16,99 €
Mehr erfahren.
- Herausgeber: John Wiley & Sons
- Kategorie: Geisteswissenschaft
- Serie: Key Concepts in Philosophy
- Sprache: Englisch
In most academic and non-academic circles throughout history, the world and its operation have been viewed in terms of cause and effect. The principles of causation have been applied, fruitfully, across the sciences, law, medicine, and in everyday life, despite the lack of any agreed-upon framework for understanding what causation ultimately amounts to.
In this engaging and accessible introduction to the topic, Douglas Kutach explains and analyses the most prominent theories and examples in the philosophy of causation. The book is organized so as to respect the various cross-cutting and interdisciplinary concerns about causation, such as the reducibility of causation, its application to scientific modeling, its connection to influence and laws of nature, and its role in causal explanation. Kutach begins by presenting the four recurring distinctions in the literature on causation, proceeding through an exploration of various accounts of causation including determination, difference making and probability-raising. He concludes by carefully considering their application to the mind-body problem.
Causation provides a straightforward and compact survey of contemporary approaches to causation and serves as a friendly and clear guide for anyone interested in exploring the complex jungle of ideas that surround this fundamental philosophical topic.
Sie lesen das E-Book in den Legimi-Apps auf:
Seitenzahl: 285
Veröffentlichungsjahr: 2014
Ähnliche
Table of Contents
Key Concepts in Philosophy
Title page
Copyright page
1: Introduction: All Things Causal
Singular vs. General Causation
Linear vs. Non-Linear Interaction
Productive vs. Difference-Making Causation
Influence-Based vs. Pattern-Based Causation
Questions
2: Causal Oomph
The Regularity View of Causation
Hume's Causation
Summary Evaluation
Advantages
Challenges
Questions
Further Reading
Study Question
3: Process and Mechanism
Causal Process Theories
Advantages
Challenges
Further Reading
Study Questions
Mechanisms
Mechanisms and Levels
Bottoming Out
Advantages and Challenges
Questions
Further Reading
4: Difference-Making
Counterfactual Dependence
Advantages
Challenges
Practice Essays
Summary
Questions
Further Reading
5: Determination
History of Causation as Determination
Advantages
Challenges
Questions
Further Reading
6: Probability-Raising
Probability-Raising Accounts of Causation
Probabilities
Probabilistic Relations
Standard Accounts
Asymmetry
Spurious Correlations
Simpson's Paradox
Advantages
Challenges
Reichenbach's Common Cause Principle
The Unclear Relationship between Probability and Singular Events
Conjunctive Forks
Questions
Further Reading
7: Manipulation and Intervention
Manipulationist Theories
Von Wright's Formulation
Menzies and Price's Formulation
Challenges
Effective Strategies
Interventionist Accounts
Causal Modeling
Questions
Path-Specific Causation
Singular Causation
Advantages
Challenges
Further Reading
8: Mental Causation
Dualism vs. Materialism
Reductive vs. Non-Reductive Materialism
Summary
Non-Reductive vs. Reductive Materialism
Summary
Questions
Bibliography
Index
Copyright © Douglas Kutach 2014
The right of Douglas Kutach to be identified as Author of this Work has been asserted in accordance with the UK Copyright, Designs and Patents Act 1988.
First published in 2014 by Polity Press
Polity Press
65 Bridge Street
Cambridge CB2 1UR, UK
Polity Press
350 Main Street
Malden, MA 02148, USA
All rights reserved. Except for the quotation of short passages for the purpose of criticism and review, no part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of the publisher.
ISBN-13: 978-0-7456-5995-4
ISBN-13: 978-0-7456-5996-1 (pb)
ISBN-13: 978-0-7456-8584-7 (epub)
ISBN-13: 978-0-7456-8583-0 (mobi)
A catalogue record for this book is available from the British Library.
The publisher has used its best endeavours to ensure that the URLs for external websites referred to in this book are correct and active at the time of going to press. However, the publisher has no responsibility for the websites and can make no guarantee that a site will remain live or that the content is or will remain appropriate.
Every effort has been made to trace all copyright holders, but if any have been inadvertently overlooked the publisher will be pleased to include any necessary credits in any subsequent reprint or edition.
For further information on Polity, visit our website: www.politybooks.com
1
Introduction: All Things Causal
At the beginning of every course I teach on causation, I like to mention three reasons for studying causation. First, it is the most important relation in the universe. As the great philosopher once said, “It surrounds us and penetrates us. It binds the galaxy together.” Second, virtually every other topic you can study depends on causation: science, history, music, business, law, medicine, and most important of all, bicycle repair.
My third reason is difficult for me to communicate without injecting some philosophical jargon, so I will leave you hanging for a while. In the meantime, you should think about reasons for yourself. Before, during, and after reading every chapter, ask yourself, “Why is causation worth studying? What is there to learn from philosophers musing about cause and effect? Is there any conclusion that could transform my conception of reality?” I can intimate to you that by spending a bit of time mulling over such questions, I have struck upon an idea that has altered my opinion about virtually everything I care about and everything I do. I hope that by introducing you to some tools for investigating causation, you will be encouraged to revolutionize your own conception of reality. If you succeed, please track me down and let me know. I would love to hear about it. We humans don't keep in touch enough.
It's going to take a modicum of effort for us to sort out what causation is and how to define it. We all have some basic ability to apply terms like ‘cause’ and ‘effect’, so we can start out by assuming that ‘causation’ is just another name for the relation between cause and effect, whatever that turns out to be. We will see soon enough that it is problematic to decipher expressions like ‘it causes’ and ‘it is a cause of’, but on the bright side, that is a sign that it's an ideal topic for drawing the sorts of distinctions that are the hallmark of good philosophy.
Before digging into details, you should be aware of the scope of a philosophical investigation of causation. Philosophy, it has been said, is in the business of understanding “how things in the broadest possible sense of the term hang together in the broadest possible sense of the term.”1 My goal is to help you to learn how philosophers have tried to understand causation insofar as it applies generally to all things. That means the focus of discussion is not on how we humans acquire our causal concepts, nor on how scientists discover causes, nor on how various historical figures have thought about causation. Instead, the goal is for you to acquaint yourself with a map of the contemporary conceptual landscape. We will be trying to answer questions like, “Which ways of thinking about causation are fruitful?” “What obstacles have been overcome in understanding causation?” and “What are the outstanding deficiencies in our current theories of causation?”
In this first chapter, I will avoid introducing any theories. Instead, I will highlight four distinctions that I have personally found to be insightful and helpful for organizing all of the information to come.
singular vs. general causationlinear vs. non-linear interactionproductive vs. difference-making causationinfluence-based vs. pattern-based causationI will now proceed through a discussion of each distinction.
Singular vs. General Causation
A singular event is a particular occurrence in the history of the universe. It can be identified in terms of its location in time and space, and it can be described in terms of the properties and relations it instantiates. When discussing singular causation, we typically designate one singular event as the effect. We then consider all the singular events that count as partial causes of that one effect. Philosophers use the term ‘cause’ for what is more accurately described as a partial cause. Singular causation is the relation that holds between any single effect and each one of its singular (partial) causes. We often express claims of singular causation in the past tense: “c was one of the causes of e,” or “c caused e,” because singular causation is usually evaluated in retrospect.
A general event is a type of event. By their nature, event types can be instantiated in multiple locations in space and time, and they can be described in terms of whatever properties and relations constitute that sort of event. We often express claims of general causation in the present tense: “C causes E,” or “Cs cause Es.” General causation is the relation that holds between an event type E and any event type C that tends to cause (or bring about) E.
(I like to use lowercase letters for singular causes and uppercase letters for general causes.)
To acquire an intuitive feel for the difference between singular causation and general causation, consider the encouraging tale of Hobo, the scruffy hound. Hobo was generously helping to reduce food waste around dumpsters in Moore, Oklahoma, when she was suddenly cornered by the resident dogcatcher. I happen to know this guy personally, and trust me, you do not want him to catch you whether you are a dog or not. Fortunately, Moore is situated in a tornado corridor, and at the very moment when Hobo had abandoned all hope of escape, a tornado touched down and hoisted Hobo high into the air, carrying her over to nearby Kitchen Lake, where she splashed down and swam to shore uninjured. Having learned a valuable lesson, Hobo migrated eastward, last anyone has heard.
In the particular fragment of history described by this very true story, Hobo was rescued by the tornado. The tornado was a major cause of Hobo's survival because dogs snared by the dogcatcher tend not to last long. More formally, we stipulate that the event of interest to us (playing the role of effect) is the dog's survival for several more years. Examining what happened in the fragment of history encompassing Hobo's narrow escape, we can rightly declare that the tornado's lifting of Hobo was one of the causes of her survival. In general, however, tornados do not cause dogs to survive for a long time; they pose a mortal danger. So, the tornado was a singular cause of survival even though tornados are not a general cause of survival.
Let's now engage in some philosophy by examining the distinction.
1. Note that in the definitions I offered for singular and general causation (marked in boldface), I used the term ‘cause’ on both sides. Therefore, I have not defined causation in terms of something non-causal. In philosophical jargon, I have not provided a reductive definition of causation. So far, I have merely attempted to distinguish the conceptual role of singular and general causes enough to convey what makes them different from each other, not to specify what each is in full detail.
2. Ordinary language carries connotations that some partial causes are more noteworthy than others, and much of the phrasing used for expressing claims of causation can obscure what is at stake in disputes over causation.
One can distinguish between background and foreground causes: The last straw you placed on the camel caused its back injury. We think of the presence of the last straw as being especially prominent and thus figuratively in the foreground even though the last straw was no heavier than each of the previous thousands of straws, which we think of as in the background.One can distinguish proximal (or proximate) causes from distal causes. The proximal causes are nearby in space and time and the distal causes are far away in space or time. For example, we say “Julius Caesar's death was caused by assassins,” not that his death was caused by his own birth. Yet, upon reflection, it becomes evident that Caesar's birth set off a causal chain that led inexorably to his demise.One can distinguish enabling causes from activating causes. We say that what caused the blaze was the cow's having kicked over the lamp. We do not say that the presence of the many wooden buildings tightly packed together caused the fire. A more perspicuous description would indicate that the density of wooden structures caused the blaze to engulf most of the city by enabling the fire to spread and that the toppled lamp was an activating or triggering cause.In all three examples, we have a tendency to deny the status of cause to the background, distal, and enabling causes, even though when we examine our reasons for doing so, we find that objectively they have the same status as events we are happy to designate as causes. In such cases, the features that distinguish them from genuine causes appear to be merely pragmatic. The partial causes we identify as non-causes might be less important to us, or they might be less amenable to our manipulations, or they might be so pervasive that it would be misleading to draw attention to them.
When philosophers discuss causation, they usually seek (or at least claim to seek) an egalitarian2 conception of causation that strives to count the background and distal and enabling causes as bona fide causes. They bracket the goal of evaluating which causes are more important and the task of ascertaining whether one of the causes deserves to be designated as the cause.
The subtleties of language and the general culture, however, allow these observations to be easily overlooked. For example, an expression of the form ‘a cause of e’ might be construed as more or less equivalent to ‘one of the causes of e’, but it might also be understood as ‘something that caused e’. If these two renderings of ‘a cause’ sound like they say the same thing, imagine a scenario where someone buys a winning lottery ticket. Is Jennifer's purchase of a lottery ticket a cause of her receiving money from the lottery agency? The purchase in this case was certainly one of the causes because she had to play to win, but the purchase was not something that caused her to win because it gave her only a one in a million chance. If we identify e's singular causes by answering, “Which events caused e?” we might arrive at different events than by answering, “Which events were among the causes of e?” The first question is usually answered by citing the events that were the most important contributing causes to e at the time they occurred. The second question is usually answered by citing events that somehow played a positive role in the overall development toward the occurrence of e. Philosophers are almost always concerned with the egalitarian conception of cause that corresponds to the second question.
3. Our linguistic idioms can also mislead us about causation to the extent that we do not separate out logical or definitional connections among events. For example, if someone answers the question, “How did Julius Caesar die?” by noting that he was assassinated, that conveys relevant information about the events that caused his death. However, it is somewhat misleading to say, “The assassination of Caesar was a cause of Caesar's death,” because assassination implies by definition a corresponding death. The definitional link is not a constituent of the causal connections among the conspirators and Caesar, and we need to ignore it when ascertaining causes.
Such definitional linkages are not rare because the term ‘cause’ is itself loaded. When we describe an event c as ‘one of the causes of e’, that presupposes that e occurred. Nothing can be a cause of an event that does not occur. So, when investigating what aspect of the cause makes its effect occur, we need to ignore the fact that its being one of the causes ensures the occurrence of its effect by definition. We do this simply by not treating “is a cause of e” as if it describes a property of c that can play a role in causation.
4. An alternative way to express the distinction between singular and general causes is to say that particular instances of causation are cases of token causation and that causal generalities concern type causation. The difference between types and tokens is meant to be widely applicable, not just to events. For example, the typographic character that comes first in alphabetic order is a type, which we sometimes call “the letter A” or just “A.” Here are five of its tokens:
A a A a a
When we say that the line of text contains only one letter, we are saying that there is only one letter-type instantiated. When we say that the line of text contains exactly five letters, we are saying that there are five tokens of this one letter-type.
Applying the type–token distinction to causation, we say that the generality that holds from smoking to lung cancer is type causation (or type-level causation) and that any particular instance of lung cancer that was caused by smoking is token causation.
Is there any reason to speak of singular vs. general causation rather than token vs. type causation? They are mostly the same distinction, but perhaps the following example highlights a nuance. We nowadays take for granted that unicorns have never existed,3 yet it seems reasonable to assert that if there had been any unicorns that walked in mud, they would have caused hoof-shaped impressions in the mud. One way to describe this possibility is to say that there is type causation between unicorns stepping in mud and the existence of hoof-marks. This type causation has no tokens because there never have been any real unicorns. On the one hand, it seems reasonable to speak of type-level causal relations among unicorn-events just because unicorns are a subtype of hoofed animal. Hoofed animals in general cause hoof-marks, and more specific hoofed species like unicorns will generally cause hoof-marks when present in muddy environments. On the other hand, when we speak of causal generalities we tend to have in mind cases where there is a repeatedly instantiated pattern of singular causation. Many individual horses have walked in mud and thereby caused hoof-marks, but no instance of a unicorn having caused a hoof-mark exists. So, perhaps we could let the term “general causation” refer to type causation that has corresponding tokens in the actual world, and make clear that “type causation” does not require that the causal regularity ever be realized.
Just to add some redundancy, philosophers have recently begun to adopt a distinction between actual causation and potential causation. As far as I can tell, ‘actual cause’ means the same thing as ‘token cause’ and the same thing as ‘singular cause’.4 Potential causation, I think, is meant to encompass causal relations that involve unicorns in addition to causal relations among existing entities.
5. Finally, most philosophical work on causation concentrates on singular causation, and there appears to be a prevailing opinion that if an adequate account of singular causes could ever be found, it would be relatively easy to construct a complementary account of general causation. The idea, perhaps, is that if we have a good rule telling us whether one chosen event is a cause of our given effect – using information about its surroundings, any operative laws, and perhaps other parameters – then that same rule will automatically identify the full collection of causes. One reason to suspect that the connection between singular and general causation is more subtle is that on the few occasions where rules for deriving general from singular causation have been specified, there has been a curious lack of detail.5
More important, there appears to be a consensus among experts that a satisfactory account of general causation cannot be provided without incorporating an account of singular causation. The worry perhaps is that a generality like “inhaling the fumes from modeling glue causes headaches” cannot hold unless there are objective enough facts about whether particular inhalations have caused particular headaches. Unlike unicorns, the causes here are not fictional.
Let's sum it all up. The main distinction so far has been laid out in terms of several terminological choices: singular vs. general causation, token vs. type causation, and actual vs. potential causation. These are basically the same. Along the way, we also noted that the sort of singular causation philosophers care about is egalitarian.
Linear vs. Non-Linear Interaction
Think about a spring-operated scale used to measure weights. On an empty scale, you place a 1 kg mass and the scale moves down 1 cm. Then, you place a 3 kg mass on the scale and it moves down 3 cm more. What is causally responsible for the current position of the scale? Uncontroversially, the placement of a total of 4 kg on the scale caused it to move 4 cm. But are both weights equally causes?
In one sense, they are. Without the smaller mass, the scale would not stretch to 4 cm. Also, without the larger mass, the scale would not stretch to 4 cm. Both masses are needed for the observed effect. We call this a case of joint causation. The two masses jointly cause the 4 cm displacement of the scale.
In another sense, they are unequal. The larger mass is responsible for three-fourths of the displacement. It is pulling down three times as much as the smaller mass.
Insofar as we are seeking egalitarian causes, the two placements of the masses are equally causes. And insofar as we are making predictions or explanations, we can use unequal quantities to represent the relative strength of the causes. The effect is proportional to the relative strength of its causes, and in the language of mathematics, we say that the effect here is a linear function of the two causes.
Is there some problem here? As far as I can tell, there isn't, but there are non-linear causal relationships as well, which we need to consider. A good example of non-linear causation is when an effect is generated by passing some sort of threshold.
Suppose you place 70 separate 1 kg masses on the scale at once, and the scale breaks, because it holds at most 50 kg. What were the causes of the scale's breaking? It seems reasonable at first to say that the entire collection of masses (being on the scale at once) caused the breakage. But what about other causes? In particular, can we ascertain whether each individual mass was a (partial) cause of the breakage?
On the one hand, we can say that no individual mass made a significant difference. If any one of the 70 had been omitted, the scale would still have broken. On the other hand, we know that all 70 together broke the scale and that they all weigh exactly the same. That suggests we should apply the principle that if some of them are responsible for the breakage, so are all the others. In that case, each mass is one of the causes of the scale breaking.
Unlike the previous linear case, there is a potential problem here for the egalitarian conception of causation. We have two lines of reasoning appealing to the equality of causes and drawing opposite conclusions. One says that because none of the individuals caused the breakage, none of them were causes. The other says that some of the single masses had to be causes of the breakage, so they were all causes. The problem is that attributing the status of ‘cause’ (in the egalitarian sense) is a binary matter. Every event is either “one of the causes” or it isn't. The framework philosophers operate under does not allow us to identify the placement of each single mass as one-seventieth of a cause or one-fiftieth of a cause or any other fraction.
One common response to such examples is to point out that some cases of causation involve overdetermination. Placing two 100 kg masses on the scale causes breakage. Because each of them alone was enough to ensure breakage and because they are identical in all relevant respects, both masses are individually culpable for the breakage. We say that these two causes overdetermined that the scale broke. No problem here.
One can now attempt to impose the same reasoning on the scenario where the breakage was caused by the 70 small masses. We can motivate attributing causal blame to each individual mass by noting that in exemplary cases of overdetermination, an event can be a cause even though it was entirely redundant. So it is no requirement for causation that if the cause had not been there, the effect would not have occurred. Furthermore, because the full complement of 70 masses caused the breakage, each of the masses being on the scale was individually and equally a partial cause of the breakage. Finally, to say that some event c was a partial cause of e is (in philosophers' parlance) the same as saying that c is one of the causes of e.
I personally find that this way of imposing the framework of egalitarian causes is acceptable as far as it goes, but it can go too far by identifying so many events as causes that most of the causal distinctions we want to draw will need to be cashed out in terms of some further theory that assesses the relative importance of the many egalitarian causes.
If we have settled on the goal of seeking a theory of causation under which every event is designated either as a partial cause of e or as a non-cause of e without any allowing for degrees of causation, then presumably we are going to simplify (in our modeling activity) some aspects of nature's causal structure. The philosopher's hope is that this simplification is not an oversimplification.
To what extent the egalitarian simplification misleads rather than leads, however, is an issue you should allow to nag you continually. You ought to wake up in the middle of the night startled and ask yourself, “Why is everyone arguing about this egalitarian conception of singular causation? What is it good for?”
Productive vs. Difference-Making Causation
It is important not to let the discussion spin out of control by becoming too abstract. Let us turn at once to a social scenario incorporating the non-linear causation we just discussed. The purpose of the following example is to illustrate two points. (1) In many cases we care about, causal responsibility should not in general be distributed among multiple causes in terms of a fraction of responsibility. (2) We sometimes have significantly different conceptions of causation driving our judgments of causal responsibility. In particular, I want to point out that we sometimes think of a cause as somehow generating or producing its effect, and at other times we think of a cause as something that makes a difference as to whether or how its effect comes about. It is not currently well known whether (and how) these two conceptions are compatible.
Once upon a time, there were two nearby islands. Paul was the lone inhabitant of one island, and he had a small boat for fishing. Vivian was the lone inhabitant of the other island, which hosted abundant stringy plants that she harvested for rope fiber. By working alone, each collected enough food for his or her own survival, but just barely and with several periods of hunger. By cooperating, however, they could gather enough food to be healthy year-round and to preserve some leftovers for insurance.
Vivian had already spent an enormous amount of time making a few large nets out of plant fiber. She haggled with Paul, who agreed to let her use his boat if she gave him half of the fish. She acceded and gathered heaps of fish from the deeper waters. Before handing any fish over, Vivian reconsidered the fairness of dividing the fish equally between her and Paul. Vivian suggested to Paul that it was not fair for him to take half of the fish because he did not do any work; he just let her use the boat. But Paul pointed out that without his boat, she wouldn't have gotten the huge stock of fish. Using his boat made all the difference between going hungry and having a huge stockpile. Vivian then noted that the same observation applied to her. Without the net, she could not have caught so many fish.
There are two conflicting arguments here concerning how to divide the cooperative surplus fairly. Vivian, we might say, was suggesting that the rights to the fish are earned by doing work and that the surplus should be distributed proportional to the amount of labor expended. Paul was countering that they are both free agents and if they both agree to a deal, what is fair is to stick to the terms: “If I hold out until you agree to give me half the fish,” Paul continues, “and you accept, it's fair for me to get half. So long as everyone consents, there's nothing more to discuss.”
I first learned about this problem of cooperative surplus while studying political philosophy. Vivian's line of argument fits in the tradition of John Locke, who argued that a person has a right to private property if he mixes in his personal labor.6 Paul's line of argument follows the classic libertarian tradition by saying, “What everyone freely agrees to is fair.”
I later realized that debates about fairness (and other ethical issues) share important features with debates about causation. The main similarity I want to point out is the non-linear causal relationship between the activities of Vivian and Paul and the total resulting amount of fish.
Let's first consider the case of a linear relationship between effort and result. When they were working alone on their own islands and their fishing conditions were equal, it might have been reasonable (at least as a crude approximation) to declare that a fair allotment of fish should have been proportional to the fraction each catches or perhaps to the fraction of work each does. If Paul chooses to work twice as long as Vivian and he catches twice as many fish, he should get his two-thirds of the total. Similarly with regard to causation, when a scale is stretched by holding a 1 kg and a 2 kg block, the 2 kg mass is responsible for two-thirds of the stretching. Furthermore, we can state these relations in terms of difference-making. We can reasonably conclude (because they are working by themselves) that if Paul had chosen not to fish at all, then Vivian would have still caught the same amount. So, Paul's actual amount of fishing activity is responsible for their having a combined catch that would have been two-thirds smaller without his effort.
For contrast, let us now consider the non-linear relationship between effort and result that exists in the original scenario. If Paul had chosen not to let his boat be used, he and Vivian would each have caught only a small amount of fish, say 1 kg each. By their working together, a total of 100 kg was caught. So, the extra amount of fish that was gathered because of Paul's letting Vivian use his boat is 99 kg. Similarly, the extra amount of fish that was gathered because Vivian constructed her nets is 99 kg. In this sense, each person individually made a 99 percent difference in the amount of fish caught. This is the problem of cooperative surplus: that it is impossible for both of them to get 99 percent of the fish.
When stuck in such a situation, one answer is to accept that both people made the same 99 percent difference in the amount of fish and split the bounty equally. That is Paul's bargaining position. Another answer is to incorporate other factors such as the amount of time and energy spent in active retrieval of the fish. Vivian argued that because Paul did not engage actively in collecting the fish, he does not deserve an equal share even if he was successful at bargaining for it.
What we have in Vivian and Paul's arguments is a divide between looking at how much difference each made to the effect and looking at how much energy or effort was expended. Vivian sees herself as more responsible for the catch because she exerted more effort and engaged in more production. Paul sees himself as equally responsible for the catch because he made just as much of a difference.
This distinction applies generally to theories of causation. Some treatments emphasize that causation is productive. Causes make their effects happen. Causes bring about their effects. Causes give rise to their effects. Causes alter and change the world. Other accounts emphasize that causation affects. Causes are difference-makers. Without the cause, the effect would not have happened. Causation is recognized by intervening in the world and manipulating causes.
This divide between a productive conception of causation and a difference-making conception will play a significant organizational role in the rest of our discussion of causation, but for now I just want to suggest that one reason there are many difficult decisions to make about how to distribute the goods of society fairly is that the goods come to us by way of non-linear causality. The simple idea that we should proportion reward for good outcomes and punishments for bad outcomes by the amount each cause contributed to the effect runs smack into the problem that in most cases, there is no objective way to proportion causal responsibility. Typically, an effect has a complicated tangle of causal relationships, and we search in vain for a formula for the degree to which one cause is more important than the other. It is a useful exercise to mull over the range of issues in our society that exhibit non-linearity owing to the underlying causal relations being non-linear.
Influence-Based vs. Pattern-Based Causation
I think the most underappreciated distinction in the history of philosophical musings on the subject of causation appears in Michael Dummett's (1964) article entitled “Bringing about the Past.” In this paper, Dummett attacks a stock argument against the possibility of influencing the past, namely the argument that “you cannot change the past; if a thing has happened, it has happened, and you cannot make it not to have happened.” He points out that this argument follows the same pattern as the fallacious argument that you cannot affect the future because there is only one way things are eventually going to turn out. Dummett agrees with the widely held view that we should not deny our ability to influence the future just because a single determinate future will eventually occur. But, then, by parity of reasoning, he draws the unorthodox conclusion that we should not dismiss the possibility of influencing the past. He thinks the stock argument – that we cannot affect the past because it has already happened – is just as bad as the argument that we cannot affect the future because it will eventually happen.
The possibility of influencing the past is a fascinating topic, and there are nuances galore to be uncovered. But, because space is short, I will just point out one extraordinarily powerful but only briefly discussed idea in Dummett's article. Here is a quotation from very near the beginning:
I think that this asymmetry [of causation being directed always toward the future and never toward the past] would reveal itself to us even if we were not agents but mere observers. It is indeed true, I believe, that our concept of cause is bound up with our concept of intentional action: if an event is properly said to cause the occurrence of a subsequent or simultaneous event, I think it necessarily follows that, if we can find any way of bringing about the former event (in particular, if it is itself a voluntary human action), then it must make sense to speak of bringing it about in order
