Sunday, April 29, 2012

Theism and the Meaning of Life (Part Two)



(Part One)

This post is the second part in a short series about the meaning of life. The series is working primarily off Gianluca di Muzio’s article “Theism and the Meaning of Life”, which appeared in the journal Ars Disputandi back in 2006. However, I’m trying to develop some formal reconstructions of Di Muzio’s arguments so as to improve my understanding of the arguments in this debate. My central contention, which is worth keeping in mind while reading this post, is that the debate is concerned with finding the necessary and sufficient conditions for a life that is worth living (i.e. meaningful = worthwhile).

The previous post was taken up with addressing William Lane Craig’s claims about the role that theism and immortality play in making life worth living. Roughly, Craig argues that without God and without immortality there can be no meaning in life. Since neither of these things are present in the atheistic-naturalistic worldview, he concludes that that worldview is practically impossible: no one could continue living in the knowledge that life was meaningless.

Burrowing a little deeper into his claims, we suggested that his argument reduced to the principle that one’s life can only have meaning if it plays a permanent role in some higher plan. I argued that this principle was flawed because, on the basis of some simple counterexamples, it appears that immortality and the existence of a higher plan are insufficient for meaning in life.

This left open the possibility that there could be some other, non-theistic, account of the necessary and sufficient conditions for meaning. In this post, we will consider Di Muzio’s proposal for just such an account.


1. The Goods-based Account of Meaning
As we saw at the end of the last post, there are really only two major possibilities when it comes to developing an account of meaning. The first possibility is to argue that everything has relative meaning. That is to say, it derives its meaning from something else. This possibility is unattractive because it leads to an infinite regress, something that Craig himself wishes to avoid. That leaves us with the second possibility: that some things are intrinsically meaningful. That is to say, they provide both the necessary and sufficient conditions for a life worth living in and of themselves.

Theists may wish to argue that God’s higher plan is the only thing that could have this kind of intrinsic meaning. Without investigating any proposed divine plans (e.g. building an eternal kingdom) in detail, we already have some prima facie reasons to doubt this: many higher plans seem to lack intrinsic meaning, so why think God’s plan is any different. So let’s explore another possibility. The possibility that there are some things, here on earth, that are intrinsically worthwhile.

Di Muzio argues that an obvious candidate for things that are intrinsically worthwhile are those things that are morally valuable. Indeed, worthwhileness and moral value seem to be natural bedfellows: to say that X is morally valuable seems to entail that it is worth doing. What’s more, this entailment relation would seem to hold even if the person pursuing that which is morally valuable has no role to play in a higher plan.

There are two ways to argue this. One is to imagine an individual life — e.g. the life of the charity worker — which pursues something that is morally valuable but which, for some reason, is excluded from God’s plan. Then ask the question: would that life suddenly seem worthless if it lacked a role in God’s plan? Di Muzio says “no”: if there is recognisable value in what the person has done, then it would be absurd to deny that their life was worthwhile. This is true even for theists because they typically recognise that generosity and charity are valuable.

Still, some hardy theists might resist the negative answer so there’s another tack one can take. Recall from earlier that God’s plan has to have intrinsic meaning lest we risk an infinite regress. So ask the question: what would make it the case that God’s plan is intrinsically meaningful? It can’t be simply because it’s his plan: that would be stipulative, arbitrary and vacuous. The only other possibility is that, in formulating and executing his plan, God is pursuing some goal or end that is intrinsically valuable. In other words, even the meaningfulness of God’s plan seems like it has to derive from the pursuit of something (some end or good) which is intrinsically valuable.

(Note: Commenter TaiChi makes an extended argument of this sort in his reductio of Craig’s position in part one of this series.)

These two arguments give support to what we can call the goods-based account of meaning. According to this account, the following principle of meaning holds true:

Access to Value Principle (AVP): If a person has access to intrinsically valuable activities in their lives, then their lives are meaningful

This principle identifies for us the sufficient condition for meaning, namely: access to value. It should be noted that access to moral value is not the only thing that is included in this account. Indeed, classically it was held that three things have intrinsic value: the good, the true, and the beautiful. How exactly they manage to do this is a topic taken up in Thaddeus Meutz’s recent article in Religious Studies. I’ll not discuss these here. All I will suggest is that any life that has access to those things would, on the goods-based account, be meaningful.

Although the AVP denies that God is sufficient for meaning, it does not deny that God is necessary for meaning. An additional argument is needed to prove this, as follows:


  • (1) If a person has access to intrinsically valuable activities in their lives, then their lives are meaningful (AVP). 
  • (2) Even if God does not exist, a person will have access to intrinsically valuable activities. 
  • (3) Therefore, even if God does not exist, a person’s life is meaningful. 
  • (4) Therefore, God is not a necessary condition for meaning.


No doubt theists will wish to deny the truth of (2). They will argue that God is the ontological grounding of all that is valuable. Without him, there can be no access to value. I’ve already discussed this claim at length elsewhere. To summarise, I do not think it is plausible. I think values neither need nor require an ontological grounding of the sort envisaged by theists. Thus, this rebuttal does not work.
So, with the help of the AVP and the preceding argument, we end up where we wanted to be: with an account of meaning that denies that God is either necessary or sufficient for meaning.


2. Immortality and Fragility
In effect, we have now successfully defeated Craig’s claim that both God and immortality are necessary (and sufficient) for meaning in life. Still, there are some lingering doubts. In particular, there are some doubts about the role of immortality in our account of meaning. Even though we now know that immortality cannot in itself make our lives meaningful — the oft-cited example of Sisyphus pushing his rock up a hill for eternity is enough to show this — might it be the case that immortality provides additional meaning? In other words, might it be true that immortality makes life more meaningful than it might otherwise be?

Many are seduced by this line of reasoning, atheists and theists alike. Consider the decision problems illustrated below. In the first decision problem, we must choose between an immortal life with no access to value (like the life of Sisyphus or an eternity in Hell), a mortal life with access to value, and no life at all. In this case, the preferable option seems obvious: the finite life with access to value would be more worthwhile than the others. But things are not so straightforward in the second decision problem. There, we must choose between an immortal life with access to value, a mortal life with access to value, and no life at all. When confronted with those options we might be tempted to say that the immortal life is to be preferred to the mortal one.



And this might provide a backdoor for the theist. How so? Well, the theist can argue that if an immortal life (with access to value) is more meaningful than a mortal one (with access to value), then God is the only being that can provide us with it. For what is heaven if not (on the common view) an eternal kingdom in which we can carry on accessing the values — love, friendship, family, the pursuit of knowledge — that we access here on earth, only forever and ever and ever.

Di Muzio frames this argument in terms of something he calls “fragility”. Basically, he argues that the problem for the atheistic view is that our access to value is fragile. It can easily be taken away or reversed by calamitous events. He gives the real example of a group of children, happily playing in a yard, being bombed to death in Sarajevo in the early 90s. Although they lived in conditions that were sufficient for meaning (Di Muzio is clear about this), they did so for a fragile and fleeting moment. Would it not be better if they were given an immortal existence in heaven?

Call this the problem of fragility:


  • (5) A life in which one’s access to value is fragile (i.e. can be taken away) is less meaningful than a life in which one’s access to value is permanent. 
  • (6) On atheistic-naturalism, everyone’s access to value is fragile (because we are all mortal). 
  • (7) On theism, at least some people’s access to value is permanent (because of immortality in heaven). 
  • (8) Therefore, life is more meaningful on theism than it is under atheistic-naturalism.


Now let’s reiterate, this argument does not say that theism is either necessary or sufficient for meaning — the AVP is still doing all the heavy lifting in that regard. But it does suggest that there is more meaning in a theistic universe because it’s possible that God can create a realm in which access to value continues forever.

As it happens, Di Muzio seems to be happy to reach this conclusion. His intent was to show that the theistic conception of meaning proposed by Craig fails, and that really the theistic view of meaning is parasitic upon the AVP. He does not need to go further and argue that theism is antagonistic to meaning. But I am not so happy to reach this conclusion. It’s not that I have any deep objections to it, it’s just that I think it’s worth exploring whether the above argument goes wrong at any point.
And, indeed, there seem to be three possible flaws to me.

First, premise (1) might be wrong. As Bernard Williams famously argued, an unending life might lose its worthwhileness by causing tedium: a dulling of one’s appreciation for that which is valuable. In a footnote, Di Muzio rejects this view, citing some more recent objections to Williams’ argument. There is a rich literature on this point that I don’t have time to explore today (perhaps some other day). Still, I raise Williams’s argument as being one way in which to rebut premise (1).

Second, premise (2) might be wrong. There are several in the transhumanist community who are actively working on projects that would help us to defeat old-age and death. The most famous of these is, of course, Aubrey de Grey, whose views I discussed a while back. These projects may be largely aspirational at this point, but they nevertheless point to the possibility that (2) could be false. Some might object that even if those projects do succeed, we would not achieve true immortality since our lives would be constrained by the lifespan of the universe. There are two responses to this: (i) our universe may be future-infinite (I’m not sure what the best current theories say on the matter) and (ii) we would still have plenty of additional time to play around with, enough to make our lives seem very worthwhile indeed (perhaps more so than life in heaven, if one takes into account the problem of uncertainty to which I allude below). A more serious concern about these life-extension projects is that they may force us to become excessively risk-averse and thereby lose access to several otherwise valuable activities. So, in a sense, they might make us more aware of our fragility.

Third, premise (3) might be wrong. After all, it’s possible that heaven is nothing like “business as usual”. What God has in store for us may not involve unending access to intrinsically valuable activities. I don’t know what else to say about this since any speculation about what heaven might be strikes me as being slightly silly. Presumably we just have no idea. But this very uncertainty might make the atheistic view (if it came with considerable life-extension) the more attractive one: better to go with what you know than take a risk on something else.


3. Conclusion
That brings us to the end of this series. To briefly recap, in part one I (with considerable help from Di Muzio) considered the arguments of William Lane Craig about theism and the meaning of life. Craig proposes that meaning depends on whether one has a permanent role to play in a higher plan. If one has such a permanent role, then one’s life is meaningful; if one does not have such a role, it it is not meaningful. Craig followed this by arguing that, on theism, there is such a permanent role to play, but on atheism, there is not.

I highlighted several flaws in this proposal. Using some thought experiments, I argued that having a permanent role in a higher plan was not sufficient for meaning. This suggested that something else provided the sufficient conditions for meaning. Following Di Muzio, a goods-based account of meaning was proffered. According to this theory, one’s life has meaning as long as one has access to things that are intrinsically valuable (the good, the true, the beautiful). Using some plausible premises, this account can then be used to show that God is not necessary for meaning.

But there is a lingering problem for the atheistic view. While it may not be the case that God is necessary or sufficient for meaning, he might still make life more meaningful than it could be by providing us with an eternal realm in which we can access value (heaven). However, as we just saw, this argument may have its own problems.

Saturday, April 28, 2012

Attempted Crimes and Changes of Mind


(Series Index)

Since my thesis was (partly) on the philosophy of responsibility, and since I teach a bit on criminal law, I occasionally like to look at some of the philosophical aspects of criminal liability. Indeed, these occasional looks have accumulated over the months, so much so that they now form a “series”. This post is part of that series. It looks at liability for attempted crimes, specifically at attempts involving a change of mind.

The post is based on an article by Gideon Yaffe, entitled “Attempt, Risk Creation and Change of Mind: Reflections on Herzog”. Yaffe is one of the best contemporary philosophers of criminal law and has written a book-length analysis of attempt liability. In the book (which I have not read), he presents an original theory of what constitutes an attempt. The aforementioned article offers a capsule summary of this theory and applies it to a rather interesting fictional scenario: Herzog’s plan to kill his wife and her lover, from Saul Bellow’s novel Herzog.

In what follows, I will first discuss some of the basic aspects of attempt liability and describe the scenario from Herzog. I will then present, in broad outline, Yaffe’s theory of attempts, before finally considering how it applies to Herzog’s change of mind. It should be noted that I will exclude a significant portion of Yaffe’s paper from the following discussion. That is the portion of the paper that considers (and rejects) the possibility that attempts should be viewed as crimes of risk creation. While interesting, that discussion would distract from my main purpose in this post, which is to consider attempts and changes of mind.


1. The Everyday Logic of Attempts
In the criminal law, a “crime” is consists of “elements”. There are two main kinds of elements in every criminal offence: actus reus (external) elements, which describe the actions and external circumstances of the crime; and mens rea (fault) elements, which cover the state of mind that accompanies the actus reus. Take the example of murder. Typically, this is thought to consist of two actus reus elements and one mens rea element. On the actus reus side we have: (a) the death of one person (b) caused by the actions of the defendant. And on the mens rea side we have: (c) the intention to kill or cause grievous bodily harm. These three elements constitute what we shall call the completed crime of murder. If a defendant happens to satisfy all three elements, and has no viable defence, then they will be liable for the completed crime.

But what happens if the crime is not completed? What happens if it is merely attempted? For instance, suppose someone buys a gun and shoots it at another person, intending to kill them, but misses. Are they not liable for something? Yes; certainly: they are liable for attempted murder. Every legal system of which I am aware includes liability for attempted crimes of this sort. So, even if the elements of the offence are not satisfied, one can be held liable for attempting to satisfy those elements.

The question then arises: what grounds do we have for holding people liable for attempts? I have considered some possibilities in a previous post. Here, I look to Yaffe’s answer, which is complex, but derives in part from the ordinary, everyday, morality of attempts. He argues that whenever we prohibit an action, it seems to follow automatically that we prohibit any attempts to perform that action. We can call this the ordinary morality of attempts (OMA) principle:

OMA-Principle: If we prohibit an action X, then we also prohibit any attempt to do X.

Support from OMA comes from thought experiments suggesting that its denial would be somewhat absurd. Consider, for instance, the father who prohibits his infant daughter from jumping up and down on the sofa. Moments later, she starts to climb onto the sofa with the intention of jumping on it. Could the child argue that she has not breached the prohibition merely because she has only attempted to jump on the sofa, but has not completed the prohibited act? In a technical sense, perhaps; but in a deeper, more morally salient sense, it would seem silly to say she could. Yaffe suggests that the criminal law tracks OMA in its prohibition of attempts as well as completed crimes.

So far so good. But now an interesting problem for attempt liability presents itself. What happens if the person begins to attempt an action, but then gets a change of mind (a sudden “pang of conscience” perhaps)? Are they still liable for the attempt? Bellow’s Herzog presents us with this problem. In the novel, we are told that Herzog is angered by his ex-wife and her lover and plans to kill them. He obtains a gun, drives to their house and watches them through the window. We are told that he imagines killing them on the drive over and while peering through the window. But then a curious thing happens. As he sees his ex-wife’s lover bathing his daughter with kindness and humour, he gets a change of heart and walks away. He never follows through on his initial intention to kill.

The sequence of events is depicted in the diagram below. The red line represents the point at which his change of mind intervenes and prevents him from completing the crime.



We want to know two things. First, is Herzog liable for attempted murder? And, second, if so, should his punishment be commuted or lessened because he changed his mind?


2. Yaffe’s Theory of Attempts
Turning our attention to the first of those questions, we must figure out what exactly does it mean for somebody to attempt something. This is where Yaffe’s theory of attempts comes in handy. The central edifice of this theory is something called the Guiding Commitment View, according to which an “attempt” can be defined as follows:

Guiding Commitment View of Attempts: An agent (A) can be said to attempt a crime (C) when the following two conditions are met: 
  • (i) A has an intention that commits him/her to every element of C; and 
  • (ii) A’s actions are guided by this intention.

The two conditions contain two important concepts — intention-based commitment and guidance by intention — which must be explained in more detail. Let’s go through them both in turn.

As Yaffe notes, the notion of intention-based commitment is broader than the notion of intending. This is because one’s intentions can commit one to a variety of actions or sequences of events that one does not directly intend. To put it another way, one’s intention-based commitments can be pluralistic: they do not have to (though they might) map directly onto a specific set of actions. For instance, suppose one intends to burn down one’s house in order to obtain an insurance payment. This intention commits one to causing the house to burn down, but that commitment can be satisfied in a number of different ways, e.g. by leaving the oven on, by neglecting to put out a cigarette, by drenching the house in petrol and throwing a match onto it, and so on. Each of these actions may be individually be intended after the original commitment.

So how is intention-based commitment to be understood in the criminal law? Yaffe suggests that, in order to figure out exactly what somebody’s intentions commit them to, we should ask ourselves a simple question: what kinds of things would be true if that intention were to play an ordinary role in motivating their actions? And, more specifically, do these things include the elements of the offence under consideration?

Let’s look now at the notion of guidance by intention. What exactly does this mean? Broadly speaking, Yaffe argues that to be guided by an intention is to act in manner that is motivated by whatever the intention demands. But applying this in practice proves tricky because action is complex and what is motivated by the intention under consideration must be distinguished from what is motivated or caused by something else. For example, suppose my intention is to ascend the stairs. I turn to you, declare “I am about to ascend the stairs!”, and then proceed to climb up the steps. In this case, I performed two actions: (i) a speech act (the declaration); and (ii) a normal act (climbing the stairs). The second of those acts was clearly motivated by the original intention; but the first was not, it seems to have some other causal history.

So how then is guidance by intention to be understood in the criminal law? Again, Yaffe proposes a simple enough test. We should ask ourselves: Did the agent initiate a causal sequence, motivated by the relevant intention, that would, in the ordinary course of events, have led to the completion of the crime? If yes, and if his original intention committed him to each element of that crime, he is liable for the attempt.


3. Did Herzog Attempt Murder?
With Yaffe’s theory in place, we can proceed to determine whether or not Herzog attempted murder. Obviously, to determine that we need to see if the two key conditions of the Guiding Commitment View have been satisfied.

As regards the first condition, the question is this: did Herzog have an intention that committed him to each element of the offence? In answering this, the insights of the omniscient author — in this case, Bellow — is of great assistance. Bellow effectively informs that Herzog had intended to cause the deaths of his wife and lover by his own actions. So it would seem he was indeed committed to each element of the offence of murder. Of course, it would be remiss not to point out that, in practice, things are not quite so straightforward. We are not omniscient authors, free to plant whatever intentions we choose in the minds of our characters. We will have to infer someone’s intentions from their outward acts (since they probably won’t tell us what they were thinking) and figuring out exactly what someone was committed to based on their outward acts is difficult. To say the least.

As regards the second condition, the question is this: did Herzog initiate a sequence that would, in the ordinary course of events, culminate in a murder? Look back to the diagram that I presented earlier on. The way I have drawn it, Herzog does indeed appear to have initiated a sequence that would have ended in the murder of his wife and her lover, but for his change of mind. The crucial question is whether this change of mind makes any difference to our analysis.

Yaffe argues that it should not. Indeed, he argues that no causal intervention into a sequence that was originally guided by an intention that committed the agent to all the elements of a crime should sway us from thinking that the crime was attempted. (At least, that’s how I read him in this article, I haven’t read his book so he may well allow for some exceptions). His reasoning takes the form of a simple reductio argument. Since every attempt involves some kind of intervention into the causal sequence — that being what distinguishes an attempt from a completion — accepting the principle that interventions nullify an attempt would mean accepting that there are no such things as attempted crimes. But clearly there are such things as attempted crimes. So the suggested principle must be wrong. To lay this out more formally:


  • (1) If it is true that interventions into an appropriately-guided causal sequence nullify an attempt, then there are no such things as attempts.
  • (2) There are such things as attempts. 
  • (3) Therefore, interventions into an appropriately-guided causal sequence do not nullify an attempt.


In this argument “appropriately-guided causal sequence” means “a causal sequence that was guided by an intention that was committed to all the elements of a crime”.

If Yaffe’s argument is correct, then it follows that Herzog did indeed attempt the murder of his wife and her lover. The fact that he changed his mind when he peered in the window does not help him, just as the fact (if it was a fact) that he was arrested by the police while staring in the window would not help him. The suggestion would thus seem to be that only a change of mind that occurred before the appropriately-motivated causal sequence was initiated would have nullified the attempt.


4. Does Herzog deserve a lesser sentence?
Yaffe’s analysis may seem harsh. To draw a parallel between the case in which Herzog prevents himself from completing the crime through a change of mind (an internal intervention) and a case in which he is prevented from completing the crime by an arresting police officer (an external intervention) will strike some as being implausible. Yaffe tries to remedy this by arguing that although Herzog has indeed attempted murder, his change of heart does not count for nothing. It will warrant some mitigation of punishment. To see this, we will have to briefly consider the possible grounds for punishment. Roughly, there are two such grounds: (i) retributive ones and (ii) consequentialist ones.

On retributive grounds, we punish people in direct proportion to their actual culpability. In cases of attempts involving external interventions, this might warrant punishing the attemptor just as severely, or marginally less severely than, the completor. This is because a person’s culpability corresponds to their state of mind and in such cases their state of mind is (typically) no different than that of the person who completes the crime. But this is not true in cases of internal intervention, assuming the internal intervention is intentional and not the result of some nervous tic or the like. In those cases, the person’s culpability has altered so they deserve a much less severe punishment, if any at all.

Similar reasoning holds if one punishes on consequentialist grounds. Consequentialists punish people either so as to deter them and others from committing similar crimes in the future. But what level of punishment will be sufficient to deter? Yaffe suggests that one way to think about it is to ask: what level of punishment would have provided sufficient reason to the person to prevent them from completing the crime? If they would have been prevented by a one-year sentence, then that would be sufficient punishment; if they would have been prevented by a two-year sentence, than that would be sufficient; and so on.

The problem comes when applying this test to those who have changed their minds. They have already provided themselves with sufficient reason to refrain from completing the crime. So would any punishment be acceptable? One may still say “yes” on the grounds that we have to think about deterring others, not just the individual in question. But still, it seems like the change of mind should serve to mitigate the burden somewhat. After all, we certainly don’t want to provide people with perverse incentives, ones that would give them no reason to change their minds. We might do this if we punish people severely irrespective of whether they changed their minds or not. If their punishment will be the same as if they completed the crime, they might be encouraged to go through with it anyway.

That brings us to the end of this post.

To summarise, Yaffe’s theory of attempts is called the Guiding Commitment View. According to this view, an agent can be said to attempt a crime if they: (i) have an intention that commits them to every element of that crime; and (ii) their actions are guided by that intention. Applying this to cases involving changes of mind, we find that those who change their minds after initiating a sequence of actions that would typically culminate in the completed crime are still deemed to have attempted those crimes. Nevertheless, the change of mind may count for something when we consider the appropriate level of punishment to impose on such people.

Wednesday, April 25, 2012

Theism and the Meaning of Life (Part One)



Since I recently re-opened the William Lane Craig can of worms, I’ve suddenly found myself with the itch to write about some other arguments he likes to make. Today, I’ll take a look at (aspects of) his comments on the meaning of life. I’ll be using the following article as my guide:

Gianluca Di Muzio “Theism and the Meaning of Life” (2006) Ars Disputandi, Volume 6 
 
This article starts by critiquing Craig’s conception of meaning and then offers an alternative conception that does not depend on the truth of theism.

While I enjoyed Di Muzio’s article, I was frustrated by the lack of formality in it (this might just be an annoying idiosyncrasy of mine, for which I apologise). In particular, I was frustrated by the failure to formalise Craig’s argument and then to map out each step in the critique of that argument. Admittedly, this is not an easy thing to do. This is for two reasons: (i) in contrast to his other work, Craig doesn’t go to the bother of formalising his own argument on the meaning of life; and (ii) his comments on meaning are quite rhetorical and enthymematic.

Despite these difficulties, I want to try to offer a decent formal analysis and critique of Craig’s argument. So although I’ll certainly be using Di Muzio’s arguments as my guide — indeed, nothing I say will be original to me — I’ll try to increase the value of those arguments through my formal reconstruction. I shall divide this reconstruction project into two parts. In this part, I will outline Craig’s argument and Di Muzio’s basic critique of it. In the second part, I will discuss Di Muzio’s alternative analysis of meaning in life.

I should add that my formal reconstruction is very much a first attempt. If people think I misconstrue the argument, or if people think that there might be a better way to formalise the argument, I'd be very happy to hear from them.


1. The Meaning of Meaning
Before getting into the details of Craig’s argument, some comments on the terminology used in this debate might be in order. The debate is typically couched in terms of “meaning” in life. But in some ways, the use of the word ‘meaning’ is unfortunate because it can lead to the conflation of “meaning in life” with semantic meaning. This conflation may not be entirely inappropriate, if there is some overlap between the two concepts, but we need to be aware of it nonetheless.

Because of this problem, alternative terminology is sometimes floated. For instance, Craig, in Reasonable Faith uses the terms value and purpose somewhat interchangeably with meaning (although he does distinguish them as well) when presenting his case against atheism. “Significance” is also a term that is occasionally used instead of “meaning”. These other terms are arguably more apt, although there is a danger here as well since, for example, references to “a valuable life” might be conflated with “a life with moral value”, which is distinct (although, again, the presence of moral value may play some part in a valuable life).

I think these terminological difficulties can be overcome. It seems clear to me that all discussions of meaning in life are primarily concerned with the question: is life worth living? In other words, is there some property (or properties) that makes existence in this universe worthwhile? The quest for meaning can be understood as the quest for this property (or properties).

In addition to this, the debate over meaning in life can be seen to be concerned with locating the necessary and sufficient conditions for worthwhileness. Indeed, the necessary/sufficient distinction can be quite useful when it comes to understanding the debate between theists and atheists on this matter. Roughly speaking, there are two dialectical jousts taking place:

The Necessary Condition Debate: Theists, like Craig, are arguing that the existence of God is a necessary condition for a meaningful life. Whereas the atheists are arguing that God’s existence is not a necessary condition for meaning.  
The Sufficient Condition Debate: Theists, like Craig, are arguing that God provides sufficient conditions for a meaningful life. Whereas atheists are arguing either: (a) that God does not provide sufficient conditions for meaning because nothing can provide such conditions (nihilism; or (b) God does not provide sufficient conditions for meaning and something else does provide such conditions.

Using this framework, we can think of the overall structure of this particular atheist-theist debate somewhat resemblant of a decision tree. At the first node, we have to decide whether God is a necessary condition for meaning. At the second set of nodes we have to decide whether (a) God provides sufficient conditions for meaning; or (b) whether atheism does. The four endpoints of this decision tree represent the four possible views one can hold on the matter, although, as it happens, I’m not aware of anyone believing that theism provides necessary but not sufficient conditions for meaning. That said, it would be open to the non-theist to argue that this is true and hence that theism, even if true, still cannot provide an adequate account of meaning in life.



As useful as this framework may be in allowing us to understand the general structure of the debate, it is, as we shall see, an oversimplification. This is because the role of immortality in meaningfulness often becomes part of the debate. As a result, two distinct decision trees confront the participants, one concerned with whether God is a necessary and sufficient condition for meaning and another concerned with whether immortality is a necessary and sufficient condition for meaning.


2. Craig’s Argument about Meaning in Life
As I said in the introduction, nowhere in his writings does Craig present a formal argument defending the conclusion that God is both necessary and sufficient for meaning. So whatever argument he is making must be distilled from what he says. This is to be charitable, of course, but charity is an important virtue in these matters.

In trying to distill Craig’s argument, Di Muzio focuses on two paragraphs from Craig’s writings. These paragraphs are quite long, so I won’t quote them in full here. I’ll just try to extract the most relevant parts from them. The relevant extracts from the first paragraph are:

“If each individual person passes out of existence when he dies, then what ultimate meaning can be given to his life?…His life may be important relative to certain other events, but what is the ultimate significance of those events? If all the events are meaningless, then what can be the ultimate meaning of influencing any of them? Ultimately, it makes no difference.” 
(From Craig, ‘The Absurdity of Life without God’)

This paragraph contains two claims. First, that the events in one’s life form part of chain of significance or meaning: the first event has meaning relative to the second event and so on. Second, that if this chain of events comes to an end, it lacks meaning irrespective of the relative meaning within the chain. In other words, the paragraph is arguing that immortality is a necessary condition for meaning. Without immortality the chain of events will come to an end.

The second paragraph centres on a thought experiment (based, apparently, on a science fiction story) involving an astronaut in a godless universe. The astronaut is faced with a choice between taking a potion that will allow him to live forever or taking a potion that will cause him to die. The astronaut chooses the second potion, hoping he will die:

But then to his horror he discovered he had swallowed the wrong vial— he had drunk the potion for immortality. And that meant that he was cursed to exist forever—a meaningless, unending life. Now if God does not exist, our lives are just like that. They could go on and on and still be utterly without meaning…So if it is not just immortality man needs if life is to be ultimately significant; he needs God and immortality. And if God does not exist, then he has neither. 
(Craig, ibid)

This paragraph also contains two claims. First, that although immortality is necessary for meaning, it is not sufficient for it. This is shown by the story of the astronaut. Second, that God is the added ingredient that turns an immortal and meaningless life into a meaningful one (Craig adds that, in any event, one couldn’t actually be immortal without God but that’s a slightly separate point). So, in other words, the paragraph ends with Craig’s major claim: that immortality plus God is sufficient for meaning.

One thing that is absent from this second paragraph is an actual reason for thinking that God would provide the special ingredient needed for meaning. Di Muzio thinks it’s pretty clear what this is though: God would have some divine plan for the cosmos and one’s life and one’s actions would have some role to play in executing that plan (for good or for ill). Craig seems to support this view in the other things he says about purpose in a godless universe, which I won’t quote from here.

In sum then, Di Muzio thinks that, for Craig, one’s life is meaningful if it has some permanent significance (role to play) in a scheme or plan that transcends (is higher than) one’s existence. To put it more pithily: one’s life has meaning if it matters in a higher scheme. Assuming this to be the principle guiding Craig’s ruminations, I offer the following two arguments as being representative of what Craig is saying.

The first is what we may call the “pro-theism” argument:


  • (1) One’s life is meaningful if and only if it has some permanent role to play in a scheme that transcends (or is higher than) oneself. 
  • (2) If Christian theism is true, then God has a plan for all of creation, namely: to bring about an eternal kingdom.
  • (3) Individual humans will have a permanent role to play in God's plan.
  • (4) God’s plan transcends (is higher than) the lives of individual humans.
  • (5) Therefore, if God exists, one’s life has meaning.


The pro-theism argument has a natural bedfellow, the anti-atheistic naturalism argument:


  • (1) One’s life is meaningful if and only if it has some permanent role to play in a scheme that transcends (or is higher than) oneself. 
  • (6) If atheistic-naturalism is true, then there is no plan or scheme that transcends an individual’s life. 
  • (7) If atheistic-naturalism is true, then all life will come to an end, i.e. no one has a permanent role to play in the cosmic drama. 
  • (8) Therefore, if atheistic-naturalism is true, one’s life has no meaning.


The two arguments are illustrated below and, together, they capture Craig's basic position on meaning in life (as it happens, Craig spends far more time on the anti-atheistic argument than he does on the pro-theism one).



Now, there are a lot of premises flying around here, which means there are many fruitful avenues of investigation and critique that could be explored. However, that would take up far too time and distract us from the main points of interest. So, for sake of argument, I’ll simply assume that everything in the anti-atheistic naturalism argument, with the important exception of premise (1), is correct even though that may not be the case. Thus, the remainder of this post will be taken up with the evaluation of the pro-theism argument.


3. A Critique of Craig’s Argument
When it comes to critiquing Craig’s argument, there is one obvious point of attack: premise (1). While the other premises could potentially be challenged (e.g. one could challenge (2) by arguing that no coherent plan has been revealed or explained to us, with some careful exegesis of biblical and theological texts), premise (1) is what really motivates the entire argument. It, after all, is what states the necessary and sufficient conditions for meaning. And since we already know that immortality is insufficient in itself for meaning, and since Craig has already conceded as much, we dedicate most of our attention to the role that the higher plan plays in the provision of meaning.

In focusing on the higher plan, we need to ask two separate questions: (i) is a higher plan, in itself, sufficient for meaning? and (ii) is a higher plan, in combination with immortality sufficient for meaning? We ask the first question because it’s possible that immortality is a red herring in this argument. And we ask the second question because, even if a higher plan is not in itself sufficient for meaning, there might be some synergistic effect such that, when combined with immortality, we have the sufficient conditions for meaning. If it turns out that the answer to both our questions is no, it will imply that Craig’s account of meaning is, at worst, flawed, or, at best, incomplete. In that event, we shall have to search for some other account.

So let’s look at the first question: is a higher plan, in itself, sufficient for meaning? The answer would clearly seem to be “no”. It’s quite possible for our lives to be inscribed within a higher plan (i.e. one that transcends our own lives) and for them to still lack meaning. Consider, if one were a Jew living in Nazi-occupied Europe during WWII, then it was true that someone had a higher plan for you, but in no sense could your life be said to be meaningful because of it. Admittedly, the higher plan in this instance is one that ends in your extermination but we’re leaving immortality to the side for now so it seems legitimate to include the Final Solution as an example of a higher plan. Nevertheless, there are examples that avoid this possibility. For instance, if one lives in a totalitarian state that has unscrupulous or oppressive aims (perhaps, North Korea) then one’s life may have a role in a higher plan, but once again that wouldn’t necessarily make it meaningful. Science fiction examples could be used to support this point. Say, for instance, the human race was bred as food or sport for some superpowerful race of aliens. Would that make the lives of individual humans meaningful?

If you are inclined to resist my interpretation of these examples, the problem may stem from confusing the sense of meaning that is being employed here. If one brings a semantic conception of meaning to bear on these examples, then it might be true to say that each of the lives in the examples has meaning because they “stand for something”. But, remember, that is not the sense of meaning that is at issue here. We are concerned with whether the lives in question are worth living. And when assessed from the standpoint of worthwhileness, these lives seem to acquire no meaning simply from the fact that they have a role to play in a higher plan. Indeed, we might be inclined to say that these lives acquire more meaning when they reject or escape the auspices of the higher plans in question. Thus, we are forced to conclude, that the mere presence or existence of a higher plan is insufficient, by itself, to provide meaning.


  • (9) The mere fact that one’s life has a role to play in a higher plan is, by itself, insufficient to make that life meaningful because one can imagine scenarios in which there is a higher plan but the higher plan does not make one’s life worthwhile.


This brings us to the second question: could the presence of a higher plan, in conjunction with immortality, be sufficient for meaning? The answer to this would also appear to be “no”. Suppose, for instance, that there exists an evil or wicked God who plans to torture each and every human being for eternity. Or suppose something like Calvinism is true but one is not among elect and is fated for perpetual purgatory or eternal damnation. In both of these cases there is a higher plan, and each individual human will play a permanent role in this plan, but in neither case could the individual lives be thought worthwhile.


  • (10) The existence of a higher plan, in conjunction with immortality, is insufficient for meaning since one can imagine scenarios in which both conditions are satisfied and life is still not worthwhile.




Taken together, (9) and (10) provide us with reason to reject premise (1). They show us that a higher plan is not sufficient for meaning either alone or together with immortality. However, they do not do two very important things. First, they do not show that the existence of a higher plan is not a necessary condition for meaning. And second, they do not show that specific higher plans are insufficient for meaning. So, a revised form of premise (1) might be possible, and this revision might provide some reason for accepting the theistic view over the atheistic one.

But let’s consider that second point for a moment. Could specific kinds of higher plan be sufficient for meaning? Sure, this is possible. But then the question arises: what kinds of specific plan? In the article, Di Muzio discusses this issue, but not quite in the same way as I do here. He suggests two ways in which higher plans could become meaningful:

(a) They could themselves part of some, still higher, plan; or 
(b) They could be intrinsically meaningful.

Di Muzio argues that (a) would lead to a infinite regress which would be unsatisfactory and would confer no advantages on the theistic position in this debate. Thus, he thinks (b) is the only live option. In order to determine whether this is advantageous to the theist, one might be inclined to assess different possible candidate theories of religious meaning (e.g. divine calling theory, redemption theory) and see whether they are intrinsically meaningful.

However, we shan’t do that here. Why not? Because the critique of premise (1) to this point, along with the suggestion that certain things can have intrinsic meaning, throws up an intriguing possibility: there could be something other than an eternal higher plan, which is compatible with the atheistic view, which provides sufficient conditions for meaning. If something like this exists, it would show that higher plans are not necessary for meaning. And if this is possible, the theistic conception of meaning in life would effectively be defeated.

We’ll consider this issue in part two.

Tuesday, April 24, 2012

Changing to Disqus

I've decided to change the commenting system on the blog to Disqus. I'm importing all comments from blogger to the new format but, apparently, this could take up to 24 hours to complete. So, for the time being, none of the old comments appear after the posts. Hopefully, that should change in due course.

Sunday, April 22, 2012

Morriston on God and the Ontological Foundation of Morality (Part 2)


(Part One)

This is the second part in a brief series on Wes Morriston’s article “God and the Ontological Foundation of Morality”. The article offers a sustained critique of William Lane Craig’s views on moral ontology. Throughout his work on moral ontology, Craig defends two key theses:

Thesis 1: If theism is true, we have a sound (ontological) foundation for morality. 
Thesis 2: If theism is false, we do not have a sound (ontological) foundation for morality.

In part one, we saw how Morriston challenged Craig’s defence of the first of those theses. In this post, we will take a look at Craig’s defence of the second thesis and Morriston’s response thereto.

Before getting into the nitty gritty, however, we must briefly revisit a complication from part one. If you recall, there are two kinds of moral facts for which Craig seeks a sound ontological foundation: (i) facts about moral value; and (ii) facts about moral duty. Officially, Craig argues that neither values nor duties have a sound foundation in a godless universe. But our focus here will be solely on his arguments in relation to moral value. This is for two reasons. First, Craig himself spends far more time on this issue than he does on duties. And second, the foundation of values seems to take primacy over that of duties. A duty is effectively an action with moral value that we are obliged to perform. Thus, to say we have a duty to do X is to presuppose that are actions and states of affairs with moral value.

So if Craig can show that moral value does not exist in a godless universe, he will effectively have shown that moral duties cannot exist either. Similarly, if Morriston can show that moral values do exist in a godless universe, he will at least have shown the possibility of a foundation for moral duties. As it happens, he does more than this. As we saw the last day, Morriston thinks there are two plausible foundations for moral duties in a godless universe. You might find that the arguments for those foundations seem more plausible after reading the discussion in this post.




1. Is Human Flourishing Morally Significant?
One popular ontological grounding for moral value, which is plausible in a godless universe, is human flourishing. Many philosophers, from Aristotle to Bentham, have argued that human happiness or flourishing, broadly construed, are sources of moral value. That is to say, actions and states of affairs that constitute or contribute to human flourishing are thought to be morally good, and actions and states of affairs that constitute or contribute to human unhappiness are thought to be morally bad.

Irrespective of whether the philosophers supporting this have believed in God or not, it seems that there is nothing essentially theistic about this conception of moral value. In other words, it seems like, even in a godless universe, there would be reason to think human flourishing was a source of value. Ask yourself the question: if it turned out that God did not exist, would you suddenly have reason not to promote human flourishing?

Interestingly, Craig’s defence of Thesis 2 (above) is almost entirely dedicated to answering this question in the affirmative. According to Craig, it is in fact true that in a godless universe, human flourishing is devoid of moral significance. Craig devises a number of arguments to support this, many of them more rhetorical than formal. However, I will attempt a formal reconstruction of some of them here.

I start with the following, which might be termed the “nothing special”-argument.


  • (1) If there is nothing special about humans, then human flourishing is not morally significant (and vice versa). 
  • (2) There is nothing special about humans in an atheistic universe. 
  • (3) There is something special about human in a theistic universe. 
  • (4) Therefore, human flourishing is morally significant in a theistic universe, but not in an atheistic one. 


I’ll talk about how Craig defends each of the key premises here in a moment. First, I want to make a general comment. As I just said, Craig is more rhetorical than formal when presenting his case for thesis 2. Nowhere is this more clear than in his use of the “nothing special” vocabulary. Craig never defines what he means by “special” and this is unfortunate since the phrase is ambiguous. Does he mean special in some objective sense? If so what could this be? Does he mean special in a subjective, observer-relative sense? Is so, whose perspective is relevant? None of these questions are answered directly and, as a result, the subsequent argument suffers. This is something that Shelly Kagan used to his advantage in his debate against Craig.

Now, that’s not to say we have no idea what Craig means by “special”; it’s just to say that whatever notion of “special” Craig is working with must be inferred from what he says. As it happens, in defending premise (2), Craig gives us a window into the workings of his mind:

“[O]n the atheistic view, there’s nothing special about human beings. They’re just accidental byproducts of nature that have evolved relatively recently on an infinitesimal speck of dust called the planet Earth, lost somewhere in a hostile and mindless universe and doomed to perish individually and collectively in a relatively short time.” (Craig and Kurtz, 2009)

As Morriston notes, this seems to boil down to:


  • (5) The following facts are true: (i) humans are tiny compared to the universe; (ii) humans haven’t been around for very long (given the age of the universe); (iii) humans owe their existence to a mindless natural process; (iv) humans die after a short time; and (v) eventually, all humans will be permanently dead. 


Therefore,


  • (2) There is nothing special about humans in an atheistic universe. 


As is clear, there is a gap that needs to be filled in the logic here. The filler will consist of some explicit principle or set of principles that allows us to say that shortness of existence, temporal and spatial “smallness”, natural origins and so forth imply a lack of “specialness”. But as Morriston notes, there is no reason to endorse any such principles. Surely, what makes humans special (or not) is what they actually are, not where they came from and how long they will be around? In this vein, he suggests that the capacity to feel pain, to love, to experience joy, to solve mathematical theorems, to write poetry and to compose symphonies are things that should make us think humans are special. And each of these is a fact about what humans actually are, not a fact about their origins and their eventual deaths. I’m not sure all of Morriston’s examples are useful — rhetorically effective though they may be — but I do agree that there is at least one important fact here, namely: the capacity for sentience (feeling pain and experiencing joy). Many strong moral principles (e.g. those covering the wrongness of killing), with robust intuitive support, make use of this fact, suggesting that it does indeed make humans morally valuable. (Note: I should add that, although ultimately endorsing a kind of nihilism, Nagel’s famous article on the absurd also makes the point that Craig’s “facts” cannot form the basis for a persuasive argument against the value of human flourishing).



2. Isn’t it all just matter re-arranging itself?
Most of Craig’s other arguments in favour of thesis 2 are simply variations on the “nothing special” theme. Still, some are worthy of independent consideration. First up is the argument based on the following, rather remarkable, passage:

“[I]f man has no immaterial aspect to his being, whether you call it a soul or a mind or whatever, then we’re not qualitatively different from other animal species. On a materialistic anthropology there’s no reason to think that human beings are objectively more valuable than rats. When a terrorist bomb rips through a market in Baghdad, all that really happens is a rearrangement of the molecules that used to be a little girl.” (Craig and Antony, 2008)

One could focus on the comparison Craig makes with animals in this quote, but that would be a mistake since it’s clear that the moral implications of materialism are really what’s at issue. If we attempt to formalise the argument that Craig is making, we notice something interesting — I think — about its rhetorical structure.


  • (6) If materialism is true, humans are just mere arrangements of molecules. 
  • (7) No mere arrangement of molecules can have moral value. 
  • (8) Therefore, if materialism is true, humans have no moral value. 
  • (9) On materialism, to blow up a little girl is simply to rearrange the molecules of which she consists.
  • (10) Therefore, if materialism is true, blowing up a little girl is a morally neutral act. 


I say this is rhetorically interesting because, once Craig has reached (8) he has proven all he needs to to show that thesis 2 is true (since he is assuming most atheists are materialists). But he still adds the argument about the little girl. Why? I conjecture this is largely for rhetorical effect, to underline the absurdity of the view he is critiquing. This is all well and good, given that this is something Craig originally said in a debate, but it’s worth being on the lookout for rhetorical ploys like this and recognising them for what they are.

In any event, rhetorical ploys to the side, is there anything to be said for this argument? Morriston thinks not. In particular, he notes that premise (7) is easily rebutted: Craig assumes that, on materialism, there is no difference between a collection of molecules forming a rock — or a rat — and a human being. But this is to massively mischaracterise the materialist view. No materialist thinks that all arrangements of molecules are the same. If the arrangement is organised in such a way that it has first person experience, is capable of planning and dreaming, of loving and laughing, then it has a very different moral character from an arrangement of molecules that has none of these things.

Furthermore, there is no reason to think that immaterialism gets a free pass when it comes to moral value. Do immaterial substances automatically come imbued with moral value? Some immaterial concepts, e.g. mathematical entities and truths, seem pretty value neutral. So why think immaterial mental parts do?

If the argument that Craig is making is that only entities with minds have moral value, then that’s all well and good (I might even agree) but that argument should be construed as being strictly neutral as between materialism and immaterialism. Materialists do not deny the existence of minds; they just have a different theory as to their origin and constitution.



3. Evolution and Moral Knowledge
There’s one final argument in Craig’s arsenal that Morriston deems worthy of consideration. This one is based on the notion that the fact of evolution somehow debunks morality. Now, as it happens, I did a fairly extensive series of posts on evolutionary debunking arguments last year in which I outlined exactly how it is that evolution might debunk morality. Craig doesn’t engage with the subtleties of that debate. This is unfortunate because it leads Craig to overestimate the power of the debunking argument and to contradict one of his own key commitments in the process. I won’t review all the details of debunking arguments here myself. Instead, I’ll just sketch some of the important aspects. Hopefully, this will be enough for people to appreciate where Craig goes wrong.

Craig’s view is that if our moral conscience is explained by evolution — in other words, if our morality is just an aid to survival and reproduction — then there is a sense in which morality doesn’t exist. In adopting this view, Craig relies on work done by the inimitable philosopher of science, Michael Ruse. Morriston had some difficulty in working out what Ruse’s actual argument was. It seems — and I’m relying entirely on Morriston’s discussion when I reconstruct this — that the argument reduced to the following:


  • (11) Evolution selects for some degree of cooperative behaviour.
  • (12) Therefore, evolution can explain morality (where “morality” = “judgments and behaviour about what is right and wrong”).
  • (13) If evolution explains moral judgment and behaviour, then the actual existence of an objective morality would make no difference to human thought and behaviour.
  • (14) Therefore, objective morality does not exist. 


As Morriston points out, there are a number of puzzling gaps in the reasoning here. First of all, the leap from (11) to (12) seems unwarranted. Morality is far more complex than mere cooperation and so even if evolution does explain cooperation we’re a long way from being able to show that all moral behaviour and judgment is, in fact, explained by evolution. Second, the leap from (12) and (13) to (14) is highly dubious. Ruse justifies premise (13) in the following manner:

“Consider two separate worlds, identical except that one has objective morality and the other does not. Humans could have evolved in both worlds to believe exactly the same thing…In short, therefore, there is a sense in which objective morality is redundant. Its existence is irrelevant to human thought and action.” 
(Ruse, "Evolutionary Ethics and Christian Morality: Are they in Harmony? 1994, Zygon)

Morriston thinks this is flawed since a moral realist is apt to argue that moral properties supervene on natural properties in such a way that a world devoid of objective value, yet in which our moral beliefs are the same as they would be in a world with objective value, is impossible. But even if that’s not right, there’s a bigger problem: there’s no way that you can get from (13) and (12) to (14) without adding an additional premise. Morriston suggests that the best candidate for this would be:


  • (15) If the truth of a proposition contributes nothing to the best causal explanation of our belief in its truth, then that proposition is not true. 


But this is a deeply flawed principle. It commits the genetic fallacy in assuming that if the origins of our belief in a proposition are suspicious then so too is the truth of that proposition. This is not good reasoning and so Ruse’s argument can’t help Craig to defend thesis 2.

This, incidentally, is a mistake that all careful proponents of the evolutionary debunking according are at pains to avoid. None of them will say that the fact (if it is a fact) that evolutionary processes explain our moral beliefs is sufficient reason for us to doubt the existence of objective morality. Instead, they will say that the fact that evolutionary processes explains our moral beliefs undercuts the warrant that we have for our belief in objective moral value. This is a subtly different thing and, it just so happens, is something Craig should be keen to deny.

Craig fans (and I use the term “fan” somewhat unorthodoxly here since I count myself as one) will know that all his claims about morality are made as part of his effort to offer a moral argument for the existence of God. The argument looks like this (it usually includes references to duties but I’ve excluded those here):


  • (16) If God does not exist, then objective moral values do not exist.
  • (17) Objective moral values exist.
  • (18) Therefore, God exists. 


This argument is supposed to provide reason for believing in the existence of God. To do that successfully it cannot simply presuppose the existence of God. That would be to beg the question. Craig is aware of this, that’s why, in defending premise (17), he is keen to point out that we all — atheist and theist alike — have knowledge of the existence of objective moral values. Which is to say, he thinks that everyone has warranted beliefs in the existence of objective moral values. Hence, Craig shouldn’t want to do anything that would undercut our warrant for believing in the existence of moral values. Hence, he should not use the warrant-form of the debunking argument.


4. Conclusion
To sum up, one of Craig’s two theses about morality claims that moral values cannot exist in a godless universe. In this post, we considered three arguments he uses to defend this thesis, as well as the responses thereto from Wes Morriston.

First, there was Craig’s argument that there’s nothing special about humans in a godless universe. Craig argued that this was due to their temporal and spatial smallness. Morriston rejected this on the grounds that the specialness of humans is grounded in facts about who they are, i.e in facts about their cognitive and mental faculties and so forth, not in facts about their natural origins and their ultimate fate.

Second, there was Craig’s claim that, on materialism, all arrangements of matter are of equal (i.e. “zero”) moral value. As Morriston noted, this argument is flawed for a variety of reasons. Not least of which is its caricaturing of the materialist position. Materialists will typically point out that there are many important differences between different arrangements of matter, some of which may be of great moral significance.

Third, there was Craig’s attempt to use an evolutionary debunking argument. As was pointed out above, if Craig is to use this argument to support his thesis, he will overreach and commit the genetic fallacy. On the other hand, if he is to use the argument appropriately, he will undermine a key premise of his moral argument. So he is caught in a bit of bind.

This brings us to the end of my exposition of Morriston’s article.

Tuesday, April 17, 2012

Morriston on God and the Ontological Foundation of Morality (Part One)



Sometime back, somebody — alas! I cannot remember when or who — asked me to write an analysis of the following article by Wes Morriston:

God and the Ontological Foundation of Morality” (2012) 48 Religious Studies 15-34

At the time I rejected the suggestion. I had read Morriston’s article when it was first published, and while I found it to be pretty good, I didn’t think it contained any revelatory insights into the putative relationship between theism and morality. More recently, however, I have had the occasion to re-read the article, and although my original interpretation hasn’t changed greatly, Morriston’s discussion has become tangentially relevant to something else I’m researching and writing at the moment and so I am much more motivated to engage with its details. Which is precisely what I’m going to do in this and the next blog post.

Morriston’s article is a sustained analysis and critique of William Lane Craig’s views on the relationship between God and morality. According to Morriston, throughout his work on this topic, Craig has defended two clear theses about this relationship:

Thesis 1: If theism is true, we have a sound (ontological) foundation for morality.

Thesis 2: If theism is false, we do not have a sound (ontological) foundation for morality.

The problem, as Morriston sees it, is that both of these theses are false. It is not true that God provides a sound foundation for morality, nor is it true that without God we would lack a sound foundation. I will discuss Morriston’s analysis of each of these claims over the next two blog posts.

In this post, I will cover Morriston’s discussion of the first thesis. I start by outlining, in a fairly quick-fire manner, Craig’s basic views on the ontological grounding of moral values and moral duties. I then look at Morriston’s critique of Craig’s views on moral values, before closing with Morriston’s critique of Craig’s views on moral duties.


1. Craig’s Moral Ontology, in a Nutshell
The universe consists of entities, activities and states of affairs. Every actually existent entity, activity and state of affairs can be called a “fact”. One of the curious features of these facts is that they sometimes have moral statuses attached to them. There are basically two types of moral status that can attach to a fact: (i) a value status, according to which the fact is held to be “good” or “bad”; and (ii) a deontic or duty-oriented status, according to which the fact is held to be obligatory, permissible or forbidden.

The central assumption guiding Craig’s moral theory is that the moral statuses we attach to facts need an ontological grounding. Without this ontological grounding moral facts would fail to be objective in the sense required by Craig (I discussed Craig’s views on objectivity at some length before). Whether this assumption is a good one is a question to which we shall return. For now, let us focus on the ontological grounding that Craig offers.

Let’s look first to Craig’s ontological grounding of moral values. According to Craig, moral values are grounded in the being of God. They aren’t simply inert abstract facts, as the moral Platonists would have us believe and which Craig thinks metaphysically “queer”, they are facts about a concrete God. Specifically, they are facts about God’s essential nature or characteristics. God, according to Craig, is essentially just, kind, generous, loving and so forth. These characteristics ultimately constitute the moral good; and the fact that they are characteristics of God thus provides the good with its ontological grounding.

Craig’s views about the ontological grounding of moral duties builds upon what he says about moral values. As regards moral duties, Craig adopts a fairly typical divine command theory (DCT): the obligatory, the permissible and the forbidden are grounded in God’s commands. Something is obligatory if God commands us to do it; something is forbidden if God commands us not to do it; and something is permissible if God neither commands nor forbids it.

Craig adds to this typical DCT the additional claim that what God commands “flows from” his moral nature. In doing so, he hopes to avoid the typical Euthphro-based objection to DCT that on this account God’s commands could be arbitrary, i.e. that God could command us to do anything, including raping and torturing children, and that if he did we would be obliged to do it. This potential arbitrariness of God’s commands seems to clash with some deep intuitive beliefs we have about moral facts, namely, that they are necessary, i.e. the same in all possible worlds. But if God’s commands flow directly from his moral nature, the arbitrariness objection is, arguably, overcome. God’s commands are no longer arbitrary, they are constrained by his morally perfect nature.

The diagram below tries to summarise Craig’s views about the ontological grounding of moral facts.




2. Morriston on Moral Values
Morriston offers two main critiques of Craig’s attempt to ground moral values in God’s essential characteristics. The critiques essentially amount to the same thing, but they differ in terms of their rhetorical strength, at least they do so in my opinion. The first, and in my opinion rhetorically weaker critique, is effectively a reformulated version of the Euthyphro dilemma. The second, and in my opinion rhetorically stronger critique, probes a bit more deeply into Craig’s conception of the grounding relationship.




2.1 - The New Euthyphro
The reformulated version of the Euthyphro works like this. Craig is claiming that things are morally good to the extent that they resemble God’s essential characteristics. In other words, as the diagram above makes clear, he is claiming that goodness is equivalent to God-likeness. But this raises a dilemma. Either God-likeness is equivalent to goodness because his essential characteristics — like kindness, lovingness, generosity and justice — are independently good, or those characteristics are good solely in virtue of the fact that they are possessed by God.

If Craig takes the first horn of the dilemma — and says that characteristics such as kindness, lovingness and so on are good independently of God — then he effectively concedes that God is not needed to ground objective moral values, they are independent, free-floating properties. This would surely not be welcome to him. But if he takes the second horn of the dilemma — and says that those characteristics are good simply because they are possessed by God — then he faces two problems. First, he appears to make the good contingent upon God’s characteristics which, arguably, could have been different. Making moral values contingent in this manner clashes with some people’s deep intuitions about the nature of the good. And second, he seems to commit himself to the claim that, in a godless universe, someone who was loving and kind would not be good. Morriston thinks this is strongly counterintuitive: we have no real reason to think that those characteristics would not be good in a godless universe.

Craig may respond that these two criticisms of the second horn of the dilemma involve counterfactual claims with impossible antecedents (call these “counterpossibles”) and that counterpossibles of this sort have, on the standard account, merely trivial or vacuous truth values. That is to say, no serious objection to his view can rely on claims like “If God had different characteristics, then those characteristics would be good” or “If God did not exist, then kindness would not be good” since it is impossible for God to have different characteristics or for God to not exist. I’ll discuss this kind of response in more detail below when I consider Morriston’s critique of Craig’s account of moral duties.


2.2 The Moral Metre Stick
The second criticism of Craig’s ontology of moral value looks more closely at the kind of ontological grounding Craig envisages. Morriston looks at a number of quotes from Craig’s debates at this point in order to determine what Craig thinks about this relationship. I will simply summarise the results of this investigation. According to Morriston, Craig’s vision of the grounding here can be understood in terms of “informative identification” and/or “paradigmatic instantiation”.

The best way to understand what an informative identification is is to look at some examples. Two are cited in Morriston’s article. The first is that “water” can be informatively identified with the chemical compound “H20”; the second is that “heat” can be informatively identified with “mean molecular motion”. To me, these are two classic examples of reductive ontological groundings, i.e. cases in which one set of properties is reduced to another. In theory, this could lead to the elimination of the first set of properties. In other words, all talk of water could be eliminated in favour of talk about H2O, or all talk of heat could be eliminated in favour of talk about mean molecular motion.

Does Craig intend to make similar claims about talk of goodness and talk of God-likeness? It’s not clear to me that he does, and the possibility that he does raises issues not addressed by Morriston in his article so I will not pursue them here. All I will say is that there are significant problems associated with reductive accounts of the good of this sort (which, I hasten to add, is not to say that all reductive accounts of the good are problematic, just that some are). Mark Schroeder’s article “Realism and Reduction” is a good place to start if you wish to explore these issues in more detail.

The notion that God provides a paradigmatic instantiation of the good is rather more interesting and, I believe, significant. The analogy Craig uses to make his case is a familiar one: the metre stick analogy. According to popular lore, the length we refer to as “one metre” is determined by reference to a standard metre stick (an iridium bar that is kept in Paris). This is the paradigmatic instantiation of the metre. But this use of this physical paradigm has a curious effect on the ontology of the metre viz. the degree to which an distance is one metre in length just is the degree to which it is similar to the length of the standard metre stick. There is nothing more to the ontology of the metre than this similarity relationship.

Craig argues that this is also true of the ontology of moral goodness. That is to say, God functions as something like a standard moral metre stick: the degree to which any person, object, event or state of affairs in the real world is good just is the degree to which it resembles God.

I have attempted to summarises the main features of the metre stick analogy in the diagram below.



There are several problems with Craig’s use of the metre stick analogy. For starters, the story about the standard metre stick is a good deal more complex than is being let on. The precise definition of a metre, and the physical paradigm that is used to define it, has changed over the years. At the moment, a metre is defined as “the length of the path travelled by light in vacuum during a time interval of 1 ⁄ 299,792,458 of a second”. It is not defined by reference to the iridium bar that is kept in Paris. The fact that the paradigm has shifted in this manner over time suggests, perhaps, that the “metre” is not simply reducible to any physical paradigm but is, instead, an abstract property. It may occasionally be helpful to identify it with a physical paradigm, but this should not detract from the fact that it is really an abstract concept, something that is ultimately separate from any physical instantiation. Arguably, this is also true of the good: the good is an abstract property, not reducible to any particular physical paradigm.

This may not be persuasive to everyone so it’s worth notin a second problem with the metre stick analogy. This is that units of measurement are arbitrary in a way that moral values are not. In order to measure things and engage in physical sciences reliant on measurement, we need some conventional system of measurement. These conventional systems can arise from completely arbitrary starting points, e.g. the length of Henry VIII’s thumb, but this does not matter: as long as the units are matters of conventional agreement, the measurement game can begin. The same is not true of moral values. We cannot just pick an arbitrary paradigm, such as God’s characteristics, and say that being good consists in similarity with those characteristics. The good is a more important, and metaphysically robust concept than that.

This leads me to suggest an alternative analogy that might be more in keeping with what we think about the good: the straightedge analogy. The analogy is inspired by Dan Dennett’s discussion of the history of the straightedge in this talk. A straightedge is a tool that is used, unsurprisingly, to draw straight edges or to check the flatness of machined surface (among other things). Over the years, many physical objects have served as paradigms of the straight edge. But the straight edge itself remains an abstract, indeed, mathematical, entity. Although our physical paradigms might get closer and closer to this abstraction, they are still not ontologically equivalent to it. I argue that something similar is true of the good: approximations of it can be found in physical paradigms, but it is not reducible to those paradigms. So there may be many reasons for thinking that God, if he exists, would serve as an excellent paradigm of the good, but this should not be taken to imply that God is ontologically equivalent to the good.

(Note: the last three paragraphs are really based on my own ideas, not Morriston’s)



2.3 - So what does ground moral values?
If the preceding criticisms succeed, then Craig has failed to show that God provides a sound ontological foundation for moral values. This raises the obvious question: if God doesn’t provide it, then what does? Morriston’s preferred answer, which has been hinted at above, is “nothing”. Moral values are abstract, sui generis facts, just as a moral Platonist would hold. Craig might object that this makes moral values metaphysically “queer” but Morriston avers that Craig’s claim that moral values just are equivalent to God’s essential characteristics is no less metaphysically puzzling.

When it comes to finding ontological groundings for facts it’s always going to be a question of which is the most appropriate stopping point. Since not every fact can have a grounding (lest we risk an infinite regress of groundings) we have to decide where to call a halt. Morriston argues that Craig has done nothing to show that God is a more appropriate stopping point than the moral properties of kindness, generosity and justice themselves.



3. Morriston on Craig’s Account of Moral Duties
After that rather elaborate discussion of moral values, my discussion of duties will seem like somewhat of a damp squib. But that’s for two reasons. First, I think the analysis of values is more fundamental and more interesting. And second, Morriston has relatively less to say about this topic.

Recall that for Craig moral duties are grounded in God’s commands, but that God’s commands are constrained by his essential nature. As a result, Craig thinks he manages to slip between the horns of the classic Euthyphro dilemma: duties really do depend on God’s commands, but God’s commands are not arbitrary.

This means that Craig thinks he need not respond to the classic “What if God commanded something terrible?” objections to DCT. Why not? Because these objections depend on counterpossibles, i.e. commands that God could never issue. This could mean one of two things:

(a) The objection relies on a counterfactual claim with a vacuous truth value according to the usual Stalnaker-Lewis semantics of counterpossibles; or

(b) The objection relies on a nonsensical claim. Asking “what if God commanded something terrrible?” is like asking “what if circles had straight edges?”. Whatever answer we give to these questions would be meaningless.

Morriston rejects both of these suggestions. He thinks that counterpossibles can be both non-vacuously true and meaningful. The example he chooses is an unusual one. He asks us to consider what would follow if an omniscient and perfectly honest being told us — per impossibile — that 2+2=5. In response he suggests that we would have to take the claim seriously since it seems to be the case that whatever an omniscient and perfectly honest being tells us has to be true. In other words, this seems to be a counterpossible claim that is neither vacuously true nor non-sensical. Hence, it is still possible for “what if God commanded X?” objections to have some effect on Craig’s DCT.

I’m not entirely sure I follow Morriston’s reasoning here. Those who are better versed in the semantics and logic of counterfactuals will be better-positioned to evaluate his argument. I’m rather more persuaded by Morriston’s second observation about counterpossibles, which is that Craig himself seems to have an inconsistent attitude toward them. In particular, it seems that Craig thinks that there is at least one counterpossible that is neither vacuously true nor nonsensical:

Thesis 2: If theism is false, we do not have a sound (ontological) foundation for morality.

This is a counterpossible because Craig seems to adopt the view that God is a necessary being, i.e. a being that could not fail to exist. But despite this Craig expends a great deal of energy defending the truth of this counterpossible. We’ll consider what he to say in part two.

If Morriston is successful in his refutation of Craig’s ontology of moral duties, the obvious question arises: what does ground moral duties? Morriston argues that there are two possibilities.

The first is that duties, like values, need no ontological grounding. After all, there are a variety of normative laws — e.g. one ought not to believe any argument that commits the fallacy of affirming the consequent — that seem to create something akin to duties but have no grounding outside of themselves.

The second is that duties could be grounded in the commands or desires of a strictly hypothetical agent, as is done in Ideal Observer-style theories of ethics. In this case, the actual existence of the ideal observer is not needed since all that matters is what the observer would command/desire if he/she existed.


4. Conclusion
To sum up, Craig thinks that if theism is true, then we have a sound ontological foundation for moral facts. Specifically, he thinks that moral values are grounded in the essential characteristics of God and that moral duties are grounded in the commands of an essentially good God.

In response, Morriston has argued that Craig’s attempt to ground moral facts in God is flawed. His grounding of moral values relies on a questionable account of the grounding relationship and is open to a revised version of the Euthyphro dilemma. Likewise, his grounding of moral duties relies on an inconsistent attitude toward the truth of counterpossibles. In addition to this, Morriston has argued that moral facts either do not need any ontological grounding or, if they do, that non-theistic groundings are available.

It should be noted that Morriston does not think that there is anything atheistic about his view. Indeed, he self-identifies as a Christian and thinks that his view, which is shared by other Christians (e.g. Swinburne) should be preferred by theists over that of Craig.

Tuesday, April 3, 2012

Bostrom on Superintelligence and Orthogonality



I know it’s been a while since I sat down to write one of these. You’ll have to forgive me: teaching commitments intervened in the interim. But that’s all coming to a close now so I should have more time for blogging over the next few weeks. And I’m going to kick-start my blogging renaissance today by taking a look at a recent article by Nick Bostrom entitled “The Superintelligent Will”. The article deals with some of the existential problems that might arise from the creation of a superintelligent AI. This, of course, is the core research topic of the folks over at the Singularity Institute.

I haven’t personally been motivated to spend much time thinking about the singularity and its potential fallout in the past, but Bostrom’s article caught my eye for a few reasons. One of them was, admittedly, its relative brevity (a mere 16 pages!) but another was that it deals directly with the concept of rationality and the normative dimensions thereto. This is a topic I’ve been interested in for some time, particularly in light of my attempts to argue for a constructivist metaethics in my PhD thesis. A constructivist metaethics, at least in my view, is one that argues that normative (moral) facts a built directly out of the basic constituents of practical/instrumental rationality.

How exactly does this tie-in with the discussion in Bostrom’s article? In the article Bostrom’s argues that even if we constructed a superintelligent machine (by “intelligent” he means capable of engaging in means-end reasoning, i.e. capable of being instrumentally rational) to have a benign ultimate goal, it might still end up doing things that are not so benign. Why is this? Because in order to achieve its ultimate goal, the AI would stumble upon certain “good tricks” of means-end reasoning that are not benign.

To be more precise, Bostrom explicitly defends two key theses in his article with an additional thesis being implicit in his discussion. The three theses are as follows (the first two names are used by Bostrom, the third is my own interpolation into the text):

The Orthogonality Thesis: Leaving aside some minor constraints, it possible for any ultimate goal to be compatible with any level of intelligence. That is to say, intelligence and ultimate goals form orthogonal dimensions along which any possible agent (artificial or natural) may vary.

The Instrumental Convergence Thesis: Agents with different ultimate goals will pursue similar intermediate or sub-goals [because such intermediate goals are either: (a) necessary preconditions for achieving the ultimate goal; or, alternatively (b) “good tricks” for achieving the ultimate goal.]

The Unfriendliness Thesis: Because some intermediate goals are unfriendly to humans, and because of instrumental convergence, even artificial superintelligences with seemingly benign ultimate goals can do things that are unfriendly to human existence.

While everything Bostrom has to say on these issues is of some interest — though, if I’m honest, none of his arguments seem sufficiently detailed to me — I’m going to zone in on what he says about the Orthogonality Thesis. This is because it is this thesis which overlaps most with my previous studies.

I’ll break my discussion down several parts. First, I’ll present my own reconstruction of the basic argument for the orthogonality thesis. Second, I’ll consider how Bostrom defends this argument from what I’m calling the motivating-belief objection. Third, I’ll consider how he defends the argument from the normative rationality objection. And fourth, I’ll consider what Bostrom has to say some about the weak constraints on the orthogonality thesis. my own (minor) criticisms of what he has to say. I should stress at the outset that my criticisms are very much preliminary in nature. They’ll need to be worked out in more detail before they have any real bite.


1. The Basic Argument for the Orthogonality Thesis
Bostrom starts his discussion of orthogonality by attempting to wean us away from our tendency to anthropomorphise about hypothetical artificial intelligences, to think that they will have similar motivations and desires to our own. Quoting from Eliezer Yudkowsky, he is keen to suggest that an artificial intelligence need not have any recognisably human motivations or desires. Unlike, say, HAL in 2001: A Space Odyssey, whose motivations and tendencies seem all too human (or do they? Sometimes I wonder).

This discussion leads directly into Bostrom’s presentation of the orthogonality thesis. He suggests that all possible agents, both natural and artificial, will take up a location in a parameter space. Although this parameter space is, in all likelihood, multi-dimensional, Bostrom says we can think of it as consisting of two major dimensions: (i) the intelligence dimension, where intelligence is understood as the set of capacities that allow the agent to engage in means-end reasoning; and (ii) the goal dimension, which speaks for itself really.

Bostrom’s contention is that these two dimensions are orthogonal, and that this orthogonality has some important implications. Here we run into a terminological quagmire. The word “orthogonal” has slightly different meanings in different contexts. For instance in mathematics, dimensions are orthogonal when they are at 90-degree angles to one another. If that was all Bostrom was claiming about the dimensions in his parameter space of possible agents it would be relatively uninteresting. Fortunately, there are other uses of the term that make better sense of his thesis. Thus, in statistics and computer science, for instance, orthogonality describes a situation in which properties can vary independently of one another. This sense of the word orthogonal tracks well with what Bostrom wants to say about intelligence and ultimate goals.

But what about an argument for the orthogonality thesis? The basic argument would appear to be this:


  • (1) For any two parameters P1 and P2, if P1 is orthogonal to P2, then the value of P1 can vary without there necessarily being any variations in P2.
  • (2) Intelligence and ultimate goals are orthogonal parameters.
  • (3) Therefore, it is possible for any level of intelligence to coexist with any type of final goal.


This really is a basic argument. Premise (1) is purely stipulative in nature, so it can’t really be challenged in this context. All the argumentative weight thus rests on premise (2). Given this, it is interesting that Bostrom chooses not to offer any formal defence of it. Instead, he tries to defend it from counterarguments, presumably in the hope that if it can weather the storm of counterargument its plausibility will increase.

Let’s see what these counterarguments are.


2. The Motivating-Belief Objection
The simplest objection to Bostrom’s claim is a popular one among certain metaethicists. It is that with higher levels of intelligence come more accurate beliefs (or their functional analogues). Furthermore, certain beliefs — e.g moral beliefs — are motivationally salient. Thus, it may be the case that a superintelligent being would have to have certain goals, and would be forbidden from having others, because accurate moral beliefs demand that it have this motivational structure.

To put this in argumentative form:


  • (4) The more intelligent an agent is, the more accurate its beliefs (moral, scientific, prudential etc.) will be.
  • (5) If a being has accurate moral beliefs, it will be motivated to act in accordance with those beliefs.
  • (6) Therefore, the more intelligent an agent is, the more likely it is to act in accordance with accurate moral beliefs.
  • (7) Accurate moral beliefs will forbid an agent from having certain goals, allow it to have certain other goals, and oblige it have yet other goals.
  • (8) Therefore, the more intelligent an agent is, the more likely it is to be prevented from having certain final goals.


Much of this argument seems sensible to me. For instance, the claim that the accuracy of beliefs goes up with intelligence (4) seems to be part of the definition of intelligence to me. Interestingly though, Bostrom thinks this might be not be true and that intelligence may not mandate the acquisition of certain beliefs, but then one has to wonder what kind understanding of intelligence he is working with. He says “intelligence” is the capacity to engage in means-end reasoning, but it seems to me like that would have to include the capacity to obtain accurate beliefs about the world.

Likewise, the claim that moral beliefs will forbid certain goals just seems to draw upon the conceptual underpinning of morality. Admittedly, the claim is not watertight. It is possible that accurate moral beliefs entail that everything is permitted, but that seems unlikely to me (although I shall return to consider the significance of nihilism again towards the end of this post).

The obvious flaw in the argument comes with premise (5). The notion that beliefs are motivationally salient is hotly contested by defenders of the Humean theory of the mind. According to them, desires and beliefs are functionally separate parts of the mental furniture. Reason is the slave of the passions. There are some decent defences of the Humean theory in the literature, but I shan’t get into them here. Bostrom doesn’t do so either. He simply notes the possibility of using the Humean theory to defend his argument from counterattack, but essentially leaves it at that. I shall follow suit, but will insert the following condensed premise into the argument diagram I am compiling:


  • (9) Humean Defence: Beliefs are motivationally inert because the Humean theory of mind is true.


In addition to using the Humean theory to defend his argument, Bostrom appeals to two further defences. I’ll discuss them briefly here. First up is the following:


  • (10) The Overwhelming Desire Defence: Even if beliefs do motivate, there may be overwhelming desires that always cancel out the motivational power of beliefs.


It’s not clear exactly where this defence should be slotted into the overall dialectic. But I suspect that it is best construed as a rebuttal to premise (5), i.e. as a claim that accurate moral beliefs can be overwhelmed by other desires. Is this true? Bostrom merely claims that it is possible, but even that seems questionable to me. If it is true that moral beliefs have motivational salience, then I would imagine that their motivational salience would be very high. Again, this seems part of the essential character of moral beliefs to me: a moral belief would provide you with a decisive reason to do or refrain from doing a certain thing.

The other defence Bostrom uses is this:


  • (11) The No-Belief Defence: It is possible to construct an intelligent system such that it would have no functional analogues of beliefs or desires.


If this is true, it would provide a decent refutation of premise (4) and so would in turn defeat the motivating-belief objection to the orthogonality thesis. But is it true? All Bostrom says about it is the following:

A third way in which it might be possible for the orthogonality thesis to be true even if the Humean theory were false is if it is possible to build a cognitive system (or more neutrally an “optimization process”) with arbitrarily high intelligence but with constitution so alien as to contain no clear analogues to what in humans we call “beliefs” and “desires”. This would be the case if such a system could be constructed in a way that would it make it motivated to pursue any given final goal.(Emphasis added)

This looks like weak argumentation to me. As is revealed by the italicised final sentence, it amounts to little more than a restatement of the orthogonality thesis. It is not, nor should it be mistaken for, a true defence of it.




3. The Normative Rationality Objection
Another counterargument to the orthogonality thesis, also addressed by Bostrom, works off a normatively “thick” account of rationality. Roughly-speaking, a normatively thick account of rationality holds that all intelligent rational agents would have to recognise that certain desires, combinations of desires or, indeed, belief-desire pairings are irrational and so they would be unable to sustain them. If this is true, then the orthogonality thesis would not go through.

Perhaps the best-known proponent of this view in recent times is Derek Parfit whose Future-Tuseday-Indifference (FTI) example captures the idea quite effectively:

A certain hedonist cares greatly about the quality of his future experiences. With one exception, he cares equally about all the parts of his future. The exception is that he has Future-Tuesday-Indifference. Throughout every Tuesday he cares in the normal way about what is happening to him. But he never cares about the possible pains or pleasures on a future Tuesday.

Parfit maintains that a practically rational agent could not seriously sustain FTI. I take it that the reason for its unsustainability has to do with a fundamental inconsistency between FTI and the agents other desires, or between FTI and the agent’s beliefs about its future existence, or between FTI and a (hypothetical) axiom of choice that requires an empathic or biased connection towards one’s future self.

So:


  • (11) It is not possible for an intelligent rational agent to sustain certain combinations of preferences or certain belief-desire pairings.
  • (12) If it is not possible for an intelligent agent to sustain every belief-desire pairing or every preference then the orthogonality thesis is false.
  • (13) Therefore, the orthogonality thesis is false.


Again, Bostrom’s response to this line of reasoning is disappointing. He says that by intelligence he means simply “skill at prediction, planning and means-end reasoning in general”, not Parfit’s normatively thick concept of rationality. But this is just to play a dubious definitional game. It seems to me like the reason Parfit’s FTI example works (if it does) is exactly because those with great skill at prediction, planning and general means-end reasoning would not be able to sustain an attitude of indifference toward their future Tuesday selves. If those skills are not involved in developing normatively thick rationality, then I don’t know what else Bostrom has in mind.

This is not to say that the FTI example works. It may still be possible for an agent to have FTI but this seems difficult to assess in the abstract (although plenty of philosophers have tried). So is it possible to create a superintelligent artificial agent with FTI? I have no idea, and I suspect we all continue to have no idea until one is actually created.


4. Weak Constraints on Orthogonality
Although I have suggested that Bostrom’s defence of the orthogonality thesis is lacking in certain respects, he himself acknowledges that there may be certain weak constraints on perfect orthogonality. The fact that he calls them “weak” constraints indicates that he doesn’t really view them as major objections to the orthogonality thesis, at least in terms of how that thesis might apply to artificial agents. But they are worth considering nonetheless.

Two of the weak constraints are relatively unimportant, at least in my opinion. One of them suggests that in order to “have” a set of desires, an agent would need to have some minimal integration between its intelligence and its decision-processes, which may in turn require a minimal level of intelligence. The other constraint is dynamical in nature. It concerns a hypothetical agent who is programmed with the final goal of becoming less intelligent. Such an agent would not sustain a high level of intelligence for long periods of time because its desire is incompatible with it (dynamical properties of this sort might be exploited by those keen on programming Friendly AI).

The third weak constraint is rather more important, particularly because it might open up the route to an alternative defence of the orthogonality thesis. The constraint is this: complex motivations, for example motivations requiring considerable time and resources to satisfy, might be incompatible with low levels of intelligence. This might be because they require a minimal amount of working or long-term memory, or something along these lines.

This sounds eminently plausible but it only applies to incompatibility at the low end of intelligence. Bostrom is not concerned about this since his ultimate concern is with the implications of creating a superintelligent agent. But, still, it does raise the question: if incompatibilities of this sort are possible at the low end, might they not also be possible at the high end? In many ways, this is exactly the question that the likes of Parfit are responding to with their normatively thick accounts of rationality. They are saying that high intelligence is not compatible with certain trivial or inconsistent desires.

But this leads me to consider an independent reason for thinking that high levels of intelligence are compatible with pretty much any kind of desire or goal. The reason comes from the (potential) truth of nihilism. As Nagel pointed out in his famous article on the topic, one of the key features of the nihilistic attitude is acceptance of the proposition that our present desires and projects are always contingent, questionable and capable of being overridden. What’s more this attitude is something that seems to be encouraged by many of our educational (i.e. intelligence-enhancing) activities. Thus, the capacity for critical thinking is often characterised in terms of the capacity to always question and challenge existing assumptions about ourselves and our relationship to the world.

To be clear, I’m not here suggesting that nihilism is true; I am merely suggesting that if it is true, and if its truth becomes more apparent as intelligence increases, it could provide powerful support for a kind of orthogonality thesis. I say a "kind of" orthogonality thesis because in its original form the thesis only focused on ultimate goals, but, of course, nihilism would imply that an intelligent being could not have an ultimate goal. But it could have any non-ultimate goal and this is presumably the same thing.


5. Conclusion
Summing up, in this post I’ve considered Bostrom’s discussion of the orthogonality thesis. According to this thesis, any level of intelligence is, within certain weak constraints, compatible with any type of final goal. If true, the thesis might provide support for those who think it possible to create a benign superintelligence. But, as I have pointed out, Bostrom’s defence of the orthogonality thesis is lacking in certain respects, particularly in his somewhat opaque and cavalier dismissal of normatively thick theories of rationality.

As it happens, none of this may affect what Bostrom has to say about unfriendly superintelligences. His defence of that argument relies on the convergence thesis, not the orthogonality thesis. If the orthogonality thesis turns out to be false, then all that happens is that the kind of convergence Bostrom alludes to simply occurs at a higher level in the AI’s goal architecture.

What might, however, be significant is whether the higher-level convergence is a convergence towards certain moral beliefs or a convergence toward nihilistic beliefs. If it is the former, then friendliness might be necessitated, not simply possible. If it is the latter, then all bets are off. A nihilistic agent could do pretty much anything since no goals would be rationally entailed by nihilism.