A Return to the Empiricist Reduction of Causes

To repeat, under the influence of empiricism the expression ‘efficient cause’ no longer came to mean ‘influence exerted by Agent on Patient on contact’ but instead came to mean more specifically ‘observable behaviour of object prior to becoming contiguous with another object’. Likewise, the effect was no longer understood as ‘change in Patient as result of the exertion of influence’ but instead as ‘observable behaviour of object after becoming contiguous with another object’. Consequently, realist empiricists that wanted to retain the notion of‘action’ as an intelligible notion were required to treat ‘action’ in terms of the precedent motion of one object (the observable correlate of influence) and ‘effect’ as the subsequent motion of another object (the observable correlate of the reception of influence). In other words, efficient causation came to be considered either as the passing over from cause to effect of some observable quantity, such as motion, or as a mere relation between motions. This idea is at the core of contemporary transmission accounts of causation (Kitcher 1989; Dowe 1992; Kistler 1998).

Arguably, whereas in philosophy the reduction of the idea that objects act on each other to the idea that causation is just a succession of observable behaviour was fairly successful, it didn’t impact as strongly on the scientific community. Just consider the following passage from James Clerk Maxwell in 1877:

The mutual action between two portions of matter receives different names according to the aspect under which it is studied, and this aspect depends on the extent of the material system which forms the subject of our attention. [...] if we confine our attention to one of the portions of matter, we see, as it were, only one side of the transaction—namely, that which affects the portion of matter under our consideration—and we call this aspect of the phenomenon, with reference to its effect, an External Force acting on that portion of matter. The other aspect of the stress is called the Reaction on the other portion of matter.


To Maxwell, influence is clearly exerted between two portions of matter, and not between events.

Everything is not always so clear-cut, though. For instance, we find that in classical mechanics the term ‘action’ is used in two different ways, one corresponding to the empiricist idea that action is the behaviour of an object prior to an interaction, and the other that action is the influence exerted by one object on another during an interaction. According to the former, ‘action’ denotes the sum of the kinetic energy of an object over an interval of time, whether or not that object actually influences anything else. This use complies with the idea that actions are to be identified with the motion of an object prior to an interaction. It is in this sense physicists may sometimes talk of pure motion as an ‘action’, but as Hertz observed:

the name ‘action’ for the integral in the text has often been condemned as unsuitable [...] these names suggest conceptions which have nothing to do with the objects they denote [in mechanics, R/]. It is difficult to see how the summation of the energies existing at different times could yield anything else than a quantity for calculation.


The other understanding of ‘action’—the way Maxwell uses the term— refers to the exertion of force that occurs on contact between objects. Basically, the first sense of ‘action’ is of the energy acquired by an object as the result of a change in its state of motion and which is then preserved in the object over time as long as nothing acts on it. Salmon would call this propagation. The second sense of‘action’ is of the influence exerted to produce a change in the state of motion of another object, which complies more clearly with the earlier standard conception of efficient cause. This is what Salmon calls production and Glennan calls constitutive production.

In any case, it is the empiricists’ refusal to deal with anything other than observable events—this holds for realist and anti-realist empiricists alike—that motivated the shift in conception of efficient causation, such that actions came to be intuitively thought of as the exertion of influence by an event on another event (of one motion on another, as it were) rather than the exertion of influence by an object on another object. In light of this development it is ironic that in physics today—as it always has, really—actions (in the latter sense of ‘action’ discussed above) are still taken to occur between bodies, not events (see Resnick, Halliday, & Krane 2002: Ch. 3), while philosophers, even those who argue in favour of the reality of forces, will insist that talk of forces acting between bodies is a derivative manner of speaking. This comes out clearly in the following passage by Bigelow, Ellis, & Pargetter, in a paper where they aim to develop a realist conception of forces as causes:

Often, we speak of forces as operating between, not events or states, but objects. On our view, this is a legitimate but derivative mode of expression. The primary relation holds between events or states; and in virtue of this primary relation there will be various derivative, indirect relations holding between the various salient individuals which are involved in the events.

(Bigelow, Ellis, & Pargetter 1988: 624)

The authors are actually claiming that physics is wrong to speak of forces operating between bodies. However, as far as I can see, philosophy is the only intellectual discipline that insists on forces operating between events rather than bodies.

There are other things to worry about in this brief passage, which relates to what I have previously discussed. For instance, they claim that the relationship holding between spatially contiguous bodies at tt is derivative of a relation that holds between the event of their becoming contiguous at tj and an ensuing event or state at t,. However, since these authors agree that causation involves the bringing into existence of effects by causes, then at tj there is no future state at t, to which the contiguous objects at tj could stand in a causal relation or which they could exert an influence on. Are they saying that the relationships between the contiguous bodies existing at t( only become causal once they have ceased to exist, notably at i2?

To my mind, the idea that the relationships holding between interacting particulars is derivative of relationships between events makes little sense unless one also assumes that events are the fundamental building blocks of reality and therefore the primary bearers of powers. On the basis of the citation above from Esfeld, seen in isolation from the rest of his writing, it would be possible to interpret him as thinking of events as the fundamental building blocks and the bearers of powers, but this would conflict with other passages in which he talks about charge and mass as powers of particles and not of the events in which those particles partake. Arguably, Esfeld resorts to talk about events exerting powers when discussing causation because it is a convenient choice in the current philosophical discourse, where events are by default treated as the relata of causation. However, he reverts back to talk of powers being the properties of things when discussing the nature of properties, because that is the received way of thinking about properties and their bearers. The most reasonable conclusion, from the perspective of substance ontology, is that events are epiphenomenal entities; they only exist in so far as substances are doing one thing or another. I’ll qualify that claim in Chapter 7.

I find it difficult to see that the idea that forces operate between events, or that events exert influence on each other, is anything other than a philosophical eccentricity that is the consequence of a temporary phase in the philosophical tradition, during which it was believed that sensory data was the only intelligible content of thoughts. As far as I can tell, none of the major schools of thought before the rise of empiricism conceived of actions occurring between events. The Aristotelians, Stoics, and Scholastics took actions to occur between two material objects whose relation is synchronous. Similarly, the Atomists, both ancient and modern, assumed that the atoms impinge on each other and thus change each other’s state of motion, and this is still a standard understanding in particle physics. In the Large Hadron Collider at CERN, they are not trying to accelerate events to smash them into other events; they try to accelerate high-energy particles to make them smash together. In chemistry the assumption is that various substances react with each other. Oxygen reacts with some fuel to combust; oxygen does not react with combustion. Furthermore, the common sense conception is that bodies act on each other: the leaden ball dropped upon a pillow acts on the pillow to make a hollow (the falling doesn’t act upon the forming of a hollow); the horse pulls the cart (not: the motion of the horse pulls the motion of the cart); the brick hits the window (not: the motion of the brick hits the breaking of the window). I find no other source of the idea that events influence other events than the reduction of the Aristotelian notion of efficient causation to what was believed to be the observable correlates of cause and effect, i.e. motion before and after two objects become contiguous. But, this means that the idea that actions occur between events clearly originates from a tradition that wanted to remove the notion of influence and substitute it with observed regularity. Causal realists should better avoid this conception of causal influence.

To illustrate fully how the idea that influence is exerted between events has infiltrated philosophy, consider that the scholastic term ‘transeünt causation’ is used today, in certain circles at least, to denote event causation and then in explicit contrast to Agent causation, in the belief that this follows the medieval use. The opposite is the case. Malebranche presented Suarez’s notion of transeünt causation as when ‘the action is in some way distinct in reality from the material cause that receives the effect’, and immanent cause as an action ‘received in the operating thing itself ... received in the same faculty by which it is elicited’ (from Oeuvres Complètes de Malebranche, translated citation from Susan Peppers-Bates 2009: 108). W.E. Johnson drew the distinction between transeunt/immanent, in the correct medieval form, as action between continuants vs. within a continuant (1924: xxiii-xxiv), as did C.D. Broad following Johnson (1968: 78), and also Dorothy Emmet who claims that medieval philosophers derived the distinction from Aristotle (1985: 77). Emmett traces the origin of the misconception about transeünt causation to Chisholm’s decision to use the term to denote event causation in contrast with agent causation (Chisholm 1966). To be fair, the blame is on his readers, because Chisholm admits that he may not be using the term in the original medieval sense. However, today, due to Chisholm’s influence, transeünt causation is often understood as event causation.

According to the standard view, while a cause can in some way be considered an event, the event in question is not itself an entity that exerts an action on the effect; the event is the exertion of influence by an Agent on a Patient. In fact, it is not really possible to conceive of the efficient cause as an entity distinct from the material cause, nor of them as wholly distinct from the formal cause, because the efficient cause is the influence exerted by one material object on another, and the characteristics of this influence are determined by the powers of the objects involved; it all comes together as one unity in an interaction, every component dependent on the other. It is in this sense that the standard view is a total cause conception; any effect is always the product of the nexus of efficient, material, and formal causes. Consequently, the necessary connection—if there is one—should be looked for in the connection between the total cause and its product, not between the efficient cause alone and the effect. Indeed, that is exactly where Hobbes thinks a necessary connection is to be found:

an entire cause, is the aggregate of all the accidents both of the Agents how many soever they be, and of the Patient, put together; which when they are all supposed to be present, it cannot be understood but that the effect is produced at the same instant; and if any one of them be wanting, it cannot be understood but that the effect is not produced.

(Hobbes 1656: Ch. X, § 3)

The conclusion is that the standard view characterises influence as occurring between objects, not events, and the necessitating cause as the sum total of causal factors present in an interaction between an Agent on a Patient. There is, however, a very serious problem with this view, first clearly identified by Bunge (1979) and later elaborated on by myself (Ingthorsson 2002, 2007, 2019), notably that it is in a certain sense incompatible with modern physics. This problem is the subject of the next chapter.

< Prev   CONTENTS   Source   Next >