Is Craig's Defence of the DCT Inconsistent? (Part One)

Tweet

He's smiling now...

Forgive me. I am going to start with a self-indulgent bit of blog history.

I started this blog over three years ago. At the time, I saw it as an outlet. I was completing my PhD and finding that I was reading lots of things that weren’t directly relevant to my research. I found a lot of that stuff interesting and I didn’t want to let it go to waste. So this blog became my information dump: when I read something interesting I would write up a blog post about it so it I would have a permanent record of how I understood the arguments it made, which I could return to at a later date.

Originally, my main extra-curricular interest was in the philosophy of religion and ethics. Consequently, the majority of my early posts tended to cover those topics. Practically none of my early posts covered material related to my own research, mainly because I used this blog to get away from my PhD.

Obviously, things have changed quite a bit since then: my posts have become longer and more complicated (neither of which is necessarily of good thing), and the subject matter has moved away from looking at philosophy of religion toward looking more and more at ethical-legal issues that happen to be the focus of my current research and teaching. Nevertheless, I remain interested in the philosophy of religion, and like to occasionally do blog posts on it.

This post is going to be an example of that continuing interest. In it, I’m going to take a look at a recent paper by Erik Wielenberg entitled “An Inconsistency in Craig’s Defence of the Moral Argument”, which unsurprisingly argues that William Lane Craig’s defence of the modified Divine Command Theory (hereafter “modified DCT”) is best by a (fatal) inconsistency.

The paper is quite short, and readily available online, but I’m going to try to add some value to it by simplifying its elements, diagramming its main argumenst, and offering some commentary of my own. To that end, I’ll spread my discussion over two separate posts. In the remainder of this post, I will do three things. First, I’ll quickly sketch the basic structure of the dialectic between Craig's moral argument and its critics. Second, I'll explore one branch of the dialectic, which challenges the role of God in the explanation of moral facts. And third, I'll consider another branch of the dialectic, which challenges the assumptions underlying Craig's moral argument.


1. The Moral Argument and its Discontents
William Lane Craig, like many others, believes that there are objective moral truths. He believes that there are states of affairs in the world of which it is true to say “that state of affairs is good/bad" and that there are actions in the world of which it is true to say “that action is right/wrong”. What’s more he believes this without thinking that the truth conditions of either statement is wholly dependent on subjective states of those who might utter them. To put it more succinctly, he believes that there are objective (mind-independent) moral values and objective moral duties.

He also believes that objective moral truths can only exist if they have some “sound foundation” or explanation. What’s more, he believes that the only possible foundation or explanation for such truths is theistic in nature. This suggests that he is committed to something like the following argument:

• (1) There are objective moral truths. That is to say: there are objective moral values and objective moral duties. 

• (2) There is a sound foundation (explanation) for objective moral truths if and only if God exists (i.e. only God provides a sound foundation for objective moral truths). 

• (3) Therefore, God exists.

Those of you who are familiar with Craig’s work will realise that this is not a perfect replication of the moral argument that he typically presents in his debates and writings. To be precise, Craig doesn’t usually frame the second premise in terms of “sound foundations” or “explanations”; rather, he frames it in terms of the existence of such facts in the first place. In other words, he says such truths cannot exist if God does not exist. But the modest framing that I have adopted above is more appropriate given the dialectic that commonly arises between Craig and his critics.

What is that dialectic? Well, obviously, as with any logical argument of the sort presented above, there are two potential sites of criticism. One could criticise premise (1) and thereby reject the notion of objective moral truths. That would be radical and discomfiting to many, but there are some who take that approach. We won’t, however, be looking at it here. The other option is to criticise premise (2). That’s the one that’s relevant here and we’ll be spending the remainder of the post looking at it.

Significantly, the criticism of premise (2) can follow at least two separate branches (there is at least one more). The first branch — which we shall call the theistic branch — challenges Craig’s contention that God provides a sound foundation for objective morality. The second branch — which we shall call the foundationless branch — challenges the assumption of premise two, namely: that we need to provide a foundation for objective moral truths in the first place. For Craig’s argument to succeed, he has to cut off both of these branches of criticism.

Wielenberg’s claim is that in trying to do so, Craig contradicts himself. Thus, the way in which he prevents the theistic branch of the criticism from taking hold is in direct contradiction to the way in which he prevents the foundationless branch of criticism from taking hold. To see this, we need to carefully trace out the dialectic that takes place along both branches. The remainder of this post tries to do so.


2. The Theistic Branch of the Dialectic
The most direct critique of Craig’s argument takes place along the theistic branch of the dialectic. As a first step, the critic can pose the question to Craig: why think God provides a sound foundation for moral truths? Craig might (if he were extremely naive, which he isn't) duly oblige by saying: through his commands God tell us what is right and wrong, and good and bad. Thus, God's commands provide the foundation we need. This is the unmodified DCT.

The critic will immediately highlight a problem with this. As noted long ago by Plato in his dialogue Euthyphro, making moral truths dependent on God’s commands in this manner seems to render them disturbingly arbitrary. Allow me to explain. Take an objective moral duty of the following sort:

Duty_ct: It is morally wrong (read: impermissible) to torture an innocent child for fun.

This would seem to be an uncontroversial example of a moral duty. What’s more, it seems like the kind of moral duty that simply has to hold true, irrespective of the circumstances. In other words, it seems like there could never be a scenario in which it is morally acceptable to torture an innocent child for fun. There is no possible world in which such torture is permissible.

But what if God commanded it? What if he said: you must torture an innocent child for my amusement. If one is a proponent of the unmodified DCT, then it would seem like one is committed to the view that if God commands it, it becomes morally acceptable. So in that case, if God issued that command, the torture of the innocent child would become permissible. The deontic status of an act is suddenly dependent on the arbitrary whim of God.

That seems unpalatable to many — including many theists who accept that God must provide the ultimate foundation for moral truth. So they’ve come up with an escape route: the modified DCT. This was (I believe) originally formulated by Robert M. Adams, but Craig has become a staunch proponent of it in latter days.

The essence of the modified DCT is that moral duties are indeed grounded in God’s commands — thus, Duty_ct is true only because God has commanded it — but that there are constraints on what God can command. To be precise, because God is essentially and necessarily good, he could never command (for example) the torture of an innocent child for fun. Thus, certain key moral duties are not dependent on some arbitrary divine whim.

Of course, the upshot of all this is that certain moral truths must hold as a matter of logical necessity. For example, the torture of an innocent child is always and everywhere wrong because God’s nature is such that he forbids it in every possible world; because it is not logically possible for a being with that nature to command otherwise. Following Wielenberg, we call such moral truths “N-Commands”:

N-Commands: God’s nature is such that there are certain things that he forbids, and certain others that he obliges, in every possible world. That is: there certain logically necessary moral duties under the modified DCT.

This is a neat solution to the arbitrariness objection, but it comes at a cost. Or so Wielenberg will argue. To see what that cost, we need to proceed to the foundationless branch of the dialectic.

(Note: You may wonder what happens to objective moral values under the modified DCT. Does God still provide the foundation for them? After all, the modified DCT solves the abritrariness problem by saying that God’s nature is such that he commands certain things as a matter of necessity. But that seems to focus solely on obligations and duties, not on values. As it happens, Craig argues that God’s nature is the grounding for moral values. Thus, God remains the foundation for all moral truths. This grounding of values in the divine nature has problems that I’ve explored before on the blog.)


3. The Foundationless Branch
The second critique of Craig’s argument follows a less direct path, but it highlights a venerable and increasingly popular view among metaethicists.

What is this view? It can be called, for want of a better name, non-theistic non-natural moral realism (NTNNMR). It doesn’t exactly trip off the tongue. I know. But it accurately describes the view, which is something. NTNNMR holds that moral truths do exist, but that they don’t really need an explanation or grounding. They simple are true. Thus, the second premise of Craig’s argument relies on a faulty assumption. Objective moral truths don’t need what Craig claims he can provide.

There is much to be said for this view. That there are certain things that are self-explanatory, self-grounding or brute is widely accepted. After all, few people think that explanations or grounding exercises can continue indefinitely. There must be some stopping points after which it makes no sense to ask for explanations or foundations. Our conception of reality must bottom-out somewhere.

But Craig nevertheless objects to NTNNMR. Why so? Let’s hear from the man himself:

If our approach to metaethical theory is to be serious metaphysics rather than just a “shopping list” approach, whereby one simply helps oneself to the supervenient moral properties…needed to do the job, then some sort of explanation is required for why moral properties supervene on certain natural states.” (Craig in Is goodness without God good enough?, p. 180)

So Craig thinks that proponents of NTNNMR aren’t serious metaethicists. Their claim that moral facts need no grounding is a case of special pleading. No serious metaethicists play this game: they all think the supervenience of moral properties on natural properties requires some explanation.

But as Wielenberg points out, this is an odd claim to make. The supervenience relation is one of logical necessity. To say that the moral property of wrongness (call this M₁ supervenes of the natural states of childhood, torture, innocence and amusement (call these “N₁ - N₄), is to say that in any two possible worlds in which N₁ - N₄ hold, so too does M₁. In other words, it is logically necessary that if N₁- N₄ is true, so too is M_1.

The upshot of this is that when Craig complains about the “shopping list” approach to metaethics employed by proponents of NTNNMR, he is complaining about a failure to explain logically necessary connections. This suggests that Craig imposes the following success condition on a metaethical theory:

Craig’s Condition: Any approach to metaethics that posits the existence of logically necessary connections must adequately explain those connections.

Now, Wielenberg exploits the appeal to this condition in his challenge to Craig. We’ll look at that the next day. But I want to close by dwelling on the substance of this condition for a moment.

To me, it is somewhat redolent of the principle of sufficient reason (PSR). The PSR states (in one form) that for every fact F there is a sufficient explanation of that fact. But Craig’s Condition is both less and more extreme than that famous principle. It is less extreme in that it only applies to the explanation of moral facts — a restriction that could be called into question by critics of Craig’s position. But it is also more extreme in that it applies to logically necessary facts. Many would hold that logically necessary facts fall outside the remit of the PSR since they can be self-explanatory. So in calling for logically necessary connections to be explained, Craig is doing something quite extreme. Of course, Craig defends this by claiming that just because a fact is necessary does not mean it cannot be explained. For example, he claims that “2+2=4” is a necessary truth, but nevertheless it is explained by the Peano axioms This suggests that at least some logically necessary facts can be explained. All that needs to be shown then is that moral facts are among those necessary facts that both have and need an explanation. But can this be shown?

Wielenberg doesn’t answer this question in his short article. That is understandable. He has other fish to fry. Still, I find it quite interesting. And as it happens I have a piece currently under review somewhere that tries to critique Craig (and some other theistic metaethicists) on this very point. That, however, is a topic for another day. For now, I shall conclude this post.

In part two, I’ll outline Wielenberg’s critique in full detail, and consider his defence of it. Stay tuned.

The Singularity: Overview and Framework Redux

Tweet

(Original Version)

I want to begin with the problem.

A couple of weeks back — the 24th of December to be precise — I published a post offering a general framework for understanding and participating in (philosophical) debates about the technological singularity. The rationale behind the post was simple: I wanted to organise the dialectical terrain surrounding the technological singularity into a number of core theses, each of which could be supported or rejected by a number of arguments. Analysing, developing and defending the premises of those arguments could then be viewed as the primary role of philosophical research about this topic. Obviously, I was aware that people are engaged in this kind of research anyway, but I had hoped that the framework might provide some utility for those who are new to the area, while also enabling existing researchers to see how their work connects with that of others.

The framework I came up with centred on three main theses: (i) the explosion thesis, which was about the likelihood of an intelligence explosion of AI; (ii) the unfriendliness thesis, which was about the likelihood that any superintelligent AI would be antithetical to human values and interests; and (iii) the inevitability thesis, which was about the avoidability (or evitability) of the creation of superintelligent, unfriendly AI. In the earlier post, I suggested that each of these theses represented a point on a spectrum of possible theses, with the other points along the spectrum being made up of weaker or contrary theses.

The post attracted a number of comments, some of which were not directly related to the substance of the post, and some of which expanded (in very useful ways) on the various ideas contained therein. One comment, however, from Eliezer Yudkowsky was somewhat critical of the framework I had proffered. Upon reflection, I think that a significant portion of this criticism was on target. The main problems with the framework as I had formulated it was: (a) it conflated notions of possibility and probability when it came to the discussion of AI development; and (b) it made the inevitability thesis seem separate from the other two when in reality it was a sub-part of both. That led to an overly-complicated framework.

You can see the problem more clearly by looking at the three theses as I originally defined them (note: the explosion thesis appears in abbreviated form here):

The Explosion Thesis: There will be an intelligence explosion, i.e. there will be AI++.

The Unfriendliness Thesis: It is highly likely that any AI+ or AI++ will be unfriendly to us. That is: will have goals that are antithetical to those of us human beings, or will act in a manner that is antithetical to our goals.

The Inevitability Thesis: The creation of an unfriendly AI+ or AI++ is inevitable. (A number of related theses were developed which suggested that it might be strongly or weakly inevitable, or maybe even “evitable” or avoidable).

Here, both the Explosion Thesis and the Unfriendliness are phrased in terms of what is going to happen or in terms of what is likely to happen. And yet, despite this, the inevitability thesis (and the associated evitability theses) is phrased in terms of what can or cannot be avoided. But isn’t that simply to say that certain outcomes either likely or not likely to occur? And thus isn’t that already included in the definition of the other two theses? A framework which conflates and blends theses in this manner is likely to do more harm than good; to confuse rather than enlighten.

That is the problem.


1. Correcting the Problem: Thoughts and Methodology
The obvious solution to the problem is to provide a better overview and framework, and that’s what I want to try to do in the remainder of this post. But before I do so, I want to say a few things about the methodology I’m adopting and the limitations of my approach.

The methodology remains committed to same goals as last time round. First, I want to organise the dialectical terrain surrounding the technological singularity into a set of theses, each of which can be supported by one or more arguments, the analysis, evaluation and evaluation of which is the role of philosophical research in this area. Second, I want to ensure that the organisation is simple without being simplistic; that it succinctly covers the key topics in such a way that the debates can be isolated in a short space of time, while also retaining some reductive complexity for those who want to think about it in more detail. Finally, I want the framework to work in such a way that those who accept particular combinations of theses will end up with distinct and recognisable positions on the technological singularity.

Those three things — identification of core theses, simplicity with reductive complexity, and the ability to locate distinct positions or views — form my self-imposed conditions of success for the organisational framework. But obviously, any such organisational exercise starts with certain biases on the part of the organiser. To be more precise, the exercise starts with the organiser having a fuzzy sense of what ought to be included within the framework, which is hopefully rendered less fuzzy, and less arbitrary in the process of formulating and refining the framework. In this instance, my own fuzzy sense of what ought to be included within the framework is threefold.

First, I think some characterisation of the technological singularity is necessary. In other words, at least one of the theses has to say something about the possibility of AI creating AI+ and AI++. That, after all, is what the “singularity” is supposed to be. This is unsurprising given that the whole terrain is already grouped under that particular heading and this is what the Explosion Thesis was supposed to capture last time out. Second, at least one of the theses should cover the risks and rewards associated with the creation of AI+. Again, this is unsurprising given that concerns about the safety of AI technologies, along with hopes as to their beneficial potential, is what provides most of the motivation for researchers in this area. This is what the Unfriendliness Thesis was supposed to capture. And third, some characterisation of the psychological, social, and political factors influencing the development of AI++ — what I will call the “strategic infrastructure of AI development” — is necessary. These factors are distinct from those relating to the pure technological possibility of AI+, but assuming possibility they will obviously play an important role in its future creation. The Inevitability Thesis was supposed to capture this aspect of the debate, but I now think it failed to do so appropriately.

So this suggests the shape that the new framework ought to take. In essence, all three of these topics ought still to be covered, but they should not necessarily be represented as distinct theses. At least, not as distinct theses in the way they were before. Instead, the framework will be reduced to two core theses — the Explosion Thesis, and the Unfriendliness Thesis — each of which comes in two versions: (i) a technical version, which relates to conceptual and technological possibility; and (ii) a strategic version, which relates to all-things-considered likelihood, given either the current or plausible forms that the strategic infrastructure could take. Arising from these two theses, will be two decision trees, the endpoints of which will represent different positions one can take on the technological singularity. The decision tree format flows naturally out of the two versions of theses, since the strategic version depends on the technical version. That is to say: technologically possibility must be addressed before we consider the strategic implications.

The end result is a framework that (I think) corrects for the primary defects in the previous version, while still satisfying the three conditions of success mentioned above. Nevertheless, the framework still has its limitations. In addition to flaws I have not spotted myself, it will not be to everyone’s tastes since it works from my own intuitive biases about what is important in this area. It also ignores some interesting topics and runs the risk of constraining speculation in this area. Consequently, despite the fact that I’m dedicating two blog posts to the exercise, I don’t want to treat it too seriously. If this version of the framework is flawed — as it no doubt will turn out to be — then so be it. I won’t waste time perfecting it since it would probably be more worthwhile to evaluate and improve upon what has been written in this area.

Anyway, onto the revised framework itself. Remember: two theses, each coming in two flavours, which can then be co-opted into a pair of decision trees, the endpoints of which represent recognisable positions one can take on the technological singularity.


2. The Explosion Thesis
We start with Explosion Thesis, which in its technical form looks like this:

The Explosion Thesis_TV: It is possible (conceptually & technically) for there to be an intelligence explosion. That is: a state of affairs could arise in which for every AI_n that is created, AI_n will create a more intelligent AI_n+1 (where the intelligence gap between AI_n+1 and AI_n increases) up to some limit of intelligence or resources.

This is roughly the same as the version of the Explosion Thesis presented in the previous post but there are two major changes. First, in line with what I said above, this version now focuses on what is technologically possible, not on what is likely to happen. Second, I’ve added the important caveat here that in order for there to be a true intelligence explosion, the intelligence gap between AI_n and AI_n+1 must be increasing (at least initially). If the gap were merely incremental, there wouldn’t really be an explosion (h/t Yudkowsky’s comment on my previous post). The actual speed of the “takeoff” is left open. Following Chalmers, I’ll use the abbreviation AI+ to refer to any AI with greater-than-human intelligence, and the abbreviation AI++ to refer to the kind of AI one might expect to arise following an intelligence explosion.

A couple of other points are worth mentioning here. First, the notion of “intelligence” is somewhat fuzzy, and likely to be contested by researchers. In order to have a fully-functional version of the explosion thesis, one would need to pin down the concept of intelligence with more precision. However, in line with my goal of formulating an overarching framework for philosophical research on the singularity, I’m leaving things a bit fuzzy and general for now. Second, the Explosion Thesis continues to represent a point along a spectrum of theses about the possibility of AI, AI+ or AI++. The lower end of the spectrum is represented by complete skepticism about the possibility of AI, with the middle of the spectrum being occupied by the AI thesis (i.e. AI is possible but nothing more) and the AI+ thesis (i.e. AI+ is possible, but AI++ is not). My own feeling is that the risk-related concerns are somewhat similar for AI+ and AI++ (though, obviously, more pronounced for the latter), hence I tend to lump those two theses together. In other words, I think that if you accept the possibility of AI+, you should ask yourself the same kinds of questions you would if you accepted AI++.

I won’t talk too much about the arguments for and against the Explosion Thesis in this post. I mentioned a few the last day, and they remain roughly the same. The classic argument is that of I.J. Good, and Chalmer’s lengthy philosophical analysis of the singularity presents a couple of arguments too. I’m going to look at those in a later post.

Moving on then to the strategic version of the Explosion Thesis, which looks like this:

The Explosion Thesis_SV: It is highly likely (assuming possibility) that there will be AI++ (or, at least, AI+).

In essence, this version of thesis is claiming that the strategic infrastructure surrounding AI development is such that, if AI++ is possible, it is likely that it will, at some stage, be created. People will be motivated to create it, and the necessary resources will be available for doing so. Although this thesis is phrased in a parsimonious manner, hopefully it can be seen how it covers issues surrounding political, social and psychological incentives, not issues of pure technological possibility. Note, as well, how in this version of the thesis I have explicitly lumped the possibility of AI++ and AI+ together. This is in keeping with what I just said about the risks of both.

To defend the strategic version of the thesis, one would need to show the absence of motivational and situational defeaters for the development of AI+ or AI++. The terminology is borrowed from Chalmers, who defined the former in terms of the psychological and social incentives, and the latter in terms of resource limitations and the potential for self-destruction associated with the creation of AI++. I’ll talk about those some other time.

How does one construct a decision tree out of these two theses? Very simply, one takes the first node of the tree to represent a choice between accepting or not accepting the technical version of the Explosion Thesis, with subsequent nodes representing choices between weaker views about the possibility of AI, and, of course, the strategic version of the Explosion Thesis. As follows:

The Unfriendliness Thesis
The technical version of the Unfriendliness Thesis can be defined in the following manner:

The Unfriendliness Thesis_TV: It is possible (conceptually and technically) for an AI++ or AI+ to have values and goals that are antithetical to our own, or to act in manner that is antithetical to those values and goals.

First things first, as I noted the last time, “unfriendliness” is probably not the best term to use here. If we were being more philosophically precise, we might define this thesis in terms of objective and subjective values (or, more simply, “good” and “bad”) since the effect of AI+ or AI++ on those values is what really matters, not their “friendliness” in the colloquial sense. Furthermore, the “us” in the definition is problematic. Who are “we” and why do our values matter so much? By using “us”, the thesis runs the risk of adopting an overly species-relative conception of value. Nevertheless, part of me likes using the label “unfriendliness”. One reason is that it maps onto the pre-existing terms of Friendly and Unfriendly AI. Another reason is that it is more catchy than alternatives like the “Detrimental to Objective Value Thesis”. So I’m going to stick with it for now.

But why is this thesis interesting? Surely what matters when it comes to AI risks is their likelihood not their possibility, particularly given how philosophers speculate about all kinds of bizarre possibilities (Twin Earths and the like). This is where the qualifier of “technical” possibility becomes important. The focus of the technical version of the Unfriendliness Thesis is not on bizarre and far-fetched possibilities, but on what is possible given what we currently know about AI architectures and the relationship between intelligence and morality. Furthermore, it is important to consider these possibilities up front because, as has been acknowledged by Bostrom and Armstrong, philosophers might be inclined to head-off discussion of AI risk by arguing that intelligence and morality are positively correlated: an increase in one leads to an increase in the other. Now, I’m not suggesting that these philosophers would be right or wrong to make those conceptual and technical arguments, quite the contrary, I’m just suggesting that those arguments should be considered independently.

That leads nicely to the strategic version of the Unfriendliness Thesis, which reads as follows:

The Unfriendliness Thesis_SV: It is highly likely (assuming possibility) that any AI+ or AI++ that is created will have values and goals that are antithetical to our own, or will act in a manner that is antithetical to those values and goals.

In other words, given existing incentive structures and methodologies within AI research and development, and given the conceptual possibilities, any AI+ or AI++ that is created is likely to be unfriendly. There are lots of arguments one could offer in favour of this thesis, some of which were mentioned the last time (e.g. the vastness of mindspace argument, or the complexity of human values argument). To those could be added the argument that Yudkowsky mentioned in his comment to my previous post: the fragility of human values argument (though I’m not sure how different that is from the others). Opposing arguments have been presented by the likes of Alexander Kruel, Richard Loosemore and Ben Goertzel. Both Alexander and Richard commented on my earlier post. They tend to feel (if I may paraphrase) that current approaches to AI construction are unlikely to yield any serious form of unfriendliness (or any unfriendliness).

Even if one thinks it is highly likely that an AI+ or AI++ will be unfriendly, one could still be cautiously optimistic. One could think that with the right changes to the existing strategic infrastructure, safe AI+ becomes more likely. This, I presume, is what research centres like Oxford’s FHI, Cambridge’s CSER and the Singularity Institute would like to see happen, and helping to achieve it is a large part of their rationale for existence (though they do not, of course, ignore the other theses I have mentioned).

The decision tree associated with this thesis follows a similar pattern to that associated with the Explosion Thesis. The first node in the tree represents the choice between accepting (or not) the technical version of the Unfriendliness thesis. The other nodes trace out other possible combinations of views. The endpoints this time are given names. ‘Optimism’ is the name given to the view of those who reject the technical version of the Unfriendliness Thesis because they think that as intelligence goes up, so too does morality. “Deflation” is the name given to those who, although accepting the possibility of unfriendly AI, deem it highly unlikely, given current incentives (etc). The view is called “Deflation” because it seeks to burst the bubble of those who think that various doomsday scenarios are likely unless we make decisive changes to the strategic infrastructure. “Cautious Optimism” is the name given to those who think that unfriendly AI is probable given the current infrastructure, but avoidable if the right changes are made. And “Doom” is the name given to those who think unfriendly AI is unlikely to be avoided.

A final note: the binary representation of possible responses to the later questions in this decision tree is obviously flawed since there is a spectrum of probabilities one could attach to the various theses. Hopefully, the reader can make the necessary adjustments for themselves so that the decision tree better represents how they think about the probabilities.

4. Conclusion
So there it is, my revised version of the framework and overview. Once again, I’ve tried to organise the dialectical terrain around a few core theses. In this case there are two — the Explosion Thesis and the Unfriendliness Thesis — but each comes in two distinct versions — a technical version and a strategic version. One’s attitude toward the various theses can be represented as choices one makes on a decision tree, the endpoints of which constitute distinct and recognisable views about the technological singularity and its associated risks. My hope is that this improves on the previous version of the framework, but I accept it may not.

Book Recommendations ♯7: The Ethics of Voting

Tweet

(Series Index)

Things have been quiet on the blogging front over the past few weeks. But I have been doing a lot of reading over the Christmas period, and hope to start back with gusto at the weekend. In the interim, I thought I might fill the gap before more substantive blogging with a new book recommendation. I’m hoping to reinvigorate this series this year.

The book I want to recommend is Jason Brennan’s The Ethics of Voting, which I first heard about over on the New Books in Philosophy podcast. As we all know, there are certain propositions about voting that tend to be widely-believed in modern democratic societies. Indeed, they border on constituting the “sacred cows” of contemporary political discourse. These propositions include:

(1) You ought to vote, i.e. there is a moral (and in some countries legal) duty to vote.

(2) Every adult should be entitled vote, i.e. the principle of “one person, one vote” applies irrespective of intellectual, rational and moral virtue. (There are, perhaps, some accepted restrictions in some countries for prisoners, though the push is very much against this, particularly in Europe)

(3) People should not buy and sell their votes.

To be sure, there is a great deal of cynicism and apathy about these propositions. Many people fail to vote, and many argue that votes are “bought” and “sold” all the time. Nevertheless, I suspect most think that these propositions represent ideal or preferred states of affairs; that the failure to live up to them represents a corruption or degradation of the political process. Certainly, one encounters social groups that encourage people to vote all the time, without telling them exactly who to vote for. Thus, for example, non-partisan campaigns that encourage the youth vote are common in many countries. The suggestion from these campaigns seems to be that the act of voting is itself desirable, irrespective of its content. (The cynic may question this: since youth voters, in particular, tend to favour more certain candidates one could argue the motives behind such campaigns are not strictly impartial, but that’s by-the-by).

The nice thing about Brennan’s book is that it subjects each of three propositions given above (along with several more) to a sustained critique. In a patient yet concise manner, Brennan argues that there is no moral duty to vote; that it is not desirable for certain people with intellectual, rational and moral vices to vote; and that it is permissible for people to buy and sell votes. These are iconoclastic claims, and Brennan’s attempt to defend them is, I dare say it, in the best traditions of philosophy. It provides much food for thought, even if one isn't ultimately persuaded by what he has to say.

Another commendable feature of the book is its style. Every argument and proposition is presented with  clarity (formal arguments and propositions throughout), and the necessarily theoretical background is sketched out for the reader. One needn’t always agree with what Brennan says, but one can at least admire the way in which he says it.

On a final note, the book receives some pretty negative reviews over on Amazon, which should always be taken with a grain of salt, but I will conclude by saying that I agree with Chad Flanders's comment in his review of the book for NDPR: "Sometimes Brennan seems to revel too much in being contrarian, which nonetheless is what makes reading the book so much fun."

My Favourite Posts of 2012

Tweet

This year was a busy one on the blogging front, though perhaps less busy than previous years. I'll just scrape past the 100 posts-barrier this year, despite having cleared it fairly comfortably in the past. Still, it's quality not quantity that counts, and the majority of my posts this year have been pretty substantive.

They have also covered a diverse range of topics, including: the ethics of the death penalty, incest, abortion and infanticide, blackmail, neuroscience-based lie detection, immortality, theism and the meaning of life, enhancement in sports and education, new natural law, the same-sex marriage debate, skeptical theism, the ontological foundations of morality and more.

Anyway, here are some of my favourites from the past twelve months. They are not arranged hierarchically, but chronologically instead. I've tried to include one post or group of posts from each month, with February being the sole exception (not enough blogging was going on back then). I haven't focused on the popularity of the posts when compiling this list, but you may like to know that the most popular post(s were the two I did on the epistemic objection to torture. They got around 8,000 individual hits. That's not a great indicator of the number of people who read them though for two reasons: (i) most people probably don't read what they click through to; and (ii) the front page of this blog is set up in such a way that you can read posts in full without clicking through to them. 

1. Enhancement and Education: Lessons from the Kobayashi Maru (Part 1, Part 2)

Tries to offer some thoughts on the enhancement and education debate by analysing the infamous Kobayashi Maru test from the Start Trek fictional universe.

2. What the After-Birth Abortion article Said (Part 1, Part 2)

Two Italian ethicists, Alberto Giubilini and Francesca Minerva, got into trouble earlier this year for an article arguing for an equivalency between abortion and infanticide. In these two posts, I took a look at what their controversial article actually said.

3. Theism and the Meaning of Life (Part 1, Part 2)

William Lane Craig argues that it's impossible for life to have any meaning in the atheistic worldview; conversely he argues that it's possible on the theistic view. In this series of posts, I look at his arguments.

4. The Epistemic Objection to Torture (Part 1, Part 2)

Some people think that torture is morally permissible in cases of extreme emergency. Typically, this is said to be because torture can help us to obtain valuable information. But is this correct? Roger Koppl argues that it isn't, for reasons related to the formal structure of an epistemic system. In these posts, I see what he has to say.

5. Skeptical Theism and the Parent Analogy

Skeptical theism is the major contemporary response to the evidential problem of evil. But its foundations are often poorly defended. This post looks at Trent Dougherty recent article about this important topic.

6. Is Death Bad or Just Less Good? (Part 1, Part 2, Part 3, Part 4)

Is there anything to the classic Epicurean argument that death is not bad for the one who dies? In this series of posts, I consider Aaron Smuts's recent defence of a modified form of Epicureanism.

7. Should Incest be Decriminalised? (Part 1, Part 2, Part 3)

Nothing provokes the "yuck" response more readily than incest, but are the arguments for its criminalisation any good? Following an article by Vera Bergelson, I consider five traditional rationales for the criminalisation of incest.

8. Would Immortality be Desirable? (Part 1, Part 2, Part 3)

Bernard Williams argued for the tedium of immortality in the 1970s in his famous article "The Makropulos Case: Reflections on the Tedium of Immorality". This series of posts looks at some recent developments in this debate.

9. New Natural Law, Homosexuality and Religion (Part 1, Part 2)

New natural lawyers frequently argue against the legalisation of same-sex marriage on the grounds that it violates the intrinsic good of marriage. They often claim that their arguments can be embraced for purely secular reasons. These two posts consider the opposing view: that their arguments are inherently religious.

10. Is the Death Penalty Irrevocable? (Part 1, Part 2)

One major objection to the death penalty is that it is not revocable. If a person is wrongly imprisoned, they can be released from jail, but if they are wrongly killed they can never be resuscitated. Is this a good objection to the death penalty? These posts try to find out.

11. Implicature and the Interpretation of the Law (Part 1, Part 2)

In everyday conversation, our utterances often contain more meaning than is present in their linguistic structure. This is due to the phenomenon of conversational implicature. Can this phenomenon affect how we interpret the law as well?

The Singularity - Overview and Framework

Tweet

Roughly (I’ll refine later on) the “technological singularity” (or “singularity” for short, and in the right context) is the name given to point in time at which greater-than-human superintelligent machines are created. The concept (and name) was popularised by the science fiction author Vernor Vinge in the 1980s and 90s, though its roots can be traced further back in time to the work of John Von Neumann and I.J. Good.

The notion of a technological singularity is treated with derision in some quarters. But that attitude seems inapposite. As David Chalmers points out in his long (and excellent) discussion, the singularity is worthy of serious philosophical and intellectual consideration. The arguments for its occurrence are themselves intrinsically fascinating, raising all sort of philosophical and scientific questions. Furthermore, if it ever did happen, it would have serious practical, ethical and metaphysical consequences. Why should their consideration warrant derision?

Many now seem to agree. The California-based Singularity Institute has long been calling for (and more recently delivering) serious research on this topic. The Oxford-based Future of Humanity Institute has several researchers who dedicate themselves either full or part-time to the issue. And the recently-launched Cambridge Centre for the Study of Existential Risk includes the possible creation of superintelligent AI as one the four greatest threats to humanity, each of which it will be studying.

I also agree. Like Chalmers, I think there are some intrinsically and instrumentally important philosophical issues to consider here, but I’m not entirely sure where I come down on any of those issues. One of the reasons for this is that I’m not entirely sure what the main issues are. I’m a philosopher (sort of - not officially!). I work best when complex topics are broken down into a set of theses, each of which can be supported by an argument (or set of arguments), and each of which is a potential object of further research and contestation. Unfortunately, I have yet to come across a general philosophical overview of the singularity that breaks it down in a way that I like. Chalmers’s article certainly attempts to do this, and succeeds in many areas, but his focus is both too narrow, and too complex for my liking. He doesn’t break down the core theses in a way that clearly and simply emphasises why the singularity is a matter of serious practical concern, something that may pose a genuine existential threat to humanity.

This post is my attempt to correct for this deficiency. I write it in full or partial awareness of other attempts to do the same thing, and with full awareness of the limitations of any such attempt. Breaking a complex topic down into discreet theses, while often useful, runs the risk of overlooking important connections, de-emphasising some issues, and obscuring others. I’m happy to entertain critiques of what I’m about to do that highlight failings of this sort. This is very much a first pass at this topic, one that I’m sure I’ll revisit in the future.

So what am I about to do? I’m going to suggest that there are three main theses to contend with when one thinks about the technological singularity. They are, respectively: (i) the Explosion thesis; (ii) the Unfriendliness thesis; and (iii) the Inevitability thesis. Each one of these theses is supported by a number of arguments. Identifying and clarifying the premises of these arguments is an important area of research. Some of this work has been done, but more can and should be done. Furthermore, each of three main theses represents a point on scale of possibilities. These other possibilities represent additional points of argumentation and debate, ones that may be taken up by opponents of the core theses. As a consequence, one’s view on the technological singularity could be thought to occupy a location within a three-dimensional space of possible views. But I’ll abjure this level of abstraction and focus instead on a couple of possible positions within this space.

With that in mind, the remainder of this post has the following structure. First, I’ll present each of the three main theses in order. As I do so, I’ll highlight the conceptual issues associated with these theses, the arguments that support them, and the scale on which each thesis lies. Second, I’ll consider which combinations of beliefs about these theses are most interesting. In particular, I’ll highlight how the combination of all three theses supports the notion of an AI-pocalypse, but that weaker versions of the theses might support AI-husbandry or AI-topia. And third, I’ll briefly present my own (tentative) views on each of the theses (without supporting argumentation).


1. The Explosion Thesis
The core thesis of singularitarian thinking is the Explosion Thesis. It might be characterised in the following manner.

Explosion Thesis: There will be an intelligence explosion. That is: a state of affairs will arise in which for every AI_n that is created, AI_n will create AI_n+1 (where AI_n+1 is more intelligent than AI_n) up to some limit of intelligence or resources.

Here, “AI” refers to an artificial intelligence with human level cognitive capacities. Sometimes, the label AGI is used. This refers to an Artificial General Iintelligence and when used can help us to avoid confusing the phenomenon of interest to singularitarians with computer programs that happen to be highly competent in a narrow domain. Still, I’ll use the term AI throughout this discussion.

The first human level AI will be numerically designated as “AI₀”, with AI₁ used to designate the first AI with greater than human capacities. Some useful shorthand expressions were introduced by Chalmers (2010), and they will be followed here. Chalmers used “AI+” to refer to any AI with greater than human intelligence, and “AI++” to refer to vastly more intelligent AIs of the sort that could arise during an intelligence explosion.

Conceptual and practical issues related to the Explosion Thesis include: (a) the nature of intelligence and (b) the connection between “intelligence” and “power”. Intelligence is a contested concept. Although many researchers think that there is a general capacity or property known as intelligence (labelled g) others challenge this notion. Their debate might be important here because if there is no such general capacity we might not be able to create an AI. But this is a red herring. As Chalmers (2010) argues, all that matters is that we can create machines with some self-amplifying capacity or capacities that correlate with other capacities we are interested in and are of general practical importance. For example, the capacity to do science, or to engage in general means-end reasoning, or to do philosophy or so on. Whether there is some truly general capacity called “intelligence” matters not. The relationship between “intelligence” and “power” is, however, more important. If “power” is understood as the capacity to act in the real world, then one might think it is possible to be super-smart or intelligent without having any power. As we shall see, this possibility is an important topic of debate.

The original argument for the Explosion Thesis was I.J. Good’s. He said:

Let an ultraintelligent machine be defined as a machine that can far surpass all the intellectual activities of any man however clever. Since the design of machines is one of these intellectual activities, an ultraintelligent machine could design even better machines; there would then unquestionably be an 'intelligence explosion,' and the intelligence of man would be left far behind. Thus the first ultraintelligent machine is the last invention that man need ever make. (LINK)

Good’s argument has, in recent years, been refined considerably by Chalmers and Hutter (there may be others too whose work I have not read). I hope to look at Chalmers’s argument in a future post so I’ll say no more about it now.

As noted at the outset, the Explosion Thesis is a point on a scale. The scale is illustrated below. Below the Explosion Thesis one finds the AI+ Thesis. Defenders of this thesis agree that greater than human level AI is likely, but do not think this will lead to the recursive intelligence explosion identified by Good. In terms of practical implications, there may be little difference between the two positions. Many of the potentially negative (and positive) consequences of the Explosion Thesis would also follow from the AI+ Thesis, though the levels of severity might be different. Next on the scale is the AI Thesis, and below that is the No AI Thesis. I think these speak for themselves. Proponents and advocates of the singularity take up some position along this scale.

The Explosion Thesis is the core of singularitarianism. Unless one accepts the Explosion Thesis, or at a minimum the AI+ Thesis, the other theses will be irrelevant. But what does it mean to “accept” the thesis? I adopt a Bayesian or probabilistic approach to the matter. This means that even if I think the probability of an intelligence explosion is low, it might still be worth dedicating serious effort to working out its implications. For example, I think a 10% chance of an intelligence explosion would be enough for it to warrant attention. Where the threshold comes for theses that are not worthy of consideration is another matter. I couldn’t give a precise figure, but I could probably give analogous probabilities, e.g. the probability of my car suddenly morphing into a shark seems sufficiently low to me to not warrant my attention. If the probability of an intelligence explosion was similar to that, I wouldn’t care about it.


2. The Unfriendliness Thesis
This thesis may be characterised in the following manner:

Unfriendliness Thesis: It is highly likely that any AI+ or AI++ will be unfriendly to us. That is: will have goals that are antithetical to those of us human beings, or will act in a manner that is antithetical to our goals.

This thesis, and its complement The Friendliness Thesis, form an obvious spectrum of possible outcomes: at one end of the scale it is highly likely that AI+ will be unfriendly, and at the other end of the scale it is highly likely that AI+ will be friendly. One’s beliefs can then be represented as a probability distribution over this spectrum.

There are a couple of conceptual issues associated with this thesis. First, there is the question of what is meant by “friendliness” or “unfriendliness”. If the concept were introduced by a philosopher, it would probably be couched in terms of objective and subjective value, not friendliness. So, in other words, they would view the important issue as being whether an AI+ would act in manner that promotes or undermines objective or subjective value. Objective value would concern what is best from an agent-neutral perspective; subjective value would concern what is best from the perspective of one or more agents. The use of the term “us” in the definition of the thesis suggests that it is some form of subjective value that is being invoked — perhaps an aggregate form — but that may not be the case. What really seems to be at issue is whether the creation of AI+ would be “good” or “bad”, all things considered. Those terms are fuzzy, for sure, but their fuzziness probably enables them to better capture the issues at stake.

If the thesis does invoke some form of (aggregate) subjective value, it raises another conceptual question: who is the “us” that is being referred to? It might be that the “us” refers to us human beings as we currently exist, thus invoking a species relative conception of value, but maybe that’s not the best way to think about it. Within the debate over human enhancement, similar issues arise. Some philosophers — such as Nicholas Agar — argue that human enhancement is not welcome because enhanced human beings are unlikely to share the values that current unenhanced human beings share. But so what? Enhanced human beings might have better values, and if we are the ones being enhanced we’re unlikely to care too much about the transition (think about your own shift in values as you moved from childhood to adulthood). Something similar may be true in the case of an intelligence explosion. Even if an AI has different values to our own, “we” may not care, particularly if our intelligence is enhanced as part of the process. This may well be the case if we integrate more deeply with AI architectures.

The main argument in favour of the Unfriendliness Thesis would be the Vastness of Mindspace Argument. That is: the multidimensional space of possible AI+ minds is vast; most locations within that space are taken up by minds that are antithetical to our values; therefore, it is highly probable that any AI+ will have values that are antithetical to our own. Of course, this argument is flawed because it assumes that all locations within that mindspace are equally probable. This may not be the case. Additional arguments are needed.

The Orthogonality Argument (which I’ve looked at before) is one such argument. It has been developed by Armstrong and Bostrom and holds that, because intelligence and final goals are orthogonal, a high level of intelligence is compatible with (nearly) any final goal. In other words, there is no reason to think that intelligence correlates with increased moral virtue. Similarly, others (e.g. Muehlhauser and Helm) argue that values are incredibly complex and difficult to specify and stabilise within an AI architecture. Thus, there is no reason to think we could create an AI with goals that match our own.

The Friendliness Thesis might be held by a certain type of moral cognitivist. One that believes that as intelligence goes up, so too does the likelihood of having accurate moral beliefs. Since these beliefs are motivating, it would follow that AI+ is highly likely to be morally virtuous. Although many reject this claim, they may point to the possibility of a Friendly AI+ as a welcome one, and give this as a reason to be favourably disposed to the project of creating AI+. More on this later.


3. The Inevitability Thesis
The third thesis may be characterised in the following manner:

Inevitability Thesis: The creation of an unfriendly AI+ is inevitable.

The term “inevitable” is being used here in the same sense in which Daniel Dennett used it in his book Freedom Evolves. It means “unavoidable” and can be contrasted with the opposing concept of “evitability” (avoidability). One may wonder why I don’t simply use the terms “unavoidable” and “avoidable” instead. My response is that I like the idea of drawing upon the rich connotations of the “inevitability-evitability” distinction that arise from Dennett’s work.

In any event, the Inevitability Thesis obviously represents a point on a scale of possibilities that includes the Evitability Thesis at its opposite end. Those who subscribe to the latter will believe that it is possible to avoid the creation of unfriendly AI+. But we must be careful here. There are different ways in which unfriendly AI+ might be inevitable and different ways in which it might be evitable. Indeed, I identify at least two types of each. As follows:

Inevitability₁: The creation of an unfriendly AI+ with the ability to realise its goals is inevitable.(Strong Inevitability Thesis)

Inevitability₂: The creation of an unfriendly AI+ is inevitable, but it need not have the ability to realise its goals.(Containment/Leakproofing Thesis)

Evitability₁: It is possible to avoid the creation of an unfriendly AI+ and to realise the creation of a friendly AI+.(The AI-Husbandry Thesis)

Evitability₂: It is possible to avoid the creation of both unfriendly and friendly AI+.(Strong Evitability Thesis)

The two types of inevitability draw upon the distinction between intelligence and power to which I alluded earlier. Those who believe in strong inevitability think that intelligence carries power with it. In other words, any sufficiently intelligent AI will automatically have the ability to realise its goals. Those who believe in containment/leakproofing think it is possible to disentangle the two. Proponents or Oracle AI or Tool AI take this view.

The first type of evitability is the view adopted by the likes of Kurzweil. They are generally optimistic about the benefits of technology and intelligence explosions, and think we should facilitate them because they can be of great benefit to us. It is also the view that people in the Singularity Institute (and the FHI) would like to believe in. That is: they think we should try to control AI research in such a way that it brings about friendly rather than unfriendly AI+. Hence, I call it the AI-husbandry thesis. Its proponents think we can corral AI research in the direction of friendly AI+.

The arguments in this particular area are complex and multifarious. They range over a huge swathe of intellectual territory including international politics, the incentive structures of academic research, human psychology and foresightedness, technological feasibility and so on. It would be difficult to give a precis of the arguments here.


4. The AI-pocalypse or the AI-topia?
So there we have it: three main theses, each of which is supported by a number of arguments, and each of which can be viewed as a point on a scale of possible theses which in turn are supported by a variety of arguments. In other words, a complex web of premises and conclusions that is ample fodder for philosophical and scientific investigation.

Some combinations of views within this complex web are significant. The most obvious is the one in which the Explosion Thesis (or the AI+ thesis), the Unfriendliness Thesis, and the Strong Inevitability Thesis are accepted. The combination of these three views is thought to imply the AI-pocalypse: the state of affairs in which AI+ take over and do something very bad. This could range from enslaving the human population, to exterminating them, to destroying the world and to many other imaginable horrors. The precise form of the AI-pocalypse is not important, the fact that it is implied (or suggested) by the combination of the three theses is what matters.

Other combinations of views are more benign. For example, if you don’t think that AI+ is possible, then there’s nothing particularly significant to worry about. At least, nothing over and above what we already worry about. If you accept the likelihood of AI+, and the Unfriendliness Thesis, and the AI-Husbandry Thesis, then you have some reason for cautious optimism. Why? Because you think it is possible for us to create a friendly AI+, and a friendly AI+ could help us to get all sorts of things that are valuable to us. And if you accept the likelihood of AI+ and the Friendliness Thesis, you should be very optimistic indeed. Why? Because you think that the more intelligence there is, the more likely it is that valuable states of affairs will be brought about. So we’d better ramp-up intelligence as quickly as we can. This is the view of AI-topians.

I suspect, though I can’t say for sure, that cautious optimism is probably the prevalent attitude amongst researchers in this area, though this is modified by a good dose of AI-pocalypticism. This is useful for fundraising.


5. What do I think?
Honestly, I have no idea. I’m not sure I could even estimate my subjective probabilities for some of the theses. Still, here are my current feelings about each:

The Explosion Thesis: Good’s argument, along with Chalmers’s more recent refinement of it, seems plausible to me, but it relies crucially on the existence of self-amplifying capacities and the absence of intelligence and resource limitations. I couldn’t estimate a probability for all of these variables. I’d be a little bit more confident about the potential creation of AI+ for the simple reason that (a) I think there are many skills and abilities that humans have improved and will improve in the future (some through collective action), (b) artificial agents have several advantages over us (processing speed etc) and (c) I see no obvious reason why an artificial agent could not (some day) replicate our abilities. So I would put my subjective probability for this at approximately 0.75. I should also add that I’m not sure that the creation of a truly general AI is necessary for AI to pose a significant threat. A generally naive AI that was particularly competent at one thing (e.g. killing humans) would be bad enough (but some may argue that to be truly competent at that one thing, general intelligence would be required, I’ll not comment on that here).

The Unfriendliness Thesis: I have criticised Bostrom’s orthogonality thesis in the past, but that was mainly due some dubious argumentation in Bostrom’s paper, not so much because I disagree with the notion tout court. I am a moral cognitivist, and I do believe that benevolence should increase with intelligence (although it would depend on what we mean by “intelligence”). Still, I wouldn’t bet my house (if I owned one) on this proposition. I reckon it’s more likely than not that a highly competent AI+ would have goals that are antithetical to our own.

The Inevitability Thesis: This is where I would be most pessimistic. I think that if a particular technology is feasible, if its creation might bring some strategic advantage, and if people are incentivised to create it, then it probably will be created. I suspect the latter two conditions hold in the case of AI+, and the former is likely (see my comments on the Explosion Thesis). I’m not sure that there are any good ways to avoid this, though I’m open to suggestions.

At no point here have I mentioned the timeframe for AI+ because I’m not sure it matters. To paraphrase Chalmers, if it’s a decade away, or a hundred years, or a thousand years, who cares? What’s a thousand years in the scale of history? If it’s probable at some point, and its implications are significant, it’s worth thinking about now. Whether it should be one’s main preoccupation is another matter.

Pornography and Speech Acts (Part Two)

Tweet

(Part One)

This is the second part in my series looking at pornography qua speech act. The series is working from Mary Kate McGowan’s article “Conversational Exercitives and the Force of Pornography”. Before getting into the topic of this post, let’s review some of the material covered so far.

The feminist legal theorist Catherine MacKinnon has argued that pornography qua speech act silences and subordinates women. In other words, she argues that pornography does not merely cause the silencing and subordination of women, but is in fact itself an act of silencing and subordination. The distinction is subtle but important because it could take pornography out of the class of legally protected speech.

MacKinnon’s argument, such as it is, lacks the clarity and logical force of a typical philosophical argument. But then again MacKinnon isn’t a philosopher (not primarily anyway). Luckily, some prominent philosophers have tried to recast MacKinnon’s argument more perspicuous philosophical garb. One of them is Rae Langton. She has suggested that MacKinnon’s argument can be understood in terms of exercitive speech acts.

An exercitive speech act is one that sets the permissibility conditions on particular kinds of actions in particular domains. For instance, a club president who says “There shall be no more smoking on club premises!” performs an exercitive speech act. Taking this concept on board, Langton’s version of MacKinnon’s argument has two stages. First, she argues that pornographic speech could be exercitive in nature, specifically it could determine the permissibility conditions in the heterosexual sociosexual arena. Second, she argues that the kinds of permissibility conditions set out by pornographic speech could be such that women are silenced and subordinated.

Both stages of Langton’s argument are controversial, but McGowan criticised the first. She argued that Langton’s argument was flawed because it relied on the Austinian concept of the exercitive. According to this concept, in order to successfully and non-defectively perform an exercitive, a variety of conditions must be met. To be precise: (i) the speaker must intend (directly or otherwise) for their utterance to have an exercitive effect; (ii) the utterance must somehow convey the desired exercitive content; (iii) the listeners must be able to appreciate the exercitive nature of the utterance; and (iv) the speaker must have the requisite authority. The problem is that few, if any, of these conditions are met in the case of pornographic speech.

But McGowan is not deterred by this. She thinks it is possible for Langton’s argument to be salvaged. This is because there is an alternative type of exercitive speech act, one which is not subject beholden to the same success conditions as the Austinian exercitive, and which may justify the conclusion that pornographic speech silences and subordinates women. This is the “conversational exercitive”.

In the remainder of this post, I will explain what a conversational exercitive is and show how it might salvage Langton’s argument. It should be noted at the outset, however, that no strong conclusion about the actual effectiveness of the revised argument will be reached. Instead, we will be considering a proposal, an intriguing one for sure, but one with many crucial details yet to be worked out.


1. What is a Conversational Exercitive?
David Lewis said that a conversation is like a baseball game. Both are rule-governed enterprises, though the conversation rather more loosely so (more on this anon). In both, the permissibility of future behaviour depends on what has gone on before. And both have a “score”, where this is understood broadly to include all those facets of the game that are relevant to its assessment and proper play.

I don’t know enough about baseball to explain this analogy coherently, so I’ll focus purely on conversations. Two examples will help to flesh out the three features discussed by Lewis (both are taken from McGowan’s article, as are all other examples in this post):

Mike’s Dog: Mike and I are having a conversation. He mentions that his dog has been to the vet. I later ask whether “the” dog is okay.

Italian Boot?: Donal and several of his friends are talking about geography. Donal says that Ireland looks like a teddy bear lying on its side. Nobody questions his claim. But later on Seamus says that Italy doesn’t look like a boot because “it’s squiggly on both sides and boots usually aren’t”.

In the first example, Mike’s mentioning of his dog alters the salience of certain facts in the conversation. Although there are many dogs in the world, the fact that he has specifically referred to his dog means it is okay for me to later refer to “the dog” without there being any confusion as to which dog I am referring to. This is interesting because it suggests that what I am entitled to say within the conversation is altered by what Mike has said. In other words, the conversational “score” is adjusted by Mike’s initial utterance.

Something similar is true in the second example. When Seamus says that Italy looks nothing like a boot because of its squigglyness, he alters the standards of accuracy that operate within the conversation. His claim is true enough, but only because the standards of accuracy have been raised. This affects future contributions to the conversation. Once again, the score has been adjusted.

Both examples show how conversations are loosely rule-governed enterprises in which the permissibility of future behaviour depends on what has gone before. In my conversation with Mike, I could not have used the expression “the dog” to refer to my own dog, or some dog other than Mike’s. Well, that’s not quite true. Of course, I could have done so, but confusion would certainly have ensued. If I wanted to introduce a new dog into the conversation, I should have done so and changed the salience facts myself. The same is true of Seamus. His assertion may be challenged on the grounds that the standards of accuracy are not that high in this conversation, but, equally, the participants could adjust to the higher standards of accuracy. That would make it impermissible to say that Ireland was like a sideways teddy bear later on.

Now we come to the crux of the matter. McGowan’s central claim is this: conversational contributions like those in Mike’s Dog and Italian Boot? are exercitive in nature. They are conversational exercitives. That is to say, like the Austinian exercitive, they enact permissibility conditions for future conversational contributions, but unlike the Austinian exercitive, they do not rely on the four conditions of success mentioned earlier.

McGowan argues for this in depth in her article, I’ll only sketch the argument here. First, conversational exercitives, unlike their Austinian brethren, don’t have to express their exercitive content. When the club president bans smoking, his utterance has to say as much, but Mike does not have to say anything like “all future references to “the” dog must be references to my dog” for his conversational contribution to have its exercitive effect. Second, this implies that conversational exercitives are not sensitive to speakers’ intentions in the same way that Austinian exercitives are. Rather, they function in a largely covert and automatic way. Third, listeners need not consciously recognise the exercitive intent or purpose of the utterance. And fourth, authority is not an issue here. All conversational participants have the authority to alter the permissibility conditions in this manner.


2. Pornography as a Conversational Exercitive
The conversational exercitive is certainly an intriguing concept, one worthy of deeper analysis and consideration, but what good does it do here? Remember, the ultimate goal is to defend something akin to MacKinnon’s argument and we haven’t really begun to approach that goal yet. Several further steps must be taken. For starters, it needs to be shown that pornographic speech is a conversational contribution. Then, it needs to be shown that it is exercitive in nature. And finally, it needs to be shown that as an exercitive it serves to silence and subordinate women (at least some of the time). What for Langton was a two stage argument is now a three stage argument.

So let’s start with stage one. Is pornographic speech a conversational contribution? McGowan argues that it is. Echoing Langton, her claim is the following:

• (1) Pornographic “speech” is part of the ongoing “conversation” in the sociosexual arena.

The inverted commas are intended to highlight the fact that we are not dealing with a paradigmatic conversation here; rather, we are dealing with something that it is akin to a conversation in all important respects. Thus, the supporting argument for (1) is analogical in nature. It suggests that the properties typical of paradigmatic conversations are also typical of what goes on in the heterosexual sociosexual arena.

Like a paradigmatic conversation, the sociosexual arena is loosely rule-governed. At any given time, in a given sexual context, certain kinds of behaviour are permissible and certain kinds are not. This is what happens in conversation: certain things can be said and certain things cannot. Further, the rules that operate within the sociosexual arena adapt to fit the behaviour of the participants. If two sexual partners are turned on by yodeling, then yodeling becomes an appropriate form of foreplay, even if this would not be appropriate in many other sexual encounters. This suggests that the sociosexual arena is, in general (and ideally), a cooperative one: one in which the individuals coordinate their activities for mutual gain. Many linguistic theorists, such as Grice, hold that this is true of conversations too. If we accept that the sociosexual arena is akin to a conversation, the notion that pornography is a contribution to that conversation becomes much more palatable.

That brings us to the next stage of the argument. Once we accept that pornographic speech is a contribution to the sociosexual “conversation”, we need to show that it can be exercitive in nature. At this point McGowan’s argument gets rather sketchy (as she herself acknowledges). But one could well imagine that pornography has an exercitive function. Suppose that two sexual partners watch pornography together, that pornography repeatedly depicts people engaging in sexual act X, and neither of them objects to this act (and maybe even signal approval), one could then say that the pornography enacts permissibility conditions for sexual behaviour. The pornography says that act X is acceptable or permissible, and it becomes difficult for the two partners to deny this in the future. Thus:

• (2) Pornographic speech, on at least some occasions, functions as a conversational exercitive within the sociosexual arena, i.e. it enacts permissibility conditions for sexual behaviour.

But even allowing for this possibility, does it enact permissibility conditions that silence or subordinate women? The argument gets even sketchier at this point, but again McGowan suggests that it is possible that, on at least some occasions, it does. For instance, pornography might repeatedly “say” that whenever women say “no” to certain kinds of sexual advance they are really coyly signalling sexual acceptance. It may thus become impossible for some women to refuse sexual advances in some contexts. This would silence them because they would be unable to perform speech acts that they ought to be able to perform.

This could also hold true for subordination. Again, pornographic material might repeatedly signal that women are permissibly treated in a subordinating and dehumanising manner, and this might become part of the “score” of the sociosexual conversation. This gives us:

• (3) The permissibility conditions enacted by pornographic speech may, on at least some occasions, silence and subordinate women.

Which allows us to reach the conclusion:

• (4) It is possible that pornography, qua speech act, serves to silence and subordinate women.

Note how weak the conclusion here actually is. It does not say that pornography does silence and subordinate women, it merely says it is possible for it do this. A lot more would need to be done to satisfy the anti-porn position held by the likes of MacKinnon. But that’s in keeping with the “modest” (her words) aim of McGowan’s article. She is merely introducing a proposal that might work in MacKinnon’s favour. She is not actually defending MacKinnon’s view, nor is she saying that pornography should be banned or otherwise restricted.

Still, McGowan thinks there might be some flaws in her argument even when understood in this weak, proposal-like form. She addresses three of them. Let’s close by looking at these.


3. Is the Proposal Flawed?
The first objection is the mere fiction-objection. It has been pointed out by others that whenever a conversational participant tells a story or joke, the normal illocutionary force of the utterances that make up that joke seem to be suspended. Thus, if I say “A rabbi, a priest and a nun walked into a bar…” as the preface to a joke, I am not taken to be genuinely asserting this to be the case. In other words, I am not claiming that three such people did walk into the bar and I am not taking responsibility for the falsity of this claim.

But if the illocutionary force of an utterance is suspended when telling a joke or story, why is this not true of the exercitive force of pornographic speech? If pornography is mere fiction, then maybe it doesn’t have the exercitive force of other non-fictional contributions to the sociosexual conversation.

• (5) Pornography is mere fiction hence it does not have the illocutionary force typical of non-fictional contributions to the sociosexual conversation.

McGowan says this is wrong. Although it is true to say that some illocutionary forces are suspended during the telling of a story, the exercitive force need not be. Furthermore, the exercitive force can extend beyond the fictional domain in which it is first presented. She gives a long example of this, which I’ll abbreviate here. Suppose I start telling a joke about a chicken pecking at a Guinness tap in bar. By doing so, I have altered the salience facts of the conversation. This renders certain future conversational contributions permissible. For example, it is suddenly okay for my conversational partner(s) to start talking about chickens and his like or dislike for them. The joke enacts permissibility conditions for future conversational contributions and these permissibility conditions hold outside of the joke-telling portion of the conversation.

• (6) Purely fictional utterances can still have an exercitive effect within a conversation, and that exercitive effect can extend beyond the fictional realm of the utterance.

Of course, that’s just one example. McGowan never shows how the same is true of pornographic speech. Indeed, this is a general problem with her proposal. Even if it is sketchy, and intended merely to provoke further debate and discussion, it is really opaque about the mechanisms through which pornographic speech achieves its exercitive effect. Or, as in this case, how it manages to avoid having a purely fictional effect.

A second, and arguably more serious, problem is that the permissibility conditions enacted by conversational exercitives are extraordinarily weak. Go back to the Mike’s Dog example from earlier on. If Mike first talks about his dog, and I later refer to “the dog” without intending it to refer to Mike’s dog, the claim is that I have violated some permissibility condition in the conversation. But so what? Mike could easily adjust to or accommodate my utterance. For instance, if I say “the dog slept in my room last night”, Mike will realise that I’m not referring to his dog since he knows his dog slept in the shed last night. As with any good conversational partner, Mike will be inclined to cooperate with me and adjust the rules accordingly. Hence, I don’t really violate any rule. Not in a serious way anyway. Couldn’t the same be true of pornography?

• (7) The permissibility conditions enacted by conversational exercitives are extremely weak: it is very difficult to violate them in a serious manner because the rules of the conversation constantly adjust to accommodate the behaviour of the parties.

McGowan responds by saying that, although this may sometimes be true, it doesn’t rule out the serious violation of permissibility conditions in some conversational contexts. This is because the violation may go unnoticed at first, which then creates tensions or problems downstream. For instance, when Mike and I start talking, respectively, about “the dog” and its various antics, we might not realise that we are talking about different dogs. In this case, the violation of the rule goes unnoticed and problems will ensue at a later stage as we become more and more confused about the dog and what it got up to. How this might work in the case of pornography is another question, but this is an interesting idea nevertheless.

• (8) It is possible for there to be a serious violation of some of the permissibility conditions enacted by conversational exercitives, e.g. as when the initial violation goes unnoticed.

The final, and related, problem is that the kinds of permissibility conditions enacted by conversational exercitives are easily reversed. Thus, to use the dog example again, though our repeated use of the phrase “the dog” to refer to different dogs might violate permissibility conditions within the conversation we are having, Mike could easily rectify the problem by taking some time out to clarify that the phrase actually refers to his dog. Again, couldn’t the same be true of pornography? In other words, even if the pornography changes the permissibility conditions within the sociosexual arena, couldn’t one of the participants reverse those changes by enacting a new set of permissibility conditions?

• (9) Conversational exercitives are easily reversed, thus any exercitive effect that pornographic speech might have can be overturned by future conversational contributions.

In response to this, McGowan argues that some conversational exercitives might be hard to reverse. For example, in the Italian Boot? case given earlier, Seamus changed the standards of accuracy within the conversation by asserting that Italy didn’t look like a boot because of its squiggly outline. It has been observed by other philosophers of language (Lewis is specifically mentioned) that once standards of accuracy are raised within a conversation, it is difficult to reverse them. (I don’t know why)

In this instance, McGowan argues that something similar is true in the sociosexual arena. Specifically, she says that once a formerly taboo sexual practice (e.g. anal or oral sex) becomes an accepted part of the sociosexual conversation, it becomes difficult to revert to the taboo. Pornographic speech may play an important role in breaking down these sexual taboos. Similarly, if it really is true that pornographic speech makes it difficult for some women (in at least some contexts) to refuse sexual advances, this might be very difficult to reverse. If “no” is really taken to mean “yes”, then it’s difficult to know what someone could do to reverse that state of affairs. Repeatedly saying “no” doesn’t solve the problem.

• (10) Certain conversational exercitives are hard to reverse, including, plausibly, the kinds of exercitives related to sexual taboos and sexual refusals.

The diagram below illustrates the various stages of the dialectic.

4. Concluding Thoughts
That brings us to the end of McGowan’s article. To briefly recap, McGowan has argued that it is possible that pornography qua speech act silences and subordinates women. This is because pornographic speech might be exercitive in nature: it might enact permissibility conditions within the sociosexual arena. To support this claim, McGowan introduced the concept of the conversational exercitive, which is distinct from the Austinian exercitive, and can more plausibly be used in this debate.

Two concluding thoughts occur to me. First, although McGowan’s proposal is pretty sketchy in this article — particularly about the mechanisms through which pornographic speech replicates the effect of the conversational exercitive — she has not been idle since it was originally published. She has written several other articles that expand upon these basic ideas. I have not read those articles. It could well be that the necessary detail is found in them.

Second, if McGowan is right to view the sociosexual arena as akin to a conversation, and to view pornography as a contribution to that conversation, it is difficult to see why pornography should be singled out for approbation. Many other contributions to sociosexual conversations — e.g. actual conversations about sex, the publication of sex tips and advice columns etc. — could be exercitive. What’s more, those contributions could well have the silencing and subordinating effect alluded to by the likes of MacKinnon. Whether that’s a sensible thought or not, I leave to the reader to decide.