Relationships of Absence
A Metaphysics Essay
April 2, 2007
Acknowledgment: The author previously submitted an earlier version of this essay for course work in metaphysics during Fall 2006. As a result, JeeLoo Liu’s comments integrally affected the production of this text.
Causation lies at the foundation of metaphysical theory. Theorists must manage the commitments that a causal theory entails, and defend them. The problem of causation by absence presents an interesting dilemma for metaphysicians. How a causal theory deals with absences affects our willingness to accept them. I review two approaches to causation by absence. First, I review Beebee’s analysis of causation by absence, and conclude that her standpoint entails a commitment to uninstantiated universals. Second, I review Lewis’ analysis of causation by absence, and conclude that his perspective involves a faulty affirmation of how absences function. Beebee’s commitment to uninstantiated universals presents less of a problem than Lewis’s theory of causation. We should side with Beebee: absences cannot play any role in a causal relation.
The basic problem of causation by absence is a disagreement over whether absences have causal power. In short, absences are really nothing: they are nonevents, and nonentities. It seems counterintuitive that absences have causal power, but we talk about absences as if they do have that power. Researchers in space must wear space suits because they believe that space causes death. Law details many cases where a person is responsible for disregarding positive duties or violating negative duties. Causation by absence is a serious issue.
Beebee utilizes causal explanation to solve the problems surrounding causation by absence. In using intuition to test various metaphysical accounts of causation, Beebee argues that the winning argument lies in distinguishing between causal explanation and causal relations. Beebee concludes that while there is no such thing as causation by absence, all our examples of causation by absence are actually causal explanations. Beebee considers relational and nonrelational theories, concludes they neither framework is adequate, and then proposes a way to differentiate between genuine causes and dubious causal absences. As we will see, Beebee’s framework defends Platonic realism.
Among these theories that Beebee considers is Lewis’ counterfactual analysis. Beebee maintains that our intuitions compel us to accept certain causes from absence and reject others. However, a pure counterfactual analysis forces the philosopher to treat all absences the same. Consider Flora’s habit of watering her neighbor’s orchids. Flora usually waters her neighbor’s orchids, but for a time, Flora stopped. Unfortunately, no one else watered the plants and they died. Our intuitions indicate that the absence of Flora’s watering caused the plants to die. We also believe that other neighbors do not have the same causal power as Flora. Intuition can differentiate between Flora and everyone else, but a pure counterfactual analysis of causation by absence cannot discriminate between causes that we would accept and causes that are irrelevant (Beebee, 2004).
Hart and Honoré’s book Causation in the Law presents another theory that may repair this problem. Beebee indicates that “what makes us single out omission as a cause but not another is abnormality” (2004, p. 295). Beebee’s objection against the abnormality standard comes from Stapleton: “There are many cases in which the abnormality criterion fails to explain our commonsense causal judgments” (Beebee, 2004, p. 295). These cases rely on norms of the abnormality standard, which require historical9 examples. In first-time examples, we cannot compare the occurrence to a historical regularity. Like Hume, the abnormality standard cannot answer Stapleton’s objection. Although our intuitions help us decide whether first-time causal absences can cause events, the abnormality standard cannot appeal to intuition (Beebee, 2004).
Another possible solution to the problem of causation by absence lies in the attempt to repair the nonrelationist account. Nonrelationists can appeal to Lewis’s discussion of possible worlds. Counterfactual dependence relies on possible worlds that match our world in every way except for the negated antecedent and consequent of a causal conditional. Beebee extends this approach by indicating that a possible solution for nonrelationists is to require that an absence may only qualify as a legitimate cause of an effect if that absent-event occurs “at a world that is reasonably close to the actual world” (Beebee, 2004, p. 298).
Beebee thinks that this proposal also fails. She argues, “Norms are doing the work by themselves, as it were, rather than merely informing our judgments about closeness of worlds” (2004, p. 299). The failure of the reasonably-close-world standard to provide a measurement of “reasonably close” forces us to fall back on our existing norms. As it stands, the reasonably-close-world standard still relies on our intuitions.
Beebee’s solution is to introduce a new category of pseudo-causation: causal explanation. This category does not meet the standards of genuine causation, and does not imply causation. Causal explanation is a new category for pseudo-accounts of causation. To make this work, Beebee needs to distinguish causal explanation from causation. Beebee quotes Lewis, saying, “To explain an event is to provide some information about its causal history” (cited by Beebee, 2004, p. 302). The “causal history of an event, he says, is ‘a relational structure’” (Beebee, 2004, p. 302). Since explanation does not identify causes, explanation is different from a relation between cause and effect.
The only task left for Beebee is to separate examples of causal explanation from causation. One source of difficulty in differentiation lies in the apparent ability of some sentences to switch between the form “A causes B” and the form “A because B,” which Beebee says is unacceptable in metaphysics. Explanation and causation should be mutually exclusive, so there should not be cases where a proposition could appear as both forms. The chief method of differentiating between causal explanation and causation involves forcing them to refer to two mutually exclusive types of statements. We can accomplish this by “saying that an event of such-and-such kind occurred, rather than that some particular event occurred” (Beebee, 2004, p. 302). This standard makes two restrictions. First, this standard restricts causal explanation to referring to universals. Second, this standard restricts causation to referring to particulars. Therefore, an example of a causal explanation would be, “The glass broke because glass is fragile.” In addition, an example of a causal relation would be, “The molecular structure of the glass caused it to break” (Beebee, 2004).
There are problems with this, since it does not satisfy the demand for a standard between explanation and causation in a non-Platonic world. Beebee’s analysis shows that causal explanations deal with universals and causal relations deal with particulars. When it comes to causation by absence, Beebee denies that causal relations link an absent cause and effect. Causation by absence occurs within causal explanations, but absences cannot appear in genuine causal relations. According to this standard, absences referred to by causal explanations must be universals, but they will never appear as particulars in causal statements. Absences detailed as causal explanations will never become instantiated by corresponding causal relations, since there are no particular causes and effects to make such a proposition possible. In other types of causation, ones that do not feature absences, causal explanations eventually give rise to causal relations when we deal with particular events. For example, fragility is a universal and plays a role in explaining why glass breaks. The universal of fragility becomes instantiated by a particular molecular structure. Unfortunately, causation by absence appears to be a case where a universal never becomes instantiated. Therefore, if Beebee intends to utilize causal explanation as a universal form of causation, then she must defend the existence of Platonic universals, because absences are truly uninstantiated universals.
I now turn my attention to Lewis’ theory of causation. Lewis’ theory contrasts competitively with Beebee’s theory of causal explanation. Lewis makes the opposite basic conclusion regarding the causal power of absences. He says that causation by absence occurs by following a nonrelational account of causation.
Lewis builds his case for causation by absence by first objecting to Menzies’ account of causation. Menzies’ most crucial position is that a causal relation is an intrinsic relation between cause and effect. This defines a set of philosophical commitments that Lewis does not want to accept wholly. Lewis, as we will see, advocates a nonrelational account of causation because of his commitment to using counterfactuals for analysis. Menzies, on the other hand, prefers a relational account of causation, and defends the commitments that the relational account entails. One of these commitments forces Menzies to derive the causal role from the relation between particular events. He accomplishes this by analyzing the instantiated relation for a given cause and effect. If the cause and effect are instantiated, then Menzies can derive the relation between cause and effect from the universal that instantiates that relation. This derivation allows Menzies to maintain that causation is an intrinsic relation between cause and effect (Lewis, 2004).
Lewis says that Menzies’ account fails to account for causation by absence. Lewis reminds us that relations require relata, and any juxtaposition that fails to produce two events/objects/entities also fails to meet the demand for relata. A relation between an event and a nonentity cannot be a relation. Since absences are nothing and provide no point of reference, absences cannot be a constituent of a relation. A causal relation is still a relation, and as such, causal relations cannot involve absences. According to Lewis, “Any relation needs relata, whether it is intrinsic or not” (2004, p. 282). Lewis argues against Menzies by using causation by absence as a counterexample. Any example of causation that involves absences cannot be relational, which presents a problem for Menzies’ relational account. Lewis wants absences to have causal power, so Menzies’ theory is inadequate (Lewis, 2004).
Lewis must find some way to analyze causation without relying on a relational analysis. A counterfactual analysis is not relational, and Lewis uses counterfactuals as the basis for analyzing the causal power of absences. Lewis says, “The counterfactual analysis escapes the problem because, when the relata go missing, it can do without any causal relation” (2004, p. 283).
How does one build a counterfactual that deals with absences? When it comes to negating an absence in a counterfactual, we should not “attend to some remarkable entity and suppose that it does not exist. Rather, we need only suppose that some unremarkable entity does exist” (Lewis, 2004, p. 283). The difference between “attending to a remarkable entity” and “attending to an unremarkable entity” is subtle, but necessary in order to capture the spirit of how counterfactuals avoid dealing with relata. Lewis says that “we do not have to reify the void in order to ask what would have happened if the void had not been there” (2004, p. 282). In essence, Lewis says that negating an absence allows us to suppose that anything might have happened instead. Lewis wants us to construct the counterfactual that deals with absences by not repeating the absence in the counterfactual. Consider Lewis’ example: ‘If I were exposed to a vacuum, then I would be killed.’ Lewis does not want the counterfactual to be: ‘If I were not exposed to a vacuum, that I would not have been killed.’ Instead, the counterfactual should be: ‘If I were surrounded by protective objects (like a space suit), then I would survive.’
This is all well and good. If we abide by this guide for constructing counterfactuals, then counterfactuals will allow us to analyze the causal power of absences. Lewis details the example of such a counterfactual that obeys these guidelines. “The void causes death to one who is cast into it because if, instead, he had been surrounded by suitable objects, he would not have died” (Lewis, p. 282). This example utilizes a counterfactual, and it constructs the negation in the form of attending to the presence of an unremarkable entity (a space suit). This example also avoids Lewis’ lack-of-relata objection, because it does not utilize a relational analysis.
Ignore Beebee’s external objection for a moment. A counterfactual analysis would work for distilling the causal power of absences as long as Lewis stays away from relations. Our assumption in utilizing counterfactuals to analyze causation is that causation does not necessitate a relation. Absences, it would seem, are the examples that support this notion. However, Lewis does not content himself with treating the causal power of absences as a non-relation.
Lewis returns to Menzies’ analysis of causation and accepts that Menzies’ intrinsic relation does apply to regular forms of causation. Lewis invents a new word, ‘biff,’ and defines that word in terms of Menzies’ account of causation (A ‘biffs’ B). Then, Lewis incorporates biff into his own theory of causation. Whenever an instance of causation occurs, then Menzies’ analysis (biff) applies. Whenever an instance of causation by absence occurs, Lewis applies a counterfactual, and then biff applies.
Lewis argues that “causation by absence is not an instance of biff. Nevertheless it can be described in terms of biff.” (2004, p. 284). A causal absence is an absence of biff, and Lewis moves toward incorporating biff with his non-relational account of causation. For each example that involves an absence, Lewis constructs the counterfactual in a way that makes the counterfactual not mention the absence. Consider the counterfactual ‘If I were surrounded by suitable objects, I would survive being in a vacuum.’ The original situation, me being exposed to a vacuum, deals with the power of an absence (the vacuum), but Lewis constructs the counterfactual in a way that masks the absence. Lewis argues that once a causal conditional ceases to mention an absence, then biff is usable.
It is here that Lewis gets into trouble. Lewis uses biff as a negative replacement for an absence, and then pretends that biff will behave counterfactually similar to positive relata. This becomes evident when Lewis starts speaking of biff-relations and different varieties of causation, all of which involve causal relations, but should be impossible due to absences.
If we look at counterfactuals closely, we notice that many of them deal with absences. Recall the earlier example of the vacuum: we discover through a counterfactual that a vacuum causes death because if a person had been surrounded by suitable objects (e.g., a space suit), then that person would survive. The absence in that example is the vacuum. The absence does not occur in the counterfactual; it transforms into a space suit. However, if we consider ordinary examples, the absence occurs in the counterfactual. Consider Descartes’ example: if I exposed wax to heat, then it would melt. According to a counterfactual account, we know this is causal because if I did not expose the wax to heat, then it would not have melted. The absence exists in the counterfactual as the absence of heat. In the wax example, the counterfactual does not relate cause and effect.
Careful choosing between an original causal statement and its counterfactual for a relation does not work. We should look at the original statement, not the counterfactual, for the relation between cause and effect. If I apply heat to the wax, then the wax melts. We are looking for a relationship between my application of heat to the wax and its melting. We are not looking for a relationship between my not applying heat to the wax and its not melting. When it comes to causation by absence, we look to the original statement for a relation, not the counterfactual. Consider again the space example. If I were exposed to a vacuum, I would die. We are looking for a relationship between my exposure to a vacuum and my death. We are not looking for a relationship between my presence in a space suit and my continued survival.
Lewis may be content with treating biff and counterfactuals as sufficient for dealing with absences for a special reason. Lewis says, “Absences are spooky things, and we’d do best not to take them seriously” (2004, p. 283). This treatment of absences out of deference to its “spooky” quality hardly justifies giving causal relations back to absences. Lewis had the correct approach by first recognizing that absences may have causal power, so long as that power is nonrelational. When Lewis approaches biff as a way to return absences to causal relations, he runs into conceptual problems. Absences cannot enter into relations, and we cannot use biff as an intellectual shortcut around that. If biff involves a relation, then the existence of a relation rules out its compatibility with absences.
Going back to Beebee: her argument is complex, not only in the sense that it is sophisticated, but because Beebee’s arguments exhibit varying levels of intellectual commitment. Beebee’s has a certain level of commitment to denying the causal power of absences. However, that level of commitment is much higher than admitting that absences have explanatory power. Causal explanation is an answer to a challenge for clarity, so if there is a problem with Beebee’s integration of absences with explanation and causation, then Beebee can sever that commitment to causal explanation without abandoning the more significant claim that absences do not have causal power. There may be a more adequate explanation for absences than causal explanation, and that hope allows Beebee to maintain her basic position on the causal power of absences.
Beebee and Lewis leave us with deciding whether absences have causal power. Lewis says counterfactuals demonstrate that absences have causal power. Beebee argues that absences can fill the role of causal explanations, but cannot be particular causes. Beebee and Lewis both run into difficulties when we analyze their philosophical commitments, but Lewis appears to have more trouble, since he departs from his own philosophical commitment to the relata-requirement of relations. Beebee can sever her commitment to the difference between causal explanations and causation. For instance, we can reject Beebee’s differentiation and still agree that absences do not have causal power. In order to repair Lewis’ arguments, we would have to dissolve the combination of Menzies’ account and counterfactuals. If we discard this combination and simply utilize counterfactuals, then we have to address Beebee’s objection to Lewis’ theory. If we had to choose whether absences have causal power, Beebee’s philosophical position is preferable for its relative coherence.
Beebee, H. (2004). Causing and nothingness. In J. Collins, N. Hall, & L. A. Paul (Eds.), Causation and counterfactuals (pp. 291-308). MIT Press.
Lewis, D. (2004). Void and object. In J. Collins, N. Hall, & L. A. Paul (Eds.), Causation and counterfactuals (pp. 277-290). MIT Press.