What Tends to Be

Essays on the Dispositional Modality

Stephen Mumford and Rani Lill Anjum

Under review

Tendencies are everywhere and all our actions and expectations reveal that we are aware of this and know how they work. We are comfortable with a world of tendencies. Yet, there is not nearly as much discussion of tendencies among philosophers as there is, for instance, of necessity and possibility. Is this because philosophers think that tendencies can be explained entirely in terms of necessity and possibility? Or is it because they do not see the tendencies we see? You might sometimes be led to believe that thinking of the world in terms of both necessity and pure contingency is perfectly natural and tendency is mysterious and little understood. This is not true. Tendency, we insist, is a distinct third modal value: and really the one that we know the best.

1. Dispositional Modality

Dispositionality is a distinctive, non-alethic, sui generis modality and this explains, for instance, why each reductive analysis of dispositions fails. Dispositionality is not the same as necessity, since having a disposition towards F does not necessitate F: indeed, if F is necessary it seems that there cannot be a disposition towards it. Dispositions should always be susceptible to prevention and interference. Dispositionality is not the same as pure possibility either, as many things are in a sense possible without their being a disposition towards them, e.g. the possibility that a struck match turns into a chicken. Instead, dispositionality is taken to be a modality between necessity and pure possibility. Our use of normative and intentional notions shows that we are familiar enough with this modality, which can be explained by the fact that we acquire it directly through experience. The dispositional modality, far from requiring a reductive analysis, is thus the modality with which we are most familiar.

2. Forebears of the Dispositional Modality

There are a number of alleged forebears of the dispositional modality. Geach attributes the view to Aquinas, for example, and Aristotle, Sellars, Bhaskar, Harré and Madden and Cartwright have all been thought to hold something similar. A closer inspection, however, reveals that while many have adopted the language of tendencies, it is not clear that they have precisely the idea of the dispositional modality. Bhaskar, for instance, lapses back into talk of necessities and this is clearly more than just talk. The key choice to be made is between A) when all conditions are ideal, a power manifests necessarily, and B) when all conditions are ideal, a power will tend to manifest. B is the dispositional modality but Bhaskar and others seem to opt for A, which we can call the conditional necessity view. It appears that even Aristotle had conditional necessity rather than dispositional modality. As J. S. Mill said, however, what explicates the notion of necessity is unconditionalness, so the coherence of such a view is challenged. And doesn’t conditional necessity collapse into triviality, where every contingent truth comes out as conditionally necessary? Dispositional modality can also be distinguished from the mixed indeterminism view of Pierce, which merely throws chance events into an otherwise deterministic world, but with the implications that events produced by chance are uncaused. If we understand how the dispositional modality view is not the same as conditional necessity nor mixed indeterminism, then we understand the former better.

3. Overdisposed

Powers can ground the facts of probability: as a tendency towards a distribution, e.g. a 50/50 propensity is a tendency towards a 50:50 distribution. However, because tendencies can come in degrees, it might be tempting to think that we should explain powers in terms of probabilities, i.e. that each power is a probability of a certain effect. This would suggest that the dispositional modality was reducible to the facts of probability. The possibility of overdisposing shows, however, that there is a difference between having a power to some degree and the probability of the power’s manifestation occurring. The mathematisation of chance has given us probabilties on a bounded scale between 0 and 1 whereas the strength of a power has to be unbounded. No matter how strong a power is, there can always be another that is stronger. This is an in-principle problem, then, of converting propensities to probabilities. Previous attempts, in terms of frequencies, ratios or degrees of belief, have not overcome this. Overdisposing means there can be more than enough for the production of an effect and yet the dispositional modality tells us that the probability of a natural event will always be less than 1.

4. Powers, Causation and Potentiality

Analytic philosophers have in recent decades rediscovered causal powers as the basis for an all-encompassing metaphysics and philosophy of nature. What recommends the powers view is its explanatory utility, including a putative explanation of real, worldly potentiality. Powers can be understood as the elements in the world that provide the grounding for its potentiality. They can be productive of their manifestations but typically do so only in certain circumstances. This will explain why there can be some potentialities that are not actualised and also why there are some constraints on what can be. The powers account also makes potentialities a possible subject of scientific investigation as powers are to a degree empirically accessible. It is important, however, that we provide a plausible account of how powers are able to bring things about: how they are able to realise some of those potentialities. A mutual manifestation model is preferred to the stimulus-response model of production. It was C. B. Martin who introduced the mutual manifestation model but it is argued that his account needs to be amended so that it resembles less mereological composition and more causation.

5. Causal Dispositionalism

Dispositions have already been proposed as a basis for causes but there have been a number of false starts. A cause should be understood as something that disposes towards an effect. This will make sense of general causal claims, which seem to permit exceptions, but also particular causal claims, which seem to employ an irreducibly dispositional modality. This is not a reductive analysis of causation but allows that the idea of cause, including its dispositional nature, can be acquired from experience. A number of common presuppositions about causation should be challenged. Both contingentist and necessitarian views will be rejected. A vector model is presented as an alternative to representing causation in neuron diagrams. Vector models allow us to consider the polygeny or effects and also how causes tend to their effects in different degrees. Using the new model, we see how causation of absence can be understood, overdetermination cases and causation by absence. The theory remains realist about both the resultant powers and the constituent powers of which they are composed. The theory is also offered as an alternative to Lewis’s popular counterfactual dependence view. Many causes make a difference but not all do, hence an amended account of causal counterfactuals is required.

6. Conditional Probability from an Ontological Point of View (with Johan Arnt Myrstad)

The RATIO analysis of conditional probability stipulates that the probability of A given B is calculated from the probabilities of A and B according to the formula P(A|B) = P(A & B) / P(B) (when P(B) > 0). This has been criticised for the cases where the RATIO analysis fails to deliver a calculation but where our intuitive judgment of conditional probability is clear. Here it is argued that there are counter-intuitive results of the RATIO analysis of conditional probability even when a calculation is delivered. This shows that conditional probabilities cannot be treated as a function of unconditional probabilities. The conclusion is broader, though. We argue that it is conditionality as such that is basic and irreducible. We give an ontological account of conditionals, which highlights a causal-dispositional connection between the antecedent and consequent conditions where the conditional has to be treated as an indivisible whole rather than compositional. This shows that the ratio definition of conditional probability, including the Kolmogorov axiomatics, favours a Humean framework and should be rejected by anti-Humeans who prefer a metaphysics of powers and propensities.

7. Carnap and the Anglo-Austrian Conspiracy against Dispositions

Carnap’s motivations for introducing reduction sentences for disposition terms are easily thought of as purely scientific and logical. However, it is argued here that Carnap’s account had primarily a metaphysical motivation. The commitment to extensional logic, which generates the problem of dispositions, has a basis in Humean metaphysics in which there is only a truth value and no modal value. The chief ontological commitment of the dispositionalists should be to the reality of the dispositional modal connection in nature, as opposed to the Humean metaphysic in which all is loose and separate. This means that the dispositionalist cannot express their most important claims using the standard apparatus of classical logic, nor any logic that builds upon the classical base. It also shows that we cannot use the standard logical apparatus for adjudicating rationally between these rival metaphysical systems as logic will inherently favour Humeanism. This illustrates the Aristotelian doctrine that metaphysics is First Philosophy.

8. What We Tend To See

There are differing standards of aesthetic taste. Agreement on beauty is not universal. But nor is it a completely random matter. There is some convergence of opinion, for instance, that a sunset, a Klimt landscape, or Mozart concerto are beautiful. While there is no unanimity of opinion, this convergence suggests that aesthetic experience is not an entirely subjective matter. There seem to be certain properties of things of which we tend to have a positive aesthetic experience. This suggests a degree of objectivity in aesthetics though not complete objectivity. It is hard to rule that someone is plain wrong if they claim something atypical gives them an aesthetic experience. Another consideration against objectivity is that we can apparently choose when to take an aesthetic experience and from what: hence an object we do not usually see in such a way can nevertheless be seen aesthetically in a voluntary act. Aesthetic experience is then best understood as a mutual manifestation of perceiver and thing perceived. Certain properties are apt to produce aesthetic experiences but only when meeting a perceiver who is also disposed to aesthetically perceive those properties. Such a dispositional theory of aesthetics thus steers a middle path between a pure subjectivism and objectivism. The prospects are then considered for extending this account to perceptual experience in general in which objects have a tendency to appear a certain way in mutual manifestation with a perceiver.

9. What We Tend to Know

Many inferences we draw are inductive in the broad sense: inferences from what we experience to what we don’t experience, e.g. from the front of objects to the backs of objects. Induction is one such type of inference and the problem of induction concerns how we know that experienced cases will be like non-experienced cases. There are good reasons why some of the traditional responses to the problem are not helpful. However, for a dispositionalist it is best to say that the problem of induction is misconceived from the start, following from a faulty philosophy of nature, without which there is no such problem. A basic Humean assumption is that laws are constituted by regularities, e.g. that All Fs are Gs. How can we know, then, that every F, whether experienced or not, will be like the ones experienced? But no natural process ever necessitates a complete uniformity. There can be general causal truths that are dispositional and thus require no constant conjunctions. Where we have constant conjunctions in nature, these will have a different source, such as essence. Inductive inferences are precisely as insecure as they should be: the insecure nature of our forward-thinking inferences reflects exactly the tendential nature of the natural world.

10. It's Not All in Your Genes (with Thomas Bøhn)

The causal mechanisms of biology for a time were thought primarily to be genes. Genes indeed play an important role but not always the one they are thought to have. Some key misconceptions are that there is a one-one correlation between genotypes and phenotypes and that possession of the gene for F guarantees or necessitates F. These misconceptions appear in the popular understanding of genetic determinism. Three examples of causation in genetics are considered as examples: Huntington’s disease, Pisum sativum and Darwin’s finches. What these cases show is that the causation involved should be characterised by complexity, context-sensitivity and a lack of necessity. These features fit well with a dispositional theory of causation in which genes can be understood as dispositions or tendencies towards their effects and this offers little support for genetic determinism. In genetics, as in other sciences, much of our task is to find the right metaphors in which to express the theory. The empirical findings tell us that metaphors such as programming, determining, coding and controlling are of little value in genetics.

11. Causation in Quantum Mechanics (with Fredrik Andersen)

Quantum mechanics seems to violate classical laws of physics in a number of ways. Its puzzling features have led some to deny that there is any causation involved in quantum phenomena for there are apparent absences of the necessity, determinism, predictability and separability that should accompany the presence of causation. However, it is argued that these are not essential features of causation per se but only of a particular, though admittedly popular, theory of causation. It is shown that there is at least one other theory of causation – a dispositional one – which does not require the features found absent in quantum mechanics. There are independent considerations in favour of the theory but it is also significant that the very same account of causation could be applicable to both quantum and non-quantum phenomena. It is clearly premature, therefore, to judge there is no causation in quantum mechanics if that judgement is based on an assumption of essential features of causation that are not really there.

12. What We Tend to Mean

A dispositional theory of meaning is developed that reconciles two plausible but seemingly conflicting claims: 1. that words have a core of meaning that they bring with them into every context and 2. that all meaning is sensitive to context. Words might dispose towards a particular or ‘literal’ meaning, sometimes quite strongly. But whether this meaning is actually conveyed when expressed will depend on a number of factors, such as speaker’s intentions, the context of the utterance and the background knowledge of the listener. It is thus argued that no meaning is guaranteed or necessitated by the words used, which is some concession towards contextualism. Composition of meaning can be understood as an analogue of polygeny in the case of composition of powers, where multiple factors combine, usually in a non-linear way. Contextual factors can be the additive interferers for meaning. However, against a complete or ‘radical’ contextualism, the dispositionalist theory would allow that there are distinct tendencies of words and sentences towards some conventionally standard meanings rather than others. Meaning can thus be taken as a mutual manifestation of words, minds and contexts.

13. Dispositions and Ethics (with Svein Anders Noer Lie)

Our ethical notions are dispositional throughout. Moral responsibility vitally requires the dispositional modality. But such responsibility is premised on two further dispositional notions: 1. agency, because we are responsible only for the actions we have caused as agents and 2. normativity, without which there is nothing we ought or ought not to do. Agency depends on their being autonomy and intentionality, which are further dispositional notions. An agent’s acts must have both autonomy and intentions behind them to so qualify as agential. And normativity rests on value. A mutual manifestation account is offered where value is created jointly by the observer and thing observed. But both agency and normativity are dispositional in their own right also. Ethics invokes dispositional notions from top to bottom. The dispositional modality also opens up a challenge to the traditional fact-value dichotomy, which is based on a modal dualist metaphysic. While one cannot infer an ought form an is, Hume’s fork leaves out the dispositional in which so much of the ethical is to be found.

14. With Great Power Comes Great Responsibility

Moore is right that causation is not the only basis for legal responsibility within the law. Omissions may also be. A theory of powers provides a basis for both. One can be responsible for something only if one had a power to bring it about and a power to prevent it. Powers come in degrees and the degree of responsibility is proportional to the degree of causation. Omissions are the agential cases of ‘causation by absence’. Absences are nothing, however, and cause nothing. They can be explained in terms of the removal of an active power being the occasion for some effect without being its cause, for its cause was the remaining powers still operating. Moore’s argument against the transitivity of causes is challenged. Causes do not peter out over time. What counts is that the cause disposes towards the effect and it may do so over any distance in time. This gives us the same result as Moore – causes are not universally transitive – but for a different reason than he gives. The connection with counterfactuals has to be amended once a move is made to the dispositional modality.

15. The Tendential Theory of Sporting Prowess

The results of sport would not interest us if either they were necessitated or they were a matter of pure chance. And if either case were true, the playing of sport would seem to make no sense either. This poses a dilemma. But there is something between these two options, namely the dispositional modality. Sporting prowess can be understood as a disposition towards victory and sporting liabilities a disposition towards defeat. The sporting contest then pits these net prowesses against each other. The stronger will tend to beat the weaker but no more than tend. This makes sense of the sporting contest in which the weaker knows they still can win. The stronger team can lose though they don’t tend to do so. The dilemma is thus escaped. Sport contains some elements of pure chance, and it may be desirable to retain some of these. But when an athlete exercises a prowess, even though they cannot guarantee an outcome because of the possibility of interfering factors, they do retain some degree of control over the outcome. Their actions still tend towards it, which differentiates such a case from pure chance.

16. Freedom and Control

The problem of free will is a problem of modality. Causation frequently is taken to involve necessity, meaning that if an action is caused then it could not have been free. There are other options but the free will problem has been hampered by a commitment to modal dualism: that there is only necessity and pure contingency. A third modality is essential for a solution to the free will problem. If we have necessity, then things couldn’t have been otherwise and the agent has no free choice, against the Principle of Alternate Possibilities (AP). But if there is complete contingency, then the agent seems to have no control over their actions, against the principle of Ultimate Authorship (UA). Both AP and UA can be salvaged, however, if we allow both causation and agency to involve the dispositional modality. It means that alternate possibilities will be entirely ubiquitous so we have AP unproblematically. The difficulty is at the same time retaining UA. But with the dispositional modality, the agent still has powers at their disposal that tend towards intended outcomes. If they succeed in achieving them, it is right that the agent retains responsibility even though there was a possibility of the outcome being prevented. Thus, AU and AP are finally reconciled.