether absences can be causes on the dispositionalist view. But an absence is nothing and cannot have any powers, so cannot be a cause of anything. On our theory, an efficient cause is an added power that takes a situation out of equilibrium. But the removal of a power could also take a situation out of equilibrium. But it is still, crucially, the remaining component vectors that produce the directed resultant and not the absent vector. The vector model nevertheless illustrates why it is useful to invoke absences in causal explanation. A similar account is given of counterfactuals, in which component causes are removed from a situation in the imagination only.

Chapter 7: The Logic of Causation
This chapter develops an upshot of the dispositional theory, which is that causal claims will have a different logic to that of identity claims and essentialist claims. All of these have been expressed in universally quantified conditionals but doing so disguises some key distinctions. A distinction is drawn between conditional and categorical claims and it is shown that the two relate differently to matters of counterexample, inductive inference and necessity.

Chapter 8: Primitive Modality
The chapter develops the idea that there is a sui generis, irreducible dispositional modality involved in causation. Causes dispose towards their effects in a way that is more than purely contingent but less than purely necessary. Although this modality cannot be analysed, we can identify some of the inferences that it sustains. In addition, we can understand the modality by analogy with normativity and intentionality. It goes too far to say that dispositionality can be explained in terms of normativity (as Lowe does) or intentionality (as Molnar does). We take dispositionality to be the more basic of these notions.

Chapter 9: Perceiving Causes
That dispositionality cannot be analysed does not automatically make it problematic. Not all concepts can be analysed and some are instead known directly from experience. This chapter considers the case for our knowledge of the dispositional modality being gained empirically. There is a particular focus on bodily experiences drawing attention to the fact that we are both causal patients and agents. This will be crucial because we need two elements to understand dispositionality: 1. that a power disposes in a certain direction, which we understand when it acts upon us, and 2. that a power can be prevented from manifesting, which we understand when we resist it.

Chapter 10: A Biologically Disposed Theory of Causation
This chapter considers some examples of biological causation and it is argued that biology exemplifies the general nature of causation developed in the book. An adequate understanding of the causal role of for instance genes seems to require acknowledgment of complexity and context-sensitivity, two key notions that are emphasised in the dispositional account. Genes, it is argued, operate in a dispositional way.

Related reading:

'Årsaker som Tendenser', Salongen, October 2011


1. Passing Powers Around

1.1 Causes and powers
1.2 Causal dispositionalism
1.3 Millian causes
1.4 Particular and general causal claims
1.5 The attack on causation

2. Modelling Causes as Vectors

2.1 The neuron and the vector
2.2 What’s so wrong with neuron diagrams?
2.3 The vector and the quality space
2.4 Composition of causes and vector addition
2.5 What the vector model explains: standard cases
2.6 Mutual manifestation and single powers
2.7 Realism about component powers
2.8 Multi-dimensional cases
2.9 The model as a heuristic

3. Against Necessity

3.1 Explaining the familiar by the obscure
3.2 ‘There is a necessary connexion to be taken into consideration’
3.3 The case for causal necessitarianism
3.4 Interference and prevention
3.5 Antecedent strengthening
3.6 Other anti-necessitarians
3.7 Is addition just subtraction in disguise?
3.8 What if there is no interferer?
3.9 Shouldn’t we just include more?
3.10 Can we exclude all interferers?
3.11 How can there be causal production without necessitation?
3.12 What if determinism is true?
3.13 Probabilistic causation and multi-directional powers
3.14 The new necessitarianism
3.15 Causal necessitarianism jettisoned

4. Reductionism, Holism and Emergence
4.1 The composition of causes
4.2 Alleged problem cases for the vector model
4.3 Some possible replies
4.4 What type of emergence do we have when powers collide?
4.5 Emergent powers

5. Simultaneity
5.1 The causal relation
5.2 Temporal priority
5.3 Problems with temporal priority
5.4 Simultaneity of cause and effect
5.5 Objections to simultaneous causation
5.6 A sweet solution

6. Explanation, Absences and Counterfactuals

6.1 Epistemology and metaphysics
6.2 Explanation
6.3 Prediction
6.4 Models
6.5 Anti-deductivism
6.6 Inductive inference
6.7 Causation by absence
6.8 Causal counterfactuals

7. The Logic of Causation
7.1 Relating cause to effect
7.2 Distinguishing causal claims
7.3 Some causal and non-causal claims
7.4 Hypothetical versus categorical
7.5 Prevention, induction, and ceteris paribus clauses
7.6 Transitivity
7.7 Context-sensitivity

8. Primitive Modality
8.1 Something in between
8.2 Dispositions and necessity
8.3 What if everything is necessary?
8.4 Dispositions and possibility
8.5 Dispositionality as natural possibility
8.6 Dispositionality and normativity
8.7 Dispositionality and intentionality
8.8 Intentionality and normativity are dispositional notions
8.9 Dispositionality as a selection function
8.10 The conditional analysis
8.11 Conclusion

9. Perceiving Causes

9.1 Primitive but empirically grounded
9.2 Causal judgements and perceptions
9.3 Where can we find the causal connection?
9.4 Bodily perception
9.5 Constant conjunction between willing and acting? 9.6 A reunificationist account of agency
9.7 The sense of proprioception
9.8 Perception and the dispositional modality
9.9 Dispositional modality is the best known

10. A Biologically Disposed Theory of Causation
10.1 A Reflective Equilibrium
10.2 Why not physics?
10.3 The central features of causal dispositionalism
10.4 Powers in biological explanation
10.5 Complexity, polygeny and pleiotropy
10.6 Context-sensitivity, plasticity and flexibility
10.7 Thresholds in biology
10.8 Emergence, holism and nonlinearity
10.9 Simultaneity
10.10 Defeasible prediction
10.11 Dispositional modality

Conclusion of this book