From the very beginnings of organised philosophical thought there has been a keen awareness of the problem of causality. In its most basic form this problem arises whenever the concept of freedom is considered. To be free is to be uncaused. This basic axiom has considerable implications. If we agree that ‘to be free is to be uncaused’ then it seems like we face two simple options as implications of this axiom. Either we deny that there is any such thing as freedom or we deny that cauality is universal. This simple axiom, that to be free is to be uncaused, produces a quite strange and difficult tension between these implications. On the one hand we might want to affirm freedom as real and present. In general we might want to go further and not only affirm the actual reality of freedom but also affirm it as something to be valued and retrieved in the face of its removal. On the other hand we want to affirm the capacity to know the world and the conection between the various facts of the world, connections of causality. Freedom breaks open the world, whilst causality constructs its’ connections and it would seem we want the strange mixture of freedom within a connected world. One of the basic responses to this type of tension has been to suggest that there are, in fact, two types of causes. Everything is thus connected through the concept of cause, but the connections are variable dependent on which of the two types of cause is in play. We thus reconcile freedom and cause by turning freedom into a type of causal force.
Like many problems in philosophy, however, a solution to one type of problem produces contradictions and confusions with regard another. It is difficult to take a position on a single problem without producing implications for other problems. Concepts never stand alone but weave into a complex web and pulling a strand at one end produces ripples throughout the arrangement. Sometimes these ripples merely reallign particular implications, make us shift our position slightly in areas that were not of immediate concern. At other times, however, these ripples rend another part of the web that we wanted to maintain and we find ourselves trying to put back together a whole that is both coherent and conducive to life. This attempt to make our thoughts and concepts coeherent lies at the heart of philosophical activity.
The attempt to answer the problem of freedom and causality by responding with a doctrine of causal types, by claiming that freedom is a type of cause, produces exactly the type of rending of the conceptual web that is incoherent. The first thing, then, is to make clear quite what the incoherence is. What’s the problem?
To begin, let’s assume the doctrine of causal types as valid. This, to be clear, posits two types of cause – the first is the basic physical causation we might associate with the reality of the world we live within, most of which we now assume to be best known through scientific methodologies of observation and experimentation; the second is the basic intentional causation we might associate with an agent or actor within the world, usually human, though not necessarily limited to that species, often expressed in terms of desires or attitudes or beliefs.
With this typology in place, albeit in a rough and ready way that allows for plenty of refinement no doubt, we are able to categorise various connections of events. Some events will be connected by virtue of physical causation, others by virtue of intentional causation – the former will be necessary, predictable, repeatable, universal and consistently identical mutatis mutandi, the latter will be free, unpredictable, unrepeatable, singular and different. The typology allows us to have our cake and eat it. As long as we maintain this basic typology we can now approach intentional causation and acknowledge that it often displays itself differently. We can acknowledge freedom in the face of human behaviour that is often remarkably predictable. It might then be argued that predictability and regularity of human behaviour is merely an appearance, capable of intricate and detailed analysis but always limited by the nature of the causality that underlies the essence of intentional agents and actors. The argument might continue by pointing out that no matter how much we might predict a humans response the possibility of an error is ineliminable in principle precisely because the very causal structure is not physical but intentional. The agent is always free to act unpredictably because the agent is always, fundamentally, beyond the very capacity of prediction. Even if we don’t act differently, we always could act differently.
This categorisation of events into two types, one of which can be understood in such a way as to allow it to be both causal and free, can be found from Plato through to Davidson. The two types of cause in Plato reconfigure as event and agent causation within one stream of contemporary philosophy1. Where, then, might there be a sense of incoherence arising from this causal typology?
One of the core advantages of this categorisation is that enables us to complicate our understand of the nature of necessity within these different realms. It enables a typology of necessity to be drawn, almost as a by-product of the typology of causes. If there are different types of cause, then we might find that to each type of cause there applies a type of necessity. We might then be able to make sense of how the physical or evental cause is a kind of necessity and operates according to a law-like predictability whilst the intentional or agental cause is a kind of necessity that does not have those characteristics. Of course it might immediately be questioned whether there is not some confusion and obfuscation at play. Someone might ask how a cause can be a cause without it also and simultaneously being necessary. Surely a cause must produce its effect or else it is not a cause. How can we have an agent causation if we allow that causation to be an ambiguous ‘might’ rather than an imposing ‘must’?
It is here that the incoherence arises. Necessity, as a concept, is at least as fundamental and basic as cause. To determine the content of our concept of necessity simply by first determining the content of our concept of causality is liable to lead to troubles if we are not very careful. Determing any concept from another is to make it dependent on that concept. The concept of goal or end, for example, is easily determinable within the realm of a chess game because the content of that game determines the content of the goal. If we were to allow this to become a general content to the concept of goal, however, we’d obviously end up in absurdities – the goal of a University degree, for example, is not ‘check-mate’. We might want to claim that the concepts of ‘necessity’ and ’cause’ are equally basic, unlike the concepts of ‘chess’ and ‘goal’ but even an agreement on this point does not imply that the one is capable of determing the other without question.
In fact the connection between necessity and cause that has developed within philosophy actually paints a very peculiar picture and one quite at odds with the assumption that necessity lies in the realm of the physical, in the realm of event causation. It might seem to the uninitiated that it is obvious that events, with their law-like behaviour, repeatably producing the same effects, offer us the fundamental model of necessity through their causal behaviour. Physical necessity would in this situation be the paradigm of any necessity. Yet philosophical argument can easily and quickly be produced that puts this conclusion into severe doubt. The nature of observation and experiment is such as to always be limited, in principle, to a finite set of cases. This is because we are finite entities, only ever capable of observing a limited, finite set of cases. We never, as Hume points out, observe necessity in the connection between cause and effect, simply conjunction between the cause and the effect. Necessity, on this argument, cannot possibly lie within the physical – necessity is a structure imposed upon the observations, not one deriving from those same observations. We organise our observations, we order our experimental results and we infer from these data-sets laws and necessities. It is not the observation that produces the law but the inference.
Necessity cannot be found within the physical, within the event, according to the Humean empiricist. Necessity is fundamentally a feature of the agents interaction with the observation, in particular their correct application of inferential structures. Neceessity is fundamentally normative. It is produced by the norms of correct or incorrect inference. These norms, if correctly applied, enable us to infer necessities such as physical laws. Of course, this also means we might be going astray and incorrectly inferring particular laws from specific observations. If, however, our inferences are correct then the necessity of the laws is as necessary as it can be. Yet how necessary can such inferences be? If necessity is produced normatively, is this production itself necessary? In other words, if I argue and infer correctly, can I produce a necessity? If this were the case, if normativity could produce necessity, then correct inferences could produce claims that were simply incapable of being denied. If normative inferential structures could produce necessity then such necessity would surely be unavoidable? Put more bluntly, if an inference is accurate and correct, nobody could deny it. Yet this would immediately produce a problem, since surely such accuracy of inference would, in producing an inevitable and necessary agreement, demolish the freedom we previously located within the agent? Responding to reason as though we could deny it is an illusion of freedom and yet without this illusion reason tastes totalitarian.
1See ‘Agency and Necessity’