Data Science and Ethics: Deontology vs. Consequentialism

Introduction

Consider the following real scenario:

• President Donald Trump has implemented a travel and refugee ban on anyone from one of several predominantly Muslim countries.  Suppose that, as a result, there are 0 terrorist incidents in the United States in the next four years, and assume that, were it not for this ban, there would have been several terrorist incidents.  
• Is the travel and refugee ban moral?
• Furthermore, suppose that the government has gathered various kinds of information about you based on your web surfing history, Facebook profile, shopping history, mobile device locations, etc. in order to fight terrorism. 
• Is the collection of such data immoral? 
• Furthermore, suppose that a machine learning model used to fight terrorism uses the religion (or race, gender, marital status, sexual orientation, etc.) of an individual to highly predict if a person will commit an act of terrorism.  Without this field, the model cannot accurately identify terrorists as well as it previously could.
• Is the use of this highly sensitive, personal, and potentially discriminatory field permissible?

Underlying this real scenario is a sort of  dilemma, pitting two desirables against each other.  On the one hand, we want good results in the aims we pursue, so the more effective a means is to achieving those results, the better.  In this case, we want protection from terrorism and safety within the United States, and so we would embrace actions that further these aims.  However, we also believe there are limits that no means can cross, that the "end does not justify the means" in every circumstance.  If you are like me, then you are probably at least a little uncomfortable with using country of origin or religion in profiling immigrants, even in the service of fighting terrorism.

This is an age old debate between the opposing normative ethical theories of consequentialism and deontology, but it is now being played out in the realm of data science. Below, we will explore the relationship between ethics and data science and discuss how ethical theory can help provide guidance about the above scenario and similar situations.

Philosophical Terms/Distinctions

Here are some key terms and distinctions to understand:

• Deontology
• The normative ethical view that an action's moral rightness or wrongness is not determined solely by the outcome of the action.  Other factors are important as well (e.g., intention of the agent, the nature of the act).
• Consequentialism
• The normative ethical view that an action's moral rightness or wrongness is determined solely by the outcome or consequences of the action.  Typically, an action is morally right if and only if it maximizes happiness, utility, or some other desired output.
• Morality vs. Legality
• Actions may be legal but immoral.  Similarly, actions may be illegal but moral.  Thus, legality is not the same as morality.
• For example, one might argue that the use of pornography is immoral, even though it is legal.  Similarly, in many countries, certain kinds of religious or political expression are illegal, but most of us wouldn't say that these kinds of expression are immoral.
• User/Individual interests vs. Company/Government/Society/Group Interests
• An individual/user's interests are distinct from the interests of a company, government, group, or even society.  While these interests might be aligned, they do not need to be aligned and may come into conflict.
• Hierarchy of rights: life, liberty, property
• There is hierarchy to rights.  In particular, the right to life is more important than a right to liberty, which itself is more important than a right to property.  Life is necessary for liberty (without life, you cannot have liberty), and liberty is necessary for property (without liberty, you cannot exercise your right to have or make decisions with respect to your property). 
• In case of direct conflict, the higher right must take precedence over the lower right.

 

Data Science Core Concept: Machine Learning

At the core of data science is the use of algorithms to predict and classify entities (e.g., people) into different categories.  This is machine learning.  Most often, the algorithm uses various fields (e.g., age) to predict an unknown but useful value (e.g., spending behavior).  The more informative the fields fed into the algorithm, the more accurate and useful the predictions will be.  Thus, it is very important to figure out which fields are the most useful in predicting or classifying, and which are not helpful.

In some kinds of "deep learning" associated with AI, the algorithms do not require the use of fields.  The algorithm constructs these on its own before making a prediction or classification.  While such algorithms often work very well, understanding why the model produced the results it did can often be nearly impossible to interpret.  That is, we don't know why the "black box" works, but it does.

Ethical Data Science: An Exploration 

So how can we apply the above philosophical terms and distinctions to the realm of data science/machine learning, and in particular, thoughtfully address the above scenario? Let's explore.

A Consequentialist Start

The reasoning behind the travel ban has the form of something like:

• If one is Muslim (or from a predominately Muslim country), then one is (or very likely to be) a terrorist.  

In the context of data science, the input field would be "Country of Origin" and/or "Religion", and the target field is "Terrorist/Not Terrorist".  Now suppose you have a sample data set that is representative of the world population with these fields, what would you expect to find?

If this is all that you have, not much.  We know that most people aren't terrorists.  Most Muslims aren't terrorists.  Most Muslims from predominately Muslim countries are not terrorists.  A model just based on these fields would probably just label everyone from these predominately Muslim countries as not being a terrorist to get a high accuracy, even though it would miss identifying any actual terrorists.  In terms of a confusion matrix, our initial model would have 0 false positives, 0 true positives, several false negatives (labeled "not a terrorist" but actually a terrorist), and many true negatives (labeled "not a terrorist" and in fact not a terrorist). 

In contrast, to avoid missing any actual terrorists, the model behind President Trump's travel ban goes to the opposite extreme and labels everyone from these predominately Muslim countries as a terrorist.  That is, the model he proposes would have many false positives (labeled "a terrorist" but not actually a terrorist), several true positives  (labeled "a terrorist" and actually a terrorist), 0 false negatives , and 0 true negatives.
 

So how would one improve this model?  With more data!  Religion and country of origin do not carry enough information on their own to provide a predictive model that correctly identifies terrorists as terrorists and correctly identifies non-terrorists as non-terrorists.  While I am not an expert at what can correctly identify a terrorist as such, I would think that, in addition to Religion and Country of Origin, fields like Gender, Race, Ethnicity, Age, Education, Economic status, Internet Activity (websites visited), Purchase History, and Travel History may be useful in more accurately identifying terrorists and non-terrorists.  Suppose that we get this information and now we have a very accurate model that correctly identifies both terrorists and non-terrorists.  Great!

In summary, in the initial set up, President Trump's model supposedly prevented all terrorist attacks.  In this it was extremely effective in the desired outcome: no terrorist attacks.  But it came at the cost of preventing many non-terrorists from immigrating to the US.  That is, it prioritized the right to life of innocent civilians over the right to liberty of prospective immigrants, assuming a rights conflict.  But is there such a conflict?  Our updated model suggests that, in theory, there isn't.  If one has enough information, one can protect the right to life of civilians AND the right to liberty of immigrants by correctly identifying both terrorists and non-terrorists. 

A Philosophical Reflection

But notice what we had to do.  We had to make use of lots of personal information collected by the government on individuals.  Is there a limit to the amount and kinds of information that the government or anyone else can collect?  Are there restrictions on how the collected information can be used?
 

I would bet that most people intuitively want to place limits on both the amount and kinds of information being collected on them, and the reasons that they would give for that intuition are often moral in nature.  For example, most people are fine with a certain amount of information being collected on them.  But as that information increases in size, they would become less and less comfortable with this.  They would likely express fears that the information will be used in inappropriate ways to target them, manipulate them, or otherwise influence them in ways that they would not want.  That is, the information could be used to harm them.

Not only is the amount of information important, the kind of information matters greatly.  Some kinds of information we aren't concerned about being publicly known (e.g., gender).  But there are other kinds of information that we would be hesitant to share (e.g., medical history, financial history) not merely because of how it could be used, but because of its personal and private nature.  We can certainly think of kinds of information that we wouldn't want anyone to know, especially private companies or governmental organizations (e.g., sexual history and preferences).  In short, the collection of some kinds of information about us would be a violation of our privacy, personal integrity, or human dignity and nature in the extreme.  These are moral concerns.

Furthermore, even if the amount and kind of data were permissible, the use of such data may not be.  In the model above, we had to make use of extremely sensitive information like race and religion to come up with this model.  While I may not object to my race and religion being known and used in some matters (e.g., demographic statistics), I would certainly have concerns with it being used in such a way as to harm myself and others, or deal with myself or others in ways I consider to be unjust.  Thus, there are morally proper uses of data.

Let's step back for a moment.  Why use data in the first place?  In short, to better achieve our aims and goals through an increase in knowledge on the subject at hand and thereby improve our decision making in achieving our goals.  On the governmental side, the use of data is intended to help in governing, whether it is to improve services, increase safety, or support other legitimate government interests.  For companies, data helps a company succeed in meeting competition, increasing profits, and other business aims. 

In contrast, an individual typically does not use data in this way but is the source of the data that governments and companies use.  An individual consents to disclosing such information in exchange for goods and services.  That is, there is a sort of marketplace transaction that occurs in which an individual "pays" for goods and services with his or her information.  Thus, we can think about that information as being part of an individual's personal property.

We do this in numerous ways every day, usually without thinking about it.  We sign up for Facebook, giving our names and other demographic information.  Through the use of its services, we also provide information on who our friends are, what we are interested in, where we live, and what we believe, knowing that online companies make money off of such information through the use of targeted advertising.  We submit to background checks to receive employment, to travel in the faster TSA-Pre lines, or obtain passports to travel abroad.  We sign up for customer discount cards at grocery stores to save money, while these stores use our purchase information to improve pricing, marketing, and ultimately, sales.  In short, there are legitimate and mutually beneficial arrangements that exchange data for goods and services.  Typically, as long as the source of data has consented to the exchange, then there isn't a moral issue.

 However, this is not always true. Even supposing we have consented to the collection and use of certain kinds of data, there still could be a moral objection.  John Stuart Mill famously argued that we do not have the liberty to sell ourselves into slavery, for we would be using our freedom to make ourselves un-free (a sort of self-contradiction).  Similarly, I would argue that there are likely certain amounts, kinds, and uses of information that, even though we were to consent to the collection and use of them, would still be immoral as they would violate our moral worth and dignity, and hence, even our consent could not morally justify the action.  This is in spite of any other benefits, goods, or services that would result for us (consider similar arguments against prostitution).

Thus, it seems that in order to avoid these moral pitfalls in collecting and using various amounts and kinds of data, one must (at least):

• limit the amount of data gathered to morally acceptable limits
• limit the kind of data gathered to morally acceptable kinds
• use the data for morally acceptable uses
• obtain consent from individuals about whom the data is gathered for the collection and use of such data
• provide just compensation for the data in the form of goods and services to the source of the data

What these general limits are I cannot say, nor is it my point here to do so.  My point here is merely to argue that it is reasonable to believe that limits do exist.  As to what those general limits are, that would require much greater analysis than I have time for here.

A Deontological Critique

But perhaps we can say something useful about the specific scenario outlined above.  Does the above example satisfy the above criteria?  Let's explore.

The amount of data does not seem to be the issue here.  For the model above to work, no great amount of data regarding any single individual is required, although a large quantity of data is required (i.e., a single record for each person in the representative sample).

The kind of data being collected is a bit more concerning.  While country of origin isn't all that personal, religion is.  In some countries, that information being made public can mean the difference between life and death, freedom or persecution.  While this information may not be as sensitive as other kinds of information, it still is extremely personal.

The use of the data seems to be the primary concern.  Religion (and Country of Origin as a sort of proxy for religion) is being used to discriminate amongst immigrants, and this is cause for great concern.  Freedom of religion is protected by the First Amendment of the Constitution as a legal right.  Now such a right might not legally apply to immigrants (as they are not citizens of the United States), but it may morally apply.  If we believe that, morally speaking, all human beings have some right to freedom of religion, belief, or conscience, and this moral right is being restricted in some way without a direct conflict with a higher right, then that is cause for concern.

In this case, the government is treating certain country of origin/religion combinations as nearly guilty, even in the absence of any actual terrorist or criminal activity.  This reminds me of the movie Minority Report, in which individuals are arrested before they commit any crime, based on the prediction/foresight that they would commit the crime if not prevented.  Even though the program drives crime down to virtually zero, we are left with the impression that such a situation is unjust, for each supposed criminal has not yet committed any crime to be punished for.  A similar example comes from just war theory as opposed to pre-emptive strikes.  In just war theory, it is presumed that a wrong has been committed that needs correcting, whereas in a pre-emptive strike, no wrong has yet occurred even though it is perceived as imminent.  However, as many people note, while under the guise of a defensive measure against a likely aggression, a pre-emptive strike actually flips the roles of aggressor and defender. 

In our scenario, because a person may be a terrorist, he or she is treated as such, even though no actual terrorism has been committed by that person.  Joined with the facts that the reason for believing that he or she is a terrorist seems to be exclusively that he or she is Muslim and/or from a predominately Muslim country, and a person's religion is extremely personal and whose freedom of religion is a moral right, this use of data clearly seems to be unjustified and immoral, even if it would bring about great good (i.e., the prevention of terrorist activity).

The only escape from this argument I can see is that one must argue that there is a direct rights conflict, and that the right to life of US citizens takes precedence over the moral right to liberty of immigrants in their immigration and religion.  Is there a direct rights conflict?  I don't think so.  One has to say that there is no effective way to prevent or fight terrorism other than to ban all immigrants from certain countries on that basis alone, which does not seem reasonable. 

What is more reasonable is that there is no effective way to fight terrorism without taking into consideration religion and country of origin as part of the data used to identify terrorists.  That seems more reasonable, but it would hardly justify banning everyone from a certain country of origin or from a certain religion, as again, most Muslims from these countries are not terrorists.  Still, it may justify using these data fields as part of the overall data used to identify terrorists.

However, as mentioned previously, other kinds of information are likely much more informative in predicting whether or not someone is a terrorist, and if this sort of information is not sensitive or extremely personal in nature, then it can and should be used instead of the more sensitive information.  Thus, it is reasonable to believe that there are effective ways of fighting terrorism that do not rely exclusively or perhaps at all on this kind of sensitive data, and if so, there is no direct rights conflict between the right to life and the right to liberty.  In short, it seems extremely unlikely that we are forced to choose between upholding the right to freedom of religion (i.e., liberty) and right to life in this matter.  Or in other words, a claim that a failure to deny the freedom of religion and immigration to immigrants leads directly to the denial of the right to life of US citizens seems extremely implausible.

Has consent been obtained from the immigrants regarding the collection and use of such data?  It is hard to say.  Some information like country of origin is known publicly and explicitly through the immigration process.  Other information is perhaps voluntarily revealed as part of the application for immigration.  Would religion be part of this revealed information?  Perhaps.  Is it appropriate to ask about this as a condition for immigration?  Maybe, but it at least appears to be somewhat suspect given its sensitive and personal nature. 

Consent has been given as part of determining one's eligibility to immigrate to the United States.  But does this also mean a consent to using the data to fight against terrorism?  This is doubtful, unless eligibility means among other things that one is not a terrorist, which is reasonable: terrorists are not eligible to immigrate to the United States.  But one could argue that this is an instance in which even  the collection and use of such information, even though consented to, is still immoral given the kind of information being used and its intended use.

Has just compensation for the data in the form of goods and services been provided?  I believe so, if the exchange is the ability to immigrate to the United States.  It seems that this can be a just and reasonable compensation for the exchange of data.

Conclusion

So what can we conclude, if anything?  Much of what I have said above is surely debatable by reasonable and well informed people.  However, I believe the deontological concerns raised above suggest that even if preventing immigration from the several predominately Muslim countries stopped all terrorist activity that would have occurred otherwise, the practice itself would still be immoral given the highly sensitive nature of the data, how it is to be used, and the likelihood of using other kinds of data that are not sensitive but are still effective in fighting terrorism.  The consequentialist good of preventing terrorist activity cannot justify the, deontologically speaking, immoral means of collecting and using sensitive data like religion in this way.  In the absence of a direct rights conflict with a right to life, other effective and morally permissible ways of fighting and preventing terrorism must be found that do not infringe on the moral rights of immigrants in their liberty to immigrate and in their freedom of religion.

While I have focused on a specific example in the above discussion, a general point can be made that ties back to data science. The above shows that there may be cases, perhaps often, in which a predictive model may have to rely on data that from a collection, kind, or use perspective, is immoral, in order to achieve a desired level of effectiveness (e.g., accuracy, lift, precision, or other measure).  That is, data scientists will on occasion have to choose between building consequentially effective models and deontologically moral, though perhaps not as effective, models.  While such a choice can perhaps be avoided when informative but not morally suspect amounts, kinds, and uses of data exist, it seems that this won't always be the case.

Thus, it is important for data scientists to think ethically when doing their work. A model's predictive capabilities cannot be the sole criteria for judging whether it can be used or implemented.  We must also reflect on the moral implications.  Is the data ethical in amount and kind? Is it being used ethically?  Have the sources of data consented to its collection and use in this way?  Have they been justly compensated?  If the answer is no to any of these, are there ways to correct these issues or to use other kinds of data to achieve the same end?  If not, we may be forced to choose between being ethical and being effective.

Moral Dilemmas and the 2016 Presidential Election

Introduction

More than any year in recent history, Americans seemed to be faced with two bad choices for President.  And it just seems to be getting worse. Clinton's email scandals continue to increase while Trump's crude talk and inappropriate behavior has caused a rupture in the Republican party.  It seems like Americans have no good choice, maybe even no morally permissible choice.  But is this true?  Do we have a moral dilemma?

Ethical Theory

As I observe what especially Republican leaders are saying about their positions on Trump, I am reminded of my philosophical ethical training and how it's concepts may be useful for understanding where we are at in the political discourse.  The concepts come from normative ethical theory, which consists of views that largely fall into three camps:

• Consequentialism: whether an action is morally right or wrong depends solely on its consequences.
• Deontology: whether an action is morally right or wrong does not depend solely on its consequences.  Intentions, the inherent nature of the act, and other factors may contribute to its rightness or wrongness.
• Virtue ethics: whether an action is morally right or wrong depends on whether a morally virtuous agent would do it.  If so, the action is right.  If not, it is wrong.  The virtuous agent will usually consider consequences, the nature of the act, and so forth, but it is NOT necessary for two virtuous agents to arrive at the same conclusion.

With these theories in mind, think on what people have said about the two candidates. You have probably heard things like:

• "I don't like Donald Trump, but I can't vote for Hillary.  Trump is the lesser of the two evils."
• "Hillary isn't great, but she is better than Trump."
• "I can't vote for either candidate.  Voting for either of them would be wrong."

Consider this article for the Washington Post.  In it, Republican leaders say the following things regarding Trump:

• Newt Gingrich:
• “It’s time for him to send targeted messages to each district and state and have Republican voters ask their candidates: ‘Are you going to help us defeat Hillary Clinton?’ And Trump should make it clear that the side effect of not helping Trump is electing Hillary Clinton.”
• Paul Ryan:
• "Ryan said Monday that he would no longer defend or campaign with Trump. "
• Marco Rubio
• “I disagree with him on many things, but I disagree with his opponent on virtually everything,” Rubio said. “I wish we had better choices for President. But I do not want Hillary Clinton to be our next President. And therefore my position has not changed.”
• Marsha Blackburn
• “You don’t go after somebody who is, as Ronald Reagan would say, your 80 percent friend. What you do is stand with them... And it is not helpful to have this kind of drama going on. What you need to do is say we have a binary choice.”

Or this article on Jeb Bush:

• "I haven't made up my mind what I'm going to do — I'm going to vote and I've made it clear I'm not going to vote for Donald Trump or Hillary Clinton," Bush told reporters. "I'm not derelict in my civic duties, I'm going to vote, but the presidency is a place where, for whatever reason, I'm not comfortable supporting either party's nominee."

As seen above, some have taken the consequentialist route by advocating for the lesser of two evils (Gingrich, Rubio, Blackburn) while others have taken a more deontological approach (Ryan, Bush). 

Moral Dilemmas

How does this apply to moral dilemmas?  Let us simply define a moral dilemma to be a moral situation in which an agent must make a decision between two (or more) seemingly immoral courses of action.  How does the agent resolve the situation and make a decision?

The consequentialist always has a way out.  In fact, there is no such thing as a true moral dilemma for the consequentialist, since the right thing to do is always whatever is best consequentially (or least bad).  It is doubtful that the choices would be equally bad on one's estimation, so one can always determine which outcome to choose, no matter how bad it is.  Some of the above Republicans have made that decision: no matter what Trump does, Clinton would be consequentially worse, so they will vote for him.

The virtue ethicist always has a way out.  If you are a virtuous agent, then no matter what you decide, as long as you are making your decision exercising your virtue, you will make a correct or virtuous decision.  There may not be a the correct decision.  I bet that many such virtuous Americans will vote.  Some will vote for Clinton while other will vote for Trump, and in a certain sense, both kinds of votes will be morally correct.  Unless we are willing to say that no virtuous person could possibly vote for Trump, or no virtuous person could possibly vote for Clinton, then we have to allow for this possibility.

But what if one has a deontological bent?  What if one believes that voting for Clinton would be morally impermissible AND voting for Trump would also be morally impermissible?  That is, both candidates have crossed a certain moral threshold such that it would be wrong to vote for either?  This seems to be the stance that Jeb Bush, Paul Ryan, and other Republicans that have stepped away from Trump are taking.

The options available to those that find themselves in such a position seem to be:

• Don't vote: if you don't vote for either of them, you are not morally culpable.
• Vote for a third party: if you can't vote for either major party candidate and you want to or believe you must vote, then you have to find a third option to vote for.

In this way, the deontologist refuses to accept the terms of the moral dilemma as a forced choice between two or more morally impermissible options.  Instead, he or she finds another alternative or refuses to make any choice at all.

Conclusion

Truth be told, I am finding myself more and more in the deontological camp on this issue.  So I will likely vote third party.  For whom?   I have no idea.  But I feel much like Jeb Bush right now, and it is very sad.  In some sense, it doesn't really matter for me as I live in Washington state and it will assuredly be won by Clinton, so the pressure is off.  But if I were in a swing state, what would I do?  Go the consequentialist route and vote for the candidate I think will do the least harm no matter how much I dislike the candidate?  Or maintain ethical purity and vote for a third party, knowing full well that similar actions by others may result in my least favorite candidate taking office?

It's a tough choice.  It's a moral dilemma.

A Defense of Normative Counterfactual Negative Causation

I wrote this paper for a Metaphysics class while completing my MA in Philosophy.  This was submitted on December 6, 2010 and remains as I submitted it (apart from formatting changes).  I probably would change a lot in this paper if I were to revisit this topic as I am not sure that I agree with myself anymore.  Perhaps someday I will write a response.
_______________________________________________________________________________


           In the last twenty years, the topic of omissions and preventions has been given much discussion in the causation literature.   David Lewis argued exclusively for their existence in one of his last published papers, while D.M. Armstrong and Helen Beebee have both strongly objected to his claims.  Recently, Jonathan Schaffer has taken up the defense of omissions and preventions and has been engaged in a back and forth debate with Phil Dowe, their most prominent critic.  While the topic is much too broad to discuss in its entirety, this paper attempts to develop a consistent defense of the existence of omissions and preventions against their usual criticisms by assuming the legitimacy of counterfactual causation and the ontological existence of normative claims.  In so doing, I will show that such negative causation, when it involves a normative component in the description of the causal event, is a genuinely metaphysical relation between events that is worthy not only of conceptual, linguistic, and pragmatic acceptance, but of ontological acceptance as well.  I will also show that such an account can guide our ascriptions of moral responsibility to agents in instances of potential omissions and preventions.

          Most theories of causation fall into two somewhat opposing categories.  One contains theories that analyze causation in terms of counterfactual dependence while the other contains singularist theories that hold causation to be an intrinsic relation between singular events.[1]  Considering the later, singularist theories argue that causal relations among particulars lead to regularities and patterns but are not dependent on them.  Two events C and E are necessarily related in virtue of their intrinsic properties and local factors.  In other words, there is a physical connection between C and E.  Given this physical connection, causal relations exist only between positive events and entities, that is, between factors that are empirically detectable.

          On the contrary, the counterfactual theory of causation holds that particular causation occurs conditionally in virtue of causal regularities and patterns; causation is an extrinsic relation between events that depends on factors external to the related events themselves.  As most famously advanced by Lewis, these causal relations can be analyzed and reduced to causal dependence, where causal dependence holds between two events C and E if and only if there is a chain of counterfactually dependent events linking C to E.  In other words, “causal dependence holds between events if and only if the counterfactual ‘if C had not occurred, then E would not have occurred’ holds” (Armstrong 445).  This counterfactual is true if and only if there is a possible world in which C does not occur and E does not occur and this world is closer in similarity to the actual world than any other possible world in which C does not occur but E does occur.  Such similarity is largely due to the laws of nature (Lewis 279).[2]

          Until fairly recently, such theories of causation have largely focused on giving an account of the relation between positive events, i.e., those events that actually occur.  However, our causal language often includes talk of negative events, that is, “events whose essence is the absence of a property or particular” (Beebee 291).  Omissions and preventions are examples of these.  Preventions are “causes of what they prevent, ‘effects’ that never occur.  Omissions are things that are not done but that have positive effects” (Armstrong 447).  For example, my failure to rescue the drowning child in the pool before me could be said to have caused that child’s death.  In this case, my omission brought about the drowning.  However, suppose that I dove into the pool and pulled the child out and it did not die.  Then my action can be said to have prevented the child’s death and that I caused the child to not die.  In each case, the supposed causal interaction involves a negative event, one that does not actually occur.  The singularist must reject the intuition that actual causation occurs in this situation.  On the other hand, the counterfactualist tends to allow that such absence or negative causation does occur and that negative events are ontologically legitimate.

           However, such an affirmation faces many problems.  First the singularist argues that negative events do not in fact exist and so they cannot figure into causal relations about the actual world.  As Lewis writes, “a relation requires relata” (Lewis 281).  Without ontological relata, a causal relation cannot be ontologically acceptable.  The proposed relation may be pragmatically, conceptually, or linguistically useful, but it will not give an account of what actually occurs in the real world since it will not actually exist.

          Second, even supposing that negative events exist, there is no causal chain linking the positive and negative events together.  For example, suppose that, through a father’s negligence, a child is hit by a car after running into the street.  Dowe points out, “it may not be true that had the father not done those actual positive things he did the accident would not have occurred since there are other ways that the father could have failed to prevent the accident” (Dowe, “A Counterfactual” 220).  Thus, it might be the case that counterfactual dependence does not exist between an omission and its cause or a prevention and its effect, and without this dependence, causation does not occur.           

          Third, one might ask where these preventions and omissions occur.  In his criticism, Dowe gives three possibilities: there, here, or everywhere.  To be located “there” is for an absence to be “located at the location of the actually occurring alternative to the absent event” (Dowe, “Absences” 12). For example, this is where the father is when his child is hit by the car.  Dowe objects that this cannot be the right location since the father is separated in space from the accident, which violates the assumed fact that causation at a distance does not occur.  To be located “here” is for an absence to be located where the absence would have occurred.  For example, this is where the father would have grabbed his child and thus prevented him from being hit.  But, Dowe objects, this is also separated in space from the accident.  Finally, to be located “everywhere” is to be “not located at all” and thus a violation of the claim that causes and effects must be located (Dowe, “Absences” 16).  Since prevention and omissions are neither here, there, nor everywhere, Dowe concludes that they do not exist. 

          A fourth charge is that if omissions and preventions are causes then the actual amount of omissions and preventions is far greater than our intuitions will allow.  Dowe points out that Kennedy’s bodyguard, you, me, and Julius Caesar all could have pushed the gun away if we had been on hand, so we are all the cause of Kennedy’s death, “and so too for every event that has ever occurred that could have been prevented by human intervention” (Dowe, “The Power” 7).  If Dowe is right, then there are too many omissions and preventions to count as causes and therefore the view that they are causes is absurd. 

          Fifth, supposing that omissions and preventions are legitimate causes, can the counterfactualist maintain a moral distinction between, say, killing and letting die?  If there is a moral distinction between some direct causes and some indirect or absent causes as if often thought, then the counterfactualist must be able to explain what that difference is and why it makes a moral difference.  If she cannot, then a singularist will have the advantage in explaining this moral intuition.

          Several distinctions need to be made that have been largely neglected in the discussion of this topic.  The first is the distinction between what “causation” means as it is commonly used and what it ought to mean.  The former focuses on usage and as such is a semantic claim about the word “causation.”  In this, the counterfactualist meaning clearly matches our usage much better than the singularist meaning because it allows for absence causation (see Schaffer “Causes need”).  Singularists may charge that this is not what “causation” ought to mean because it does not match the metaphysical reality of causation, but that is a separate question from what it means in practice.  It is important to not conflate these two distinctions. 

          Second, one needs to distinguish between “a cause” and “the cause.”  Almost nothing can claim to be the cause of any event.  Instead, most events have several causes that together cause them.  When we single out a cause as being the cause, it is often done for practical purposes that serve human needs and interests.  However, we want to give an account of causation that is not merely pragmatic but objectively true; we want an ontological analysis of causation.  Thus, it may be appropriate to speak of a cause as being the primary cause of an event if the majority of that event’s causes can be traced back to this cause, but it will still be considered a cause, one of many causes, of the event.[3]

          Third, one needs to distinguish between negative causation by agents and negative causation by events.  We should not presuppose that an example of agent causation can be translated into a generalized formulation of event causation.  Within negative causation by agents, one also needs to distinguish between different kinds of normative agents and which normative claims relevantly apply to them.  For example, my oven thermometer may fail to register the correct temperature and thereby cause my turkey to be burned, but I would not hold it morally responsible for that result.  That is, I would not ascribe to it the property of moral blameworthiness for its failure to function as it ought to.

          Fourth, one should note that causal responsibility is distinct from normative responsibility.  It does seem to be always true that if A is (morally) responsible for B, then A caused (or was a cause) of B.  We only hold people morally responsible for an ensuing event if they were somehow a cause of it and it does not make sense to hold someone responsible for an event for which they were not a cause.  However, it is clear that it is not always true that if A caused (or was a cause of) B, then A is (morally) responsible for B (e.g., my breathing may have been a necessary condition for the occurrence of a tornado, but usually I am not morally responsible for the damage it causes).  But the conditional is not always false either.  Why, then, do we hold others morally responsible in some cases but not in others when in both cases they are a cause of the ensuing event?  That is, why does an agent acquire the property of being morally blameworthy and why is her action singled out among the many causes of an event in some instances but not in others?  Since neither intention nor primary causality is necessary for moral responsibility, it seems that causality is turned into responsibility by the violation of a norm; what is caused is normatively significant and hence its initiator acquires a normative property.  Thus, we can say that an agent A is responsible for B if A is a cause of B and A violates a norm related to B’s occurrence. 

          However, while this is easily applied to agents, it is more difficult to apply it to events in general and omissions and preventions in particular.   One can immediately object that events are not “responsible” for other events and that causation, like these events, is not normative.  This is the position that Helen Beebee takes.  Beebee says that a normative analysis of causation “doesn't look like the sort of analysis we ought to be giving of the metaphysics of causation by absence” (Beebee 296).  She goes on to assert that “nobody within the tradition of the metaphysics of causation that I'm concerned with here thinks that causal facts depend on human dependent norms,” and that adding normativity “would in any case make the truth of causal claims turn out to be a relative matter” (Beebee 297).  In other words, Beebee believes that causation is an objective metaphysical relation that is not dependent on human subjectivity.  In order to maintain this objectivity, we cannot admit relative normative claims into our analysis.

          However, it seems possible to maintain the objectivity of causation even if we permit moral claims and other normative notions into our analysis.  If normative claims are not relative or dependent on humans but instead are true in virtue of ontologically real normative properties or statistical regularities in the world, then these claims can factor into an objective analysis of causation.  Indeed, the counterfactual analysis of causation is based on Hume’s regularity analysis.  Or, if one does not want to make causation an intrinsically normative relation, one can locate the normative component in the events that are related and thereby maintain the objective status of causation.  Thus, the counterfactual analysis of causation can easily accommodate a normative component in an acceptable way.

          Beebee brings up another related objection.  She claims that “how often something has to happen in order to count as ‘normal’- and hence for its absence to count as abnormal - is always going to be rather fluid” (Beebee 297).  She is concerned that there is no genuine unity to our concept of “normal.”  McGrath counters this worry by noticing that we speak of something as normal by referencing an appropriate objective standard (McGrath 138).  For example, with promises, we appeal to a moral standard; with alarm clocks, a design or artifactual standard; with hearts, a design or biological standard; with rain, a weather-cycle standard based on statistical regularities and laws of nature (McGrath 138).  These standards are different for each domain and pose different requirements for what counts as normal.  However, common to each example is a measurement of how well it matches the standard in its domain, and this commonality gives unity to the notion of “normal.”  Given the foregoing reasons, it seems possible to have an objective notion of “normal” and hence, normative claims can play a role in causal claims.   

          At this point, we can consider the first objection: negative events do not in fact exist and so they cannot figure into causal relations about the actual world.  Schaffer offers a solution, as suggested by Lewis, that we can “regard the absence description as a way of referring to a present event” (Schaffer, “Causes need” 212).  For example, the father’s failure to prevent his child from running into the road is a way of describing what the father actually did (e.g., take a nap, read a book).  Similarly, supposing that the father did restrain the child and thereby caused the accident to not occur, the event of the accident’s non-occurrence refers to the positive event of whatever the child did (e.g., sit down, stay on the lawn).  However, this leads to our second objection that there is no causal link between the negative event and the positive event.  For instance, suppose that the father took a nap and the child ran into the street and was hit.  It is not necessarily true that, if the father had not taken a nap, then the child would not have been hit.  The father could have read a book or watched TV instead of napping.  Thus, there is no causal dependence between the referred-to positive event and its supposed effect.   

          We need a way to turn the positive event “the father’s taking a nap” and the negative event “the father’s not being watchful” into one positive event that gives counterfactual dependence.  The solution is to include the normative component of the negative event into the positive event or to give a positive event that is normative.  For instance, the event “the father’s taking a nap” is not an ordinary nap-taking.  In context, this event violates a moral norm since it is a negligent nap-taking.  It would be better to describe the event as “the father’s taking a nap and not being a watchful father.”  Thus, “the father’s taking a nap and not being a watchful father” caused the accident because, if the father had not been taking a nap or had been a watchful father, then the accident would not have occurred.  We might also characterize the event as an instance of negligence in which the father’s taking a nap is a description of that instance.[4]  In other words, the true event is not the father’s taking a nap, but the father’s being negligent.  Instead, the nap describes how he was negligent.  Thus, the father’s negligence caused the accident because, if the father had not been negligent, then the accident would not have occurred.   This is true in the most similar possible world to the actual world, and so there is causal dependence and hence causation.[5]

          If normative claims and normative properties are ontologically real, then this event is a real event.  To say that a father is “watchful” or “negligent” is to attribute real moral properties to him and to say that he fulfills or does not fulfill his real moral obligations. In addition, the father did take a nap, so this event is positive in nature; it really exists.  The actually occurring event provides the positive relata and the normative component or property of the event provides the counterfactual dependence. Together, they secure a genuine causal relation between these two events.  This causal relation is ontologically real and not merely linguistic, pragmatic, or conceptual.

          If the foregoing is correct, we can easily address the other problems lodged against this position.  First, one can appeal to the normative component or property of an event to limit the number of omissions and preventions.  Even Lewis subtly appeals to this when he claims that someone cast into the void dies because “he would […] have been surrounded by the life-sustaining objects that normally surround us” (Lewis 282).  Normative claims can pick out which agents and events have a normative relationship to the ensuing event.  For instance, it is not my nap, nor your nap, nor anybody else’s nap the causes the child to be hit by a car.  It is the father’s nap and his lack of positive moral properties in addition to his moral obligation to his child that makes the father a cause of the accident.  I may have napped, and I may not have good moral character, but since I do not stand in a relevant normative relation to this child, I am not a cause of the accident.

          Second, we can say that the omission occurred where the referred-to positive event occurred (e.g., wherever the father is taking a nap when the accident happens), since it is this event that stands in a causal relation to the effect.  Similarly, a prevention occurs at the positive point of prevention (e.g., where the father grabs the child to prevent it from running into the road).  Dowe will object that in both cases, the cause is separated in space from its effect.  But this only poses a problem if causes are physically connected to their effects (as Dowe assumes).  The counterfactual theory of causation does not require a cause and effect to be physically connected, but only that a chain of counterfactual dependence exists between them.  Since this criterion is met, causation at a distance is legitimate.

          Third, the counterfactualist can make a distinction between positive and negative causation that is morally relevant by referencing the normative component of the events involved.  For example, suppose it is true that “one should never kill an innocent person.” Then the event “the physician’s administration of a lethal dose” caused the event “the patient died.”  The moral norm can be added to the first event, and though it does not make any causal difference, it makes a difference with respect to responsibility.  In this case, the physician is morally responsible for the patient’s death.  However, the event “the physician pulled the plug” caused the event “the patient died” only if there is a moral norm against pulling the plug.  Assuming there isn’t, then this is not a genuine case of causation, that is, the pulling of the plug did not cause the patient to die.  Instead, the ailment that the patient was suffering from caused the patient to die.  But if there was a moral norm against pulling plugs or against letting people die, then this would count as genuine causation.  We could say that the event “the physician’s pulling the plug and his not being a moral practitioner” caused the event “the patient died” because if the physician had not pulled the plug or had been a moral practitioner, then the patient would not have died.  We could also go on to say that the physician was morally responsible for the patient’s death since his pulling the plug violated a moral norm and caused the patient’s death.

          In his conclusion, Lewis asks: “What is causation? As a matter of analytic necessity, across all possible worlds, what is the unified necessary and sufficient condition for causation? - It is somehow a matter of counterfactual dependence of events (or absences) on other events (or absences)” (Lewis 287).  This paper has attempted to defend this position specifically in instances of absences.  It has been argued that if normative claims are true in virtue of ontologically real normative properties or statistical regularities in the world, then these claims can factor into an objective analysis of causation.  On such an analysis for omissions and preventions, one can combine the normative component of a negative event with its positive referring event to create a positive event that is normative.  This normative and positive event can then figure into causal dependence and hence causation itself.  Thus, omissions and preventions are, properly construed, examples of genuine causation.

Works Cited and Consulted

Armstrong, David. “Going Through the Open Door Again.” Causation and Counterfactuals. Ed.

John Collins, Edward Hall, and Laurie Paul.  Cambridge, MA: MIT Press, 2004. 445-58. Print.

Beebee, Helen. “Causing and Nothingness.” Causation and Counterfactuals. Ed. John Collins,

Edward Hall, and Laurie Paul.  Cambridge, MA: MIT Press, 2004. 291-308. Print.

Dowe, Phil. “Absences, Possible Causation, and the Problem of Non-locality.” Phil Dowe:
          Selected Work on Causation. N.p., Jan. 2009. 1-23. Web. 1 Oct. 2010. <http://homepage.
          mac.com/pdowe/Research/cause.html>.

---. “Causes are Physically Connected to their Effects: Why Preventers and Omissions
          are not Causes.” Contemporary Debates in Philosophy of Science. Ed.          
          Christopher Hitchcock. Malden, MA: Blackwell Publishing, 2004: 189-96. Print.

---.  “A Counterfactual Theory of Prevention and ‘Causation’ by Omission.” Australasian Journal of
          Philosophy 79.2 (2001): 216-26.  Philosopher’s Index. Web. 1 Oct. 2010.

---. “The Power of Possible Causation.” Phil Dowe: Selected Work on Causation. N.p.,
          Jan. 2009. 1-22. Web. 1 Oct. 2010. <http://homepage.mac.com/pdowe/Research/
          cause.html>.

Lewis, David. “Void and Object.” Causation and Counterfactuals. Ed. John Collins, Edward

Hall, and Laurie Paul.  Cambridge, MA: MIT Press, 2004. 277-90. Print.

Lowe, E.J. A Survey of Metaphysics. New York: Oxford University Press, 2002. Print.

McGrath, Sarah. “Causation by Omission: A Dilemma.” Philosophical Studies 123 (2005): 125-
          48.  Philosopher’s Index. Web. 1 Oct. 2010.

Schaffer, Jonathan.  “Causation by Disconnection.” Philosophy of Science 67 (2000): 285-300.
          Philosopher’s Index. Web. 1 Oct. 2010.

---. “Causes need not be Physically Connected to their Effects: The Case for
          Negative Causation.” Contemporary Debates in Philosophy of Science. Ed. Christopher          
          Hitchcock. Malden, MA: Blackwell Publishing, 2004: 197-216. Print.

--------------------------------------------------------------------------------

[1] For the purposes of this paper, I will ignore probabilistic theories and I will be assuming a largely deterministic universe, although my argument seems to be neutral between determinism and indeterminism.

[2] Lewis ultimately rejected this specific formulation in favor of his influence account of causation.  However, my argument does not depend on the specific details of a counterfactual account of causation, but only on the general framework that is common to all such accounts.  That is, that causation is a matter of counterfactual dependence between events.

[3] This point may be contested as some may insist that there is always a “the cause” of an event.  However, my argument does not seem to require taking a side on this issue.  Rather, the point is made in order to be clearer about the terms being used in the debate.

[4] Thanks to Andrew Morgan for offering this point.

[5] This account can avail itself to the same modifications and strategies that all counterfactual theories of causation use to avoid cases of preemption and other similar issues.  However, this is a side issue to the main point of this paper, which assumes the legitimacy of counterfactual theories of causation in general.