Abstract
Professors Grimm, Bennett, and Vollrath have defended a particular individuation of actions in the same way1 —namely, (i) identifying doing-C by A-doing with, simply doing-A (which we might label the Anscombe-Davidson thesis) by (ii) explaining how such a doing-of-A can become, or change into, doing-C. That is, recognizing that an event can change after its occurrence—what I would deem “grandfather change”, the acquisition of a “delayed characteristic” (Bennett) or a “posthumous property” (Vollrath)— provides a defense of the Anscombe-Davidson action identity thesis. In this chapter, I will primarily consider Grimm’s work, since it is the most sustained and detailed defense of the eventchange maneuver. My criticisms apply equally well, of course, to Bennett and Vollrath.
An earlier version of this chapter was read to the American Philosophical Association, San Diego, California, March 23, 1979.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
Notes
Grimm, 1977; Bennett, 1973; Vollrath, 1975.
I disagree, however, with M. Brand’s (1977) conclusion that all events necessarily occurring (occurring in all possible worlds) in exactly the same spatial-temporal regions are identical. This position strikes me as making typical events instances of too many event-kinds. In this paper, however, I say little that provides an argument against Brand.
P. 221. Also, see the paragraph beginning “Suppose Boone shot...” (p. 224, left column). I agree with all of what Grimm says in it, except what he concludes in the last sentence. Also on p. 224 (right column), Grimm says “Boone killed the bear if and only if Boone caused the bear’s death by doing something or other of a kind appropriate to killing. In this case he shot him. Thus, since Boone’s shooting the bear became the cause of the bear’s death at tj [right!], it became Boone’s killing the bear at tj [wrong!].” (My inserts!) The ‘appropriateness’ is not (nor does it become) identity.
I distinguish (and this whole chapter provides a fuller defense for it) between the event Boone is agent of or responsible for—such as a shooting—and the event of his being the agent of it. Both such events can be suitably referred to by different uses of a phrase like “Daniel’s shooting”. Since I am more concerned herein with applying such distinctions to killings and “killing”s than to shootings and “shooting”s, I shall refrain from insisting on the distinction for “Daniel’s shooting” below (and for related phrases like “the shooting”, “Leo’s being shot”, etc.). Most of the time below when I refer to Daniel’s shooting (or Leo’s being shot) I do mean to refer to the event itself (the ‘being shot’, the shooting) aside from who is responsible for (agent of) it. In particular, it is this event (the shooting itself, whether or not Daniel is the agent of it) that non-agentively causes death—as in (3) below, for example. It may appear that my disagreement rests on a particularly refined point of grammar which, even if it is correct, has no philosophical importance. If becoming a causing can be plausibly identified (even against the grammatical grain) with becoming a cause, then an ontological reduction may be obtained—e.g., no additional event other than the eventual cause for an effect may be required and, thereby, one source for the multiplication of events (as in Goldman, 1970, Thomson, 1971, Thalberg, 1975, and others) is removed. (For non-philosophical reasons for wanting very clear conditions for events and actions, cf. p. 263 of Beardsley, 1975.) On the contrary, I claim, the perhaps refined point does make a philosophical difference. To identify (for philosophical reasons) becoming-a-cause with becoming-a-causing will be based on an equivocation and a hoped-for ontological economy will be obtained only at the price of confusion (not to mention being unfair to events).
At the very least (4) entails (5), but there could be some question about whether (5) entails (4). For example, if the meaning of (5) is broader than that of (4)—there are more kinds of causing-to-dies than there are killings, say—then (5) entailing (4) would have to be defended on other than narrowly semantic grounds. However, notice that even if (4) and (5) express distinct propositions that are not even necessarily equivalent, still the event reported in a use of (4) might be the very same event reported in a use of (4); i.e., event identity does not require propositional identity (identity of seemingly different propositions reporting the events). This fact is largely what’s behind the claim that an event can be an instance of more than one event-kind.
The terminology for propositions permits me to entertain more clearly, if not uncontroversially, certain things—e.g., that (4) and (5) on certain typical uses of them express the same proposition, i.e., P(4) = P(5).
Both Grimm and I are primarily considering particular, individual events—events as individuals, as contrasted with event-kinds of which the individuals are cases. Our usage contrasts, for example, with Chisholm’s. Chisholm’s notion of ‘event’, one species of ‘states of affairs’, would seem to be what I call an event-kind; cf. his 1976 and 1977. For related discussion on propositions, events, and facts, see my 1981 [Chapter 5 above].
As in, for example, Rosenberg (1977) and Stern (1978), as versus Beardsley (1975) and Mackie (1974).
To take an analogous case, Mary might be the child of Daniel, so that there was an event of Daniel becoming the father of Mary. But only metaphorically could we say that Mary is the child of Daniel’s fathering her. Daniel’s fathering her might produce her, but the relation between Daniel and Mary is one thing—viz., being or becoming father to her (her being or becoming his child)—and the relation between the event of his fathering her and her is another —viz., not itself fathering, but one of containment (containment of some kind or other—obviously not, of course, class membership or whole-part). Mary is a part or a constituent of Daniel’s fathering her. Analogously, we might say— metaphorically or idiomatically—Daniel’s killing Leo caused Leo’s death. Indeed, we do say such things—such metaphorical uses of “cause” parallel to that of “fathered” in saying “The Vietnam War ‘fathered’ many unwanted orphans”. Yet, we must not be misled by that. It’s not a non-metaphorical, straightforward use. Similarly, Daniel’s killing Leo may ‘cause’ Leo’s death, but only metaphorically. Strictly speaking, that killing contains it, or involves, entails, implies, presupposes, or implicates it.
Remember, from footnote 5 above, that “E(5) = E(4)” would have to be true if, as I am often inclined to believe, P(5) = P(4). Of course, even though (P(5) = P(4)) entails (E(5) = E(4)), while (E(5) = E(4)) does not entail (P(5) = P(4)), still it is possible that E(5) = E(4) even when P(5) ? P(4). Some sentences expressing different propositions can sometimes report the same event, if an event can, as Grimm and I believe, be an instance of more than one event-kind.
The oddness, of course, is apparently one of redundancy or repetitiveness, perhaps not so striking as an oddness like that of a purported entity like a green idea in a purported state or process of sleeping furiously. But it’s still a noticeable oddness.
It also seems clear that Grimm (like Bennett, Vollrath, Davidson, and others) won’t be overly impressed with the oddnesses I have pointed out (or other awkwardnesses or absurdities we may be forced to by the Anscombe-Davidson identity thesis or the Grimm maneuver for defending it). For providing palliatives to ease the grammatical or logical pain in such cases seems to be what such theorists are after. Thus, if I also argued that (i) the absurdity of Daniel’s pulling the trigger being also (by parity of reasoning) what Daniel’s killing Leo reduces to, follows from (ii) Grimm’s conclusion, I suspect that Grimm (like Vollrath, cf. the latter’s p. 330) would seek a palliative for that absurdity. Well, maybe such consequences aren’t sufficient to refute the view. But I, for one, believe there clearly is a point where the weight of awkward consequences is large enough to crush any hope for defending the Anscombe-Davidson thesis.
Some other complex events are simple compound events, such as the execution of Sacco and the execution of Vanzetti constituting the conjunctive compound execution of Sacco and Vanzetti. And notice that even though there are negative events in addition to conjunctives ones—e.g., the negative event of Hitler not having been executed for his crimes (which is a state of Hitler, at least)—there don’t appear to be disjunctive events. That is, some analogue of DeMorgan’s Law does not apply to events—i.e., Non-(Ei + Ej) ? (Non-Ei or Non-Ej). The purported ‘disjunctive’ event of, say Hitler’s being executed or Hitler’s committing suicide would not play a role in causation like conjunctive and negative events would (viz., a genuine ineliminable role). For every disjunctive event I can think of, when playing a role of a cause or an effect, simply reduces to a disjunction of causations (e.g., Hitler’s being executed or Hitler’s committing suicide causing the end of German Naziism reducing to either his execution causing it or his suicide causing it) which is not always the case with conjunctive and negative ‘complex’ events. (If Ei+Ej caused Ek, then it’s not necessarily true (hardly ever true, I think) that Ei caused Ek and Ej caused Ek.) More interesting ‘complex’ events are the non-compound ones. For example, in addition to those expressed by causative verbs (“John grew the corn”, which if true suggests that there was the event of John causing the corn’s growing, the corn’s growing being a constituent of the event of John causing it to, which someone might want to identify with certain of John’s farming activities—circularly described, the corn-growing ones—causing the corn to grow), adverbial modifications sometimes involve complex events. For example, the Titanic’s sinking rapidly might be cited as the main contributory cause of the resulting large loss of life. In that case, it is the event (albeit a ‘state’-like event) of the Titanic’s sinking (Es, say) being rapid (that event having that property or being in that state) that caused the large loss of life. For if it (the sinking) had been slow enough, say five days, then all could have been saved. The relevant ‘state’-like event is Es-being-rapid, an event distinct from Es (though presupposing it)—not, as Beardsley (1975) would have it (via his modification sets), identical to it. (For additional discussion of these matters, cf. Chapter 8 above.)
“The question can be raised as to whether A1’s causing B1 is itself an event, over and above the events A1 and B1. If it is, one may ask when that event occurred.” Grimm (1977), p. 223.
“If it’s supposed that A1s causing B1 is itself an event, then how is it related to A1 and B1?Does it itself have causes and effects? If so, what are they?” Ibid., p. 223.
“Affirmative answers lead to regress.” Ibid., p. 223.
Certainly, Thomson (1971) has made that very clear. My own position, incidentally, is expressed in her remarks “There are any number of true answers to the question when A killed B. [e.g., in March 1971, or last weekend] ...there is no true answer to the question when A killed B that gives a time-stretch smaller than the minimal one that includes both the time of A’s shooting of B and the time of the death of B.” (Ibid., p. 122; my insert.) The time of a killing can be generically specified (and cf. Vollrath’s description of it as ambiguous, op. cit., p. 333). The proper response to a demand for more specificity than the minimal time stretch including both shooting and dying (e.g.) is not to try to give it, but to “turn the question aside” (Thomson, op. cit., pp. 122–23)—though I think this is an understatement. The question for additional specificity is answered, I believe, by giving time of the cause-event and time of the effect-event (after giving a time region as the generic answer). To be specific about killing vis à vis its time simply is to give times of its constituents, not to give something it doesn’t have (which is an instant, say, of time within the time stretch which is when the killing ‘really’ took place). Vollrath (op. cit.) seems to see this. So, it’s difficult to see why he ends up defending the Grimm position.
Cf. M. Brand, op. cit., p. 333. Thomson mentions this too (op. cit., p. 132). Also, problems with being precise aren’t new. Do various specifications of how to measure the coastline—e.g., coast-guarding-wise as versus real-estate-wise—show that there aren’t any coastlines? There are, in addition, some other troubles with temporal aspect of Grimm’s position. He claims that events are, or are typically thought to be, momentary. He mentions this in discussing change of events (p. 223, right column; also see the first line of right column, p. 221). He claims that non-momentary events—e.g., processes such as the melting of a piece of butter— might change during their occurrence, which is not the kind of change he has in mind for events becoming causes. However, events changing during their occurrences—non-“grandfather change”—is a delusion. So, Grimm’s allowing for it in order to set it aside and consider the “hard case” is a mistake. If a certain piece of butter melts, for example, from ti (when solid) to tj (when liquid), then variation in certain qualities of the melting (such as melting rate) are simply not changes of the event while it occurs. The event is what occurs, though we may not fully learn what that is until after it’s entirely over for a process. If, during the occurrence, the rate of melting changes, that is not a change of the event. For the event simply contains those changes and does not change into and out of them like an object changing (into and out of a position, for example). If during tk-tl (a proper sub-duration of the butter melting duration ti-tj) the rate of melting is va, and during tm-tn (another proper sub-duration of ti-tj which is distinct from, non-overlapping with, and wholly after tk-tl) the rate of melting is vb (where vb ? va), then the rate of melting has indeed changed. But what is it exactly that has changed? It’s the butter’s melting rate. For one duration, it (the butter) has rate-of-melting va. For a later one, it has rate vb. There is, indeed, a change— from one rate to the other. But that is a sub-event—the change from va to vb—within the whole event (a process) of the butter melting. The butter-melting event itself, however, does not change. It consists (in part) of that change, among many others.
Here are some briefly stated objections and replies to this example: (a) Objection: What happened is that Ei caused Ek, not that Ei caused Ej. Reply: Maybe. However, if that is so, then another kind of example is obtained—something causing a different agentive causing (and, therefore, still satisfying (I) below). For it would appear that Ek itself has an underlying agentive causal form; roughly, “Ek“ = “John’s causing (the wine glass in his hand to fall)”. (b) Objection: It wasn’t John who stained the rug by dropping his wine glass on it (Ej), but Mary. Mary stained the rug by causing John to drop his glass (which she did by scaring him). Reply: Maybe this is correct, if Mary’s agency somehow carries through to the final event (a topic I return to below in conclusion). But either it does or it doesn’t. If it doesn’t, then John remains the agent (and the example stands). (Remember that intention is beside the point. One may do something (be responsible for it) unintentionally. Certainly, this example shows that best. Neither Mary nor John needed to intend to stain the rug, but surely at least one of them did.) If Mary’s agency does carry through, then this does affect the example. But then an example of another sort is generated—e.g., of a causing having an effect. That is, Mary’s causing John to drop his glass (by startling him) caused the rug to be stained; i.e., Mary’s causing Ek by Ei caused El. (c) Objection: All such alleged cases of causes of causings will disappear (via a Grimm maneuver) when it is seen that causing reduces to the cause (like the killing reduces to the shooting); i.e., Ei‘s causing Ej will simply reduce to Ei because Ei becomes the cause of Ej. Reply: This is implausible on any way of looking at it and helps show with a clearer example why causings can’t be reduced to causes, killings to shootings. If Ei‘s causing Ej reduces to Ej itself (Mary’s saying what she did being or becoming identical with her utterance’s causing John to drop his glass and thereby stain the rug), then an infinite regress arises (of a sort implicit in Criticism I above). Let E* = Ei‘s causing Ej. But if E* = Ei (by a Grimm maneuver), then E* =(Ei‘s causing Ej) causing Ej = ((Ei‘s causing Ej) causing Ej) causing Ej = ... ad infinitum.
Some effects that are seemingly effects of an event which is itself an event causing an event—i. e., apparent cases of type (IV)—turn out just to be effects of effects (effect chains). For example, heat’s being applied to it (Ej) might cause some water to boil (Ek). Then it might be thought that that—Ej’s causing Ek— causes a whistle on the container of the water to sound (El). But that would be a mistake. Such happenings would not conform to (IV), but simply be the chain of Ej‘s causing Ek and Ek‘s causing El. Indeed, any candidate for satisfying any of (I)-(IV) may on analysis reduce to a chain of causes or effects. All I claim is that some examples, when properly understood, don’t vanish into chains (though I grant that my favorable examples all entail or involve closely related chains of causes and effects).
There is, indeed, another way to describe the situation—viz., the bullet’s leaving the rifle, followed by its hitting the boulder, followed by its hitting the hiker, where each such event is connected causally to what follows or precedes it (and so we simply have a causal chain of simple events). I am not claiming that the latter alternative is inaccurate, nor that a legitimate chain of causes and effects aren’t revealed by describing it the way I did at first. I only claim that my way of describing it (to fit (III)) is not inaccurate. And, it’s not clear at this point whether the two complex events so described (the instance of (III) and the chain) are or are not absolutely and exactly identical. Also, it might be noted that the way I have derived an example of a non-agentive causing having a cause amounts to taking what one might consider an accompanying circumstance (i.e., El) and calling it a cause of a complex event (i.e., of Ej’s causing Ek). I don’t deny this. Ei could be so considered; i.e., Ej caused Ek in the circumstance Ei (rather like some heat applied to some water causing it to boil at a certain temperature in the circumstances of constant pressure on the water). However, reformulating such circumstances as causes of complex events is not illegitimate, I claim.
Grimm’s are inappropriate enough to suggest a deeper ungrammaticality in the expressions. Yet, I fear that the ungrammaticality is not quite fully specifiable with existing techniques of natural language analysis; i.e., there are controversies —cf. M. Shibatani (ed.) The Grammar of Causative Constructions (Vol. 6 of Syntax and Semantics), Academic Press, 1976.
This does not require that P(5) = P(4); see footnotes 5 and 10 above. If “killing” is simply more specific in meaning than “causing...to die”—following Shibatani, “The Grammar of Causative Constructions: A Conspectus”, especially pp. 28–31, in Shibatani (ed.) op. cit., pp. 1–42—then one and the same event could instantiate a killing and a causing to die in that the former is an instantiation of a more specific kind and, thereby, also an instantiation of any more generic kind the specific kind falls under.
The grammatical facts are as follows. In both (3) and (6), on the appropriate uses herein presumed, “Daniel” is one way or another a grammatical subject. Moreover, the same term is the grammatical agent required by a following verb—i. e., “cause” in (6) has the grammatical feature of requiring a grammatical subject that is agentive (and so “cause” is the agentive “cause”). And in (3), “Daniel” is the grammatical subject of “shooting”, which is likewise agentive. However, there is a difference. In (3), “Daniel” is not the subject of the main verb of the sentence, but only of an embedded sentence that has been nominalized. The surface subject of (3) is “Daniel’s shooting Leo”. That complex subject is not grammatically marked agentive and the “cause” of (3) is, thereby (since it is grammatically well-formed), non-agentive. These grammatical facts reveal that (i) the causation expressed or reported via an appropriate use of (3) is non-agentive causation, and (ii) the causation expressed or reported via an appropriate use of (6) is agentive causation. One event causing another is non-agentive causation. An agent causing an event is agentive causation. (Vollrath, op. cit., pp. 331–332, seems to be as clear as anyone about this, but fails to see its moral.)
However, to avoid infinite regress, it may be that some basic actions are agent causations that don’t entail (or require) such closely related non-agent causings.
Daniel may well blame himself for helping Leo commit suicide albeit unwittingly. He may think he should have been paying more attention to surrounding activities during target practice, or regret not being quicker to respond to something in his line of fire, etc. Some may object to my example by saying that the circumstances show that Daniel did not shoot Leo in the appropriate agentive sense of “shoot”. That would be to the point. Suffice it to say here that the example can be further modified so that (say) Daniel was shooting Leo intentionally (in order to inflict a very superficial wound to his forearm so that Leo could have an excuse for avoiding the draft). Then permit Leo to exercise his suicidal tendencies and make (3) true. Then the shooting was the right sort (genuinely agential, even intentional), though Leo killed himself.
The research for this chapter was supported in part by a grant from the Syracuse University Research and Equipment Fund (administered by the Senate Research Committee). I am extremely grateful to Cindy D. Stern for important criticisms of an earlier draft, as well as for helpful suggestions and two of the crucial examples used above.
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 1997 Springer Science+Business Media Dordrecht
About this chapter
Cite this chapter
Peterson, P.L. (1997). The Grimm Events of Causation. In: Fact Proposition Event. Studies in Linguistics and Philosophy, vol 66. Springer, Dordrecht. https://doi.org/10.1007/978-94-015-8959-8_10
Download citation
DOI: https://doi.org/10.1007/978-94-015-8959-8_10
Publisher Name: Springer, Dordrecht
Print ISBN: 978-90-481-4856-1
Online ISBN: 978-94-015-8959-8
eBook Packages: Springer Book Archive