In William Devlin’s “Salmon and Causality: The Problem of Arbitrary Selection,” he nicely explains Salmon’s theory of causal processes, and he lucidly explains how there are apparent problems with Salmon’s theory. Devlin shows that if we follow Salmon’s theory it leads to the postulation that causation is arbitrary.

I must admit at the outset of this commentary that I am a presentist when it comes to the ontology of time, and thus I have a natural proclivity for arguing against any theories of non-simultaneous (durationless) causation, such as Salmon’s, since in the presentist’s universe there are only present moments, and thus no non-identical moments of time that are connected across a duration—such as is the case with non-simultaneous causal connections between events (that are alleged to exist by so many metaphysicians and other philosophers and scientists).

I enjoyed Devlin’s article, but there is an apparent problem with his attack on Salmon: if it can be shown that causal relations and causal processes do not exist, then there is no need to show that Salmon’s position reverts from causal process back to causal connection as Devlin has shown, since both causal connections and causal processes do not exist in the first place.

An Argument for the Impossibility of Salmon’s Causal Processes (and Russell’s Causal Lines)

Devlin shows how Salmon’s theory explains causation in terms of worldlines rather than spacetime points. Events are spacetime points (spatially and temporally unextended), but causal processes are collections of points (1-dimensional). It is important to note that Salmon apparently considers his processes as atomic (partless) and irreducible:

 

It is my conviction that… a radically different notion [of causation] should be developed… One of the fundamental changes which I propose in approaching causality is to take processes rather than events as basic entities. (Salmon 1993, 154-155) (Emphasis added.)

 

I see no other choice but to take Salmon literally on his use of the word “basic” here, and thus to consider processes as being irreducible, extended items, and thereby some sort of extended simple (see McDaniel 2002) which cannot be cut into parts. Therefore, a worldline is an irreducible causal line that exists in a spacetime manifold (or is an irreducible piece of spacetime). 

            I will next discuss that this appears to involve problems for both Devlin and Salmon. An extended causal process (causal line), P_C, can only be considered at the entire spacetime manifold that it is located is at, call it p₁p₂p₃. On this account, P_C cannot be considered at a part (sub-location) of p₁p₂p₃, such as the basic (pointy) locations, p₁, p₂, or p₃, or the non-basic (non-pointy) sub-locations p₁p₂, or p₂p₃. On this account, the statement,

 

“P_C is located at p₁p₂p₃,”

 

is true, and the statements about P_C being at any non-basic subspace of p₁p₂p₃ (i.e., space p₁p₂, or space p₂p₃), or at the individual basic subspaces of p₁p₂p₃ (locations p₁, p₂, and p₃), are all false, such as the statements,

 

“P_C is located at p₁”,

 

“P_C is located at p₂”, or

 

“P_C is located at p₃,” and so on.

 

I will next show that there is a problem here in considering a simple (partless) entity (P_C) being located at a non-simple spacetime manifold. Put in similar words: there is a problem with the idea that there can be a non-basic manifold containing irreducible, noncomplex (partless) P_C. (I will later show that if P_C is a piece of the spacetime manifold rather than an entity that occupies spacetime, there are also serious problems with that account of P_C also.) The problem is so serious that it apparently shows that causal relations and causal processes do not exist, contrary to what is believed by the many philosophers who hold that they do.

If P_C is partless (noncomplex), it is a single entity. If P_C is describable by a statement, since it is partless then the entirety of P_C is describable by the statement. For example, if P_C is located at (at least) two topological locations, such as p₁ and p₃, P_C would be describable by the statements, “located at p₁”, and, “located at p₃”. If P_C is located at p₃, and if p₁≠p₃, then by being at p₃, P_C is describable by the statement, “not located at p₁”, since the topological location p₃ is occupied by something since P_C occupies p₁p₂p₃, and this something is not the same something that is at p₁. This could be said of any location that P_C occupies that is not p₁. If P_C occupies more than two locations, and for that reason is located at three locations, p₁, p₂, and p₃, then at locations p₂ and p₃ P_C would be describable by the statement, “not located at p₁”. This apparently involves contradiction, however, since the relation is one, partless entity, and if it is “located at p₁”, and if it is “not located at p₁”, each of these statements must describe the entirety of P_C, and that shows that P_C is impossible. It appears that I need not go further to show that there are no causal processes or causal relations across a non-simultaneous topological region, but since there may be many objections, I will further clarify my points.  

P_C is at p₁p₂p₃, but aspects of P_C at p₁, p₂, or p₃ cannot be discussed, since there are no such aspects of P_C that are not identical to the whole of P_C and which are located only at  p₁, p₂, or p₃ (or at p₁p₂ or p₂p₃ ). Nevertheless, it seems that it can only be the case that the individual basic building blocks of p₁p₂p₃ (p₁, p₂, or p₃) can only be occupied by something to do with P_C. By this I merely mean that when we consider p₁, p₂, or p₃, individually (or p₁p₂ and p₂p₃ individually), we apparently can only conclude that they are not unoccupied with (respect to P_C). The reason that p₁, p₂, or p₃ must be considered to be occupied by something to do with P_C is because the entirety of p₁p₂p₃ that P_C is at is a topological region that is made up of more fundamental spacetime locations, and if P_C is at a non-basic location (such as p₁p₂p₃) and accordingly occupies the entire manifold, it must also be the case that the item occupying p₁p₂p₃ results in each of the more basic locations (p₁, p₂, p_3, p₁p₂, or p₂p₃) that make up p₁p₂p₃ also being occupied.

A location would not be occupied at all if none of its sub-locations that compose it were occupied. In other words, if P_C occupies a non-basic topological region (p₁p₂p₃) but does not occupy the more fundamental non-basic locations p₁p₂ or p₂p₃, or p₁, p₂, p₃ of p₁p₂p₃, then P_C does not occupy the entire location (p₁p₂p₃). For these reasons, P_C‘s being at p₁p₂p₃ must also lead to all of the basic and non-basic sub-locations of p₁p₂p₃ being occupied. But this poses a serious problem for P_C, since it cannot have aspects at sub-locations of p₁p₂p₃ that are not identical to the whole of P_C.

The argumentation given to this point need not apply only to P_C. If one postulated that there are non-complex causal relations between events that are relations that are temporally and spatially extended between two non-simultaneous events, the reasoning above would also show fatal problems to do with such relations since the reasoning above, in the end, really shows that there can be no noncomplex connection of any sort between any two items that are (believed to be) at a temporal or spatial distance from one another. If one imagines that an unextended connection between events (a connection located at non-simultaneous events, but not between them) is possible, the reasoning above also applies to them, since the argumentation only requires that the items connected across time and/or space be connected by a simple relation. It appears that the reasoning above shows that there are fatal problems with any connection between any two objects not collocated in space or time, which would appear to vindicate an extreme form of nominalism, an extreme form of conceptualism, or a blob theory (reality is structureless).

The reasoning given above, where non-basic regions were discussed as being composed of basic sub-regions and non-basic sub-regions, holds for any non-basic topological region, since any non-basic topological region is made up of more fundamental topological regions. If it were the case that a non-basic region, such as p₁p₂p₃, were not made up of more fundamental regions, then an extended and non-basic spacetime manifold would not be made up of anything, and it would not be a topological region at all. For these reasons, a non-basic topological region can only be composed of more fundamental topological regions, and/or of basic atomic topological locations, and a process occupying a non-basic topological region must accordingly result in the more fundamental or basic locations also being occupied. P_C, for these reasons, cannot be located at p₁p₂p₃, since it cannot be located at any of the sub-regions that make up p₁p₂p₃. This appears to be a fatal problem for the coherence of P_C: no sub-regions of the P_C‘s entire spatial location (p₁p₂p₃) can have anything to do with P_C, and for that reason, the P_C cannot be a spatially located entity at all, which is impossible. It appears that the basic structure and features of this argument just given to do with P_C’s inability to be located at a complex spacetime region, is merely an issue to do with the way a partless item has a size of some sort, and has different regions that do contrary things. Therefore, this argument can be applied to any item that is theorized to have a size, but no parts, such as a Democritean atom, a Planck cell, a Planck unit of time, and so on.

Lastly, if one wishes to maintain that P_C can merely be a part of spacetime, rather than occupied in space time, that is of no avail, for the following reasons. Spacetime is composed of spacetime points, and it is considered to be a mereological whole only because of the topological metrical relation, typically called connectedness, alleged by mathematicians and philosophers to connect the spacetime points. If this is the case, then P_C itself would be a topological manifold, and thus it would be composed of spacetime points and the metric relation that holds them. But note that if the relation is basic, then it in fact occupies more than one spacetime point due to the fact that it coincides with the spacetime points that it interrelates (the relation must be where the points are in order to ‘interact’ with them and therein to interconnect them). Therefore, P_C, if it is itself a piece of the spacetime manifold (or is any sort of topological continuum), is subject to the argumentation given above in this commentary.

 

Works Cited

Grupp, Jeffrey, 2005. "The R-Theory of Time, or Replacement Presentism: The Buddhist Philosophy of Time." The Indian International Journal of Buddhist Studies (IIJBS). No. 6. Pages 51-122.

 

McDaniel, Kris, 2002, "Extended Simples, Shape, and Space" at the 2^nd Annual Princeton-Rutgers Graduate Student Philosophy Conference.

 

Phillips, Stephen H., 1995, Classical Indian Metaphysics, Open Court, Chicago.

Pines, Shlomo. 1997. Studies in Islamic Atomism. Jerusalem: The Magnes Press.

 

Salmon, Wesley, “Causailty: Production and Propagation”. In Sosa and Tooley, 1993, Causation, Oxford: New York. 154-171.