Sunday, March 31, 2013

The Logic of Indirect Speech (Part Three)

I was just thinking we could take of it right here, in Brainerd


(Part One, Part Two)

Over the course of the past two blog posts, I have been looking at different aspects of Pinker, Nowak and Lee’s (PNL’s) theory of indirect speech, as presented in their article “The Logic of Indirect Speech”. PNL’s goal is to offer some explanation for the prevalence of indirect speech acts in human social discourse.

The motivating reason for doing this is that indirect speech poses something of a paradox. If people do not directly say what they mean — i.e. if they try to implicate what they mean — they adopt a form of locution that is inefficient and runs the risk of being misunderstood. So why bother? Why not just say what you really mean?

PNL argue that part of the answer to this question comes in the shape of plausible deniability. What indirect speech gives us, that direct speech cannot, is the ability to deny certain implications of our speech. This is useful in a variety of domains. In part one, we looked at the simple example of the police bribe, where denying that you are offering a bribe can be of great value. Similarly, in part two, we looked at the case of relationship negotiation, where plausible deniability allowed one to hover between different relationship types and thereby avoid giving offence.

This seems like a decent start to the explanation of indirect speech, but there is, however, a final puzzle that needs to be addressed. Why are there so many indirect speech acts whose implications cannot be plausibly denied? For example, expressions such as “would you like to come up and see my etchings?” or “this is a real nice place, it would be a shame if something happened to it” are indirect, but everyone knows what they really mean. The conventional implication is so widely understood that it seems odd that anyone would use them. So why do they stick around?

In the remainder of this post, we will look PNL’s solution to this puzzle. As we shall see, their argument is that even when the conventional implication of the phrase is widely understood, the phrase has some utility. To appreciate their argument, we need to do three things. First, we need to appreciate the digital nature of language. Second, we need to understand the concept of common knowledge. And third, we need to see how indirect speech blocks the route to common knowledge.


1. Language as a Digital Medium
PNL argue that language is tacitly perceived to be a digital medium. In using this term “digital” they mean to draw upon the digital/analog distinction. Roughly, a digital signal is one in which something has either one value or another, there is no intermediate state; whereas an analog signal is one that is continuously variable. The classic example here comes from musical recordings. In an analog recording, a signal is taken from a recording device and directly laid down on (say) magnetic tape in an continuously varying sine wave. In a digital recording, the signal is sampled and laid down in a discrete discontinuous wave.

PNL say that language consists entirely of discrete signals — morphemes, words and sentences — which are combined in order to communicate meaning. There is a reasonable amount of evidence underlying this, some of which is discussed in Pinker’s book The Stuff of Thought. To give a quick example, phenomena that are continuous in the real world (e.g. space and time) are digitised by language (e.g. “a moment in time”), resulting in a communicated meaning that is discrete and discontinuous.

None of this is to say that there aren’t a huge variety of shades of meaning present in language. Anyone with a passing familiarity with interpretive practices knows this. It is, however, to say that words and sentences are typically perceived to have a discrete meaning. In other words, people typically believe you to be saying one thing or another, not both.

This has three consequences for indirect speech, according to PNL. They are:

Overt Propositions serve as Focal Points: A focal point is a location within a physical space that two rational agents, acting independently, can coordinate on. The concept comes from Schelling’s work on game theory. He gave the example of two agents who wish to meet each other in New York on a particular day, but have not agreed a time and place. He argued that Grand Central Station at 12 noon would be an obvious focal point. PNL claim that something similar is true of overt or direct propositions. They are focal points within semantic space and two rational agents can coordinate on their meaning, acting independently. Indirect propositions, on the other hand, are not focal points and thus cannot serve the same coordinative function, even when their implications are 99% certain.
Playing to an Audience: According to some psychological theories (e.g. Goffman’s) people are always playing to a hypothetical audience,, when they say or do things, even if they have a real audience. Thus, when I talk to you I am obviously trying to say something to you, but I might also be playing to a hypothetical audience of eavesdroppers. This has an interesting effect on my use of indirect speech acts. Although the implications of those speech acts might be very clear in the particular context of my conversation with you, they might not be clear to the hypothetical audience. Thus, I may continue to use them, even when their implications are certain, because I have that audience in mind.
Blocking the Route to Common Knowledge: Because language is typically perceived to mean one thing or another, an indirect speech act — which hovers between at least two interpretations — can block the route to common knowledge. This has an interesting psychological effect on how people interpret the indirect speech, even if they are sure what the implication actually is.


This final suggestion is of most interest to me and so I’ll spend the remainder of the post fleshing it out. But I can’t resist passing some comment on the first suggestion. While I like the idea that overt propositions serve as focal points, I wonder whether this really helps to resolve the problem we started with. After all, focal points are largely a matter of convention. Grand Central Station serves as a focal point only to people who share a common set of assumptions. Similarly, a focal point of this sort could change over time. Thus, if Grand Central Station were knocked, another location might become a focal point. If this is right, then why can’t focal points change within semantic space too? Why can’t an initially indirect proposition come to treated like an overt proposition. Indeed, I have hard time believing that this hasn’t already happened when it comes to allegedly indirect locutions like “would you like to come up and see my etchings?”


2. Blocking the Route to Common Knowledge
If we are going to understand the “blocking the route to common knowledge”-argument, we’ll first need to understand what common knowledge really is. This is a technical concept from the literature on game theory, which can be defined in the following manner:

Common Knowledge: A proposition P is common knowledge among a group of agents G, if everyone in G knows P, they all know that they know P, they all know that they all know that they know P, and so on ad infinitum.

An example would be the following. You and I are sitting at a desk opposite each other. I take a banana out of my pocket and place it on the desk, announcing the fact as I do so, knowing that you can see and hear me. In this instance, there is common knowledge that there is a banana on the desk. Why? Because you and I both know that there is a banana on the desk. I know that you know that there is a banana on the desk, and you know that I know that there is a banana on the desk. Furthermore, I know that you know that I know that there is a banana on the desk, and you know that I know that you know that there is a banana on the desk. (And so on ad infinitum).

If you are confused by the recursive phrases — “I know that you know that I know…” — in this example, you are not alone. It is a difficult concept to wrap your head around. One way to get a better grasp on it is to contrast common knowledge with the closely related, but importantly different, concept of “mutual knowledge”. This is perhaps best illustrated by an example. This one is taken from Ben Polak’s lectures on game theory, available from the Open Yale website (lecture 2 if you must know).
Suppose we have two people facing forwards, looking in the same direction, but not at each other. Suppose that a third person approaches these two people from behind and then places a pink hat on each of their heads. This third person then gets the two people to face each other. Is the fact that at least one person is wearing a pink hat common knowledge?



The answer is: no. If you look at the diagram above, and follow my reasoning, you can see that the conditions for common knowledge are not met here. Person X knows that person Y is wearing a pink hat; and person Y knows that person X is wearing a pink hat. But X doesn’t know that Y knows that X is wearing a pink hat, nor does Y know that X knows that Y is wearing a pink hat. This is because neither of them knows what kind of hat they are wearing.

This actually gives us everything we need to understand PNL’s argument about indirect speech. For it is the distinction between common and mutual knowledge that allows them to conclude that indirect speech blocks the route to common knowledge, even when its implications are virtually certain. To put it most succinctly they argue that indirect speech gives us mutual knowledge, but not common knowledge.

Take a scenario in which A says to B: “would you like to come up and see my etchings?”. We can assume that A knows what the implication of this really is, and we can also grant that B knows what the implication is. Thus, there is mutual knowledge of the implication. But there isn’t common knowledge of the implication. Why not? Because B doesn’t know that A knows what the implication is, and A doesn’t know that B knows what the implication is. This is true even if there is 99% certainty of the implication on both sides. The lingering uncertainty blocks the route to common knowledge. This wouldn’t be true in the case of a direct proposition. If A had said to be “would you like to come up and have sex?” there would be common knowledge that A had made a sexual advance.
I think this is a pretty interesting argument.


3. Conclusion
To sum up, in this series we’ve looked at PNL’s theory of indirect speech. According to PNL indirect speech is used — despite its inefficiency — because it provides the speaker with a valuable commodity: plausible deniability. This is true even if the indirect speech act has an implication which is widely understood.

In this post, we explored some of the reasons for this. The primary argument advanced by PNL is that because language is tacitly perceived to be a digital medium — i.e. a medium in which words digitise continuous properties, and a proposition has either one meaning or another — the value of indirect speech is maintained even where its implications are widely known. One of the main reasons for this is that the use of an indirect speech act blocks the route to common knowledge.

Saturday, March 23, 2013

The Logic of Indirect Speech (Part Two)

.... I was just thinking we could take care of it right here, in Brainerd.


(Part One)

An indirect speech act is one in which a speaker does not say what they really mean. Such speech acts are prevalent in a variety of social contexts. On the face of it, this poses something of conundrum as these speech acts are inefficient and prone to being misunderstood. So why are they so common?
In their article, “The Logic of Indirect Speech”, Pinker, Nowak and Lee (hereinafter “PNL”) try to develop a theory that explains the prevalence of indirect speech. They do so by eschewing classic cooperation-based theories of implied meaning, in favour of a conflictual and strategic theory. As they see it, people rely on indirect locutions because doing so allows them to achieve desirable outcomes when interacting with others who may have different or unknown interests.

We looked at the first part of their theory the last day. This argued that indirect speech is used in certain settings because it provides the speaker with a valuable commodity, namely: plausible deniability. The classic illustration of this came in the shape of the Briber’s Dilemma, a shorthand name for the decision problem in which a speeding driver must decide whether to bribe a traffic cop. In this scenario, an implied bribe can be the rational choice because it provides the speaker with the best of both worlds: the scrapping of their ticket if the cop is dishonest, coupled with no risk of arrest for bribery if the cop is honest.

The thing is, it’s easy to see the value of plausible deniability in such a context. The costs and benefits are readily discernible, and so the rationality of indirect speech reveals itself with a minimum of fuss. It’s more difficult when the costs and benefits are opaque or hidden. This is where the next part of PNL’s theory comes into play. They argue that the phenomenon of relationship negotiation can help us to unearth the hidden value of plausible deniability in other social contexts.

In the remainder of this post, I look at PNL’s discussion of relationship negotiation. I do so in three parts. First, I discuss the three main types of social relationship (according to Alan Fiske). Second, I explain the problem of relationship ambiguity. And third, I explain why indirect speech in general — and plausible deniability in particular — can play a valuable role in addressing the problem of relationship ambiguity.


1. The Three Main Types of Relationship
We all know that human social intercourse is mediated through a variety of relationships. In my capacity as a university lecturer, I stand in a particular kind of relationship with my students. The relationship comes with a set of norms and standards that most people try to respect. It is, however, a very different kind of relationship from the one I have with my friends and family, both of which have a distinctive set of norms and standards associated with them.

Mapping out all the possible relationships and the concomitant norms and standards sounds like an exhausting task. Fortunately, social scientists have been trying to do this for some time, and although I imagine there is no perfect taxonomy of relationship-types, Alan Fiske’s taxonomy is the one used by PNL. Fiske argues that all human relationships (across all societies) fall into three distinct categories:

Dominance Relationships: These are relationships in which one party exerts dominance or authority over another, and the other defers to the dominant party. This relationship can be seen in the animal kingdom where dominance hierarchies are common. As PNL put it, this relationship is governed by the ethos “Don’t mess with me”.
Communal Relationships: These are relationships in which resources are shared freely between the parties, resulting in almost perfect equality. As PNL put it, this relationship is governed by the ethos “What’s mine is thine and what’s thine is mine”.
Reciprocal Relationships: These are relationships that lie somewhere in between the other two. They are characterised not by dominance, nor by free sharing of resources, but by tit-for-tat exchanges. As PNL put it, the governing ethos is “You scratch my back, I’ll scratch yours”.

Fiske actually identifies a fourth relationship type (“Market Pricing”) which is found only in industrial societies. It is not discussed by PNL so I will ignore it (perhaps to my own cost).

The key thing about each of the three relationship-types listed above is that in addition to having a governing ethos, there is a distinctive style of communication associated with each. Thus, in the dominance relationship, imperatives, direct requests and non-verbal dominance displays are common. In the communal relationship, nonverbal cues, physical contact and coordinated activities are common. And in the the reciprocal relationship, actual tit for tat exchanges are the dominant signal, but these can be negotiated through explicit verbal contracts.


2. The Problem of Relationship Ambiguity
If at any given time everyone was aware of the exact relationship-type they happen to be in, then there may be no role for the plausible deniability of indirect speech. The distinctive communicative style associated with each of the three relationships would provide perfect information to the respective parties, and they could thereby say what they mean, provided that what they say fits within the ethos of the particular relationship.

The problem, however, is that relationships can be ambiguous. This arises from mixed beliefs and mixed signals. For example, sometimes I think that the relationship I have with my students is one of dominance and authority. But at other times I think this wrong and that it should be viewed as a relationship of reciprocity or maybe even communality. It’s quite possible that the students are unsure of this too. Some of the verbal and nonverbal cues I send out might indicate that it is, indeed, a relationship of authority, and they may act accordingly. But, at other times, I may send out verbal and nonverbal cues that suggest a relationship of communality and friendship. This may confuse them, resulting a less certainty about the appropriate form of conduct.

As PNL point out, relationship ambiguity of this sort can lead to aversive emotions, such as the feeling of “awkwardness”, as well as undesirable outcomes. This arises most commonly when there is a mismatch between the two parties regarding the type of relationship they are in. To quote:

When relationships are ambiguous, a divergent understanding between the parties can lead to the aversive emotion we call “awkwardness”. There are awkward moments in a workplace or university when an underling or student makes a transition from a subordinate (dominance) to something closer to a friend (communality). Good friends (communality) are advised not to engage in a business transaction (reciprocity), like the sale of a car or a house, which can endanger the friendship. The ambiguity between dominance and sex (a kind of communal relationship) is the battleground of sexual harassment conflicts… 
(PNL, p. 836)

These last two examples — the friends falling out over the sale of a car, and the work colleagues involved in a sexual harassment suit — indicate just how serious the problem of relationship ambiguity can be. Is it possible that indirect speech plays a role in minimising this problem?


3. Plausible Deniability in Relationship Negotiation
This is what PNL argue. Turning first to something called “politeness theory”, they point out that language serves at least two functions (probably more). The first is to state a proposition such as “it is now raining” or “I will pick up my dry-cleaning”; the second is to negotiate and maintain relationships with others. Indirect speech is particularly important in performing this second function because it allows people to straddle the boundaries between two types of relationship, without giving rise to awkwardness or undesirable consequences.

To see this, go back for a moment to the Briber’s Dilemma from part one. In this scenario, the key problem from briber’s perspective was his lack of certainty as to cop’s values. If the cop was dishonest, he could safely offer an overt bribe, but if the cop was honest, he could not. The indirect bribe allowed him to make a pitch for the benefits of the overt bribe, without incurring its negative consequences. If the dishonest cop understood the implication of what the briber was saying, he would take the bribe. But even if the honest cop had suspicions about the implication, he would not act upon it because he knew it would be difficult to prove his case in court. The indirect bribe provides plausible deniability.

The lack of certainty as to the cop’s values has an analogue in cases relationship ambiguity, namely: lack of certainty as to which relationship-type we happen to be in. If I think I stand in a relationship of authority to my students, and they think the same thing, then I can safely adopt the conversational idioms of an authority figure without offending anyone. But if our views diverge, things become trickier. Indirect speech becomes our saviour. Instead of demanding things directly from students, I can issue polite requests (“If you could find time to do the advance reading, that would be awesome”) which allows them to take the hint, without giving rise to awkwardness.


4. Conclusion
To sum up, the plausible deniability of indirect speech is a valuable commodity, not only in the quasi-legal setting of the Briber’s Dilemma, but also in the everyday art of relationship negotiation. Indirect speech allows conversational parties to avoid unpleasant outcomes arising from divergent views about the type of relationship they happen to be in. This gives one good reason for the prevalence of indirect speech.

But surely something is still missing? As PNL highlight in their paper, many indirect locutions — including the classic “Would you like to come up and see my etchings?” — do not seem to involve much in the way of plausible deniability. After all, everyone knows what the implied meaning of that offer really is. Explaining the prevalence of these blatant indirect locutions is a task we take up in the last part of this series.

Friday, March 22, 2013

The Logic of Indirect Speech (Part One)

.... I was just thinking we could take care of it right here, in Brainerd.

I’ve been a fan of Steven Pinker’s work for some time. And although I like to think I never fall into the cognitive trap of hero worship, I do think his books and articles bristle with fascinating, thought-provoking and oftentimes controversial ideas. I also think he is among the best science communicators, writing in a clear and accessible style, without talking down to the reader. I guess this isn’t too surprising given that his next book is going to be about style in written communication.

Pinker’s primary academic specialism is in the field of psycholinguistics. Roughly, this is the study of the psychological (and perhaps neurological) mechanisms underlying the human ability to use and comprehend language. As part of his work in this field, Pinker (along with some co-authors such as James Lee and Martin Nowak) has been exploring the logic of indirect speech.

Indirect speech is a significant psychological and cultural phenomenon. An indirect speech act is one in which a speaker does not say what he or she really means. Instead, they rely on commonly recognised implied meanings to communicate their true intentions. Pinker asks the obvious question: why do people do this? Especially when it seems like it would be easier and more efficient to use direct speech.

As it happens, the phenomenon of indirect speech is something that interests me too. I maintain an active interest in linguistic theory and the law. One thing that has emerged from this is a concern with the role of implicature in the interpretation of legal texts. In other words, I wonder to what extent implied meanings can (and more importantly should) affect the interpretation of the law.

Because of this, I’m interested in Pinker’s work on the logic of indirect speech. To be sure, his goals are quite different from mine. Whereas I’m interested in the normativity of indirect speech, he is interested in developing an explanatory theory of indirect speech. Still, these interests can overlap so I thought it would be worth taking a look at some aspects of his theory of indirect speech. This will make a nice break for me from the heavily ethical and normative work that I usually examine on this blog.

So, up for consideration over the next few blog posts will be Pinker, Nowak and Lee’s (hereinafter “PNL”) paper “The Logic of Indirect Speech”, which appeared in the 2008 edition of PNAS. This piece is almost entirely theoretical in nature. Pinker and Lee have followed it up more recently with an experimental study that finds some support for their theory, but I won’t be looking at that here. I’ll focus purely on the theory.

This theory is broken into three parts. Part one examines the link between indirect speech and plausible deniability. Part two looks at the role of indirect speech in relationship negotiation. And part three tries to explain why indirect speech is used even when the implied meaning is obvious (and thus the speech no longer even seems indirect). I’ll look at part one today, leaving the other two parts to subsequent posts.


1. The Briber’s Dilemma
The first thing to note about indirect speech is that it throws up something of a paradox. To use one of Pinker’s favourite examples, the classic line “Would you like to come up and see my etchings?” is widely (though not universally) recognised as an indirect sexual request. If translated into direct speech it might read something like “Would you like to have sex with me (or engage in some cognate activity)?”. The question is why do people use the indirect locution when the direct locution is available to them. After all, the direct locution is more efficient and less vulnerable to being misunderstood. If you really want to have sex with someone why not just come out and say so?

Now, clearly we all see the problem here. We know, instinctively, that there are many advantages to the indirect locution. The challenge is to explain exactly what those advantages are, and why the use of indirect speech, far from being paradoxical, is actually a highly rational act. To do this, PNL introduce us to a game theory-style conundrum* involving a speeding driver (“speedster”) and a traffic “cop” (to use the crude American term). The game goes through two iterations, one in which indirect speech is not an option, and the other in which it is.

In the first iteration, we must imagine that the speedster has been pulled over to the side of the road and is about to be issued a speeding ticket. The speedster has two options: (i) say nothing; or (ii) offer an overt bribe (e.g. “If you scrap the ticket, I’ll give you 50 bucks!”). These two choices lead to different outcomes, depending on the state of the world. In this case, the “state of the world” refers to the values of the cop. If the cop is honest, then he won’t accept a bribe. If he’s dishonest, he will accept a bribe. This creates a significant decision problem from the perspective of the speedster. If the cop is honest and the speedster offers a bribe, then he could be arrested for bribery. This would be a worse outcome than getting a speeding ticket. But if the cop is dishonest and the speedster offers a bribe, he could avoid the ticket. What should he do?

I’m going to represent this problem on a decision matrix. To make it more concrete, I’m going to assume that there is a 50-50 chance that the cop is dishonest, and that costs to the speedster are as follows:

Cost of Bribe = $50
Cost of Ticket = $100
Cost of Arrest for Bribery = $1000

That gives us the following payoff matrix:



As you can see, given the probability of dishonesty and the costs of the various outcomes to the speedster, the most rational choice from their perspective is to say nothing. The expected cost of saying nothing is $100, whereas the expected cost of an overt bribe is $525.

Now let’s run through a second version of the game. This time instead of having only two options, the speedster has three options: (i) say nothing; (ii) offer an overt bribe; or (iii) implicate a bribe (e.g. “could I make a donation to the policeman’s benevolent trust?”). The inclusion of this third option changes the nature of the problem. This time round it becomes important to know how the dishonest and honest cops will react to the implicit bribe. For ease of analysis, I’m going to assume that the dishonest cop will automatically recognise it as a bribe, and readily accept it. Contrariwise, the honest cop may or may not recognise it as a bribe, but in either event won’t act upon it. In other words, he’ll just proceed to issue the speeding ticket.

Assuming that the costs and probabilities are the same as the previous version, we get the following payoff matrix:




As you can see, this time around the rational choice is rather different. The implicit bribe is the option that yields the lowest expected cost ($75).

This gives us a partial explanation for the persistent and prevalent use of indirect speech. It shows how, in certain contexts, indirect speech is the rational option. PNL argue this is because indirect speech provides us with a valuable commodity: plausible deniability.


2. The Logic of Plausible Deniability
To see the value of plausible deniability, we need to confront some of the limitations in the preceding model. As I have presented it, the Briber’s Dilemma ignores several significant variables. Let’s work our way through each of them now.

For starters, it ignores the fact that recognition of the implied meaning of the indirect bribe will not be perfect. In other words, the model I have presented assumes that the dishonest cop will be able to recognise the implied meaning with probability 1; and that, likewise, there is no chance that the honest cop will act on the implied meaning and arrest for bribery. This needs to be corrected in any proper model of the scenario.

In a somewhat similar vein, the model sketched above assumes that indirectness is a discontinuous property. In other words, it assumes that there is one, and only one, level of indirectness that can be present in an implied bribe. In reality, the indirectness of the bribe is likely to vary along a graded scale. One can try to imply a bribe with an extremely indirect locution (e.g. “it’s a lovely day isn’t it officer, it’s shame to have these unpleasantries”), or slightly less indirect bribe (e.g. “I was hoping we could sort it out here in Brainerd”), all the way up to a direct bribe. Picking the appropriate level of directness is something the rational briber will need to keep in mind.

Finally, and perhaps most importantly, the model fails to highlight the significance of the directness threshold and the role it plays in the decision-making of the dishonest and honest cop. To put it bluntly, in order for an indirect bribe to be rational, it must be the case that the dishonest cop accepts the bribe at a lower level of directness than the honest cop requires before making an arrest for bribery. If the level of directness is the same for both, indirect bribing cannot be a rational strategy.

For sake of argument, imagine that the locution used to imply a bribe is “could I make a donation to the policeman’s benevolent trust?”. Suppose that for the dishonest cop this implies a bribe, one that he is all too ready to accept. But to the honest cop, things are less certain. He kinda-sorta thinks it might be a bribe, but he’s not sure enough to make an arrest for bribery. After all, the level of proof at trial is pretty high, and if there’s any reasonable doubt as to what the speedster really meant, a conviction will not be secured. If this is the case, indirect bribing is the rational strategy. On the other hand, if the honest cop is willing to make an arrest at the same level of directness as the dishonest cop requires to recognise a bribe, indirect bribing is not rational.

PNL’s model of rational bribing takes into account these three factors. In doing so, it makes a better case for the rationality of indirect speech. As they frame it, it is the fact that the dishonest cop is willing to act on a different level of directness than the honest cop that gives rise to the phenomenon of plausible deniability. And it is this phenomenon of plausible deniability that makes indirect bribing rational.


3. Conclusion: Expanding the Explanation
To summarise, PNL are concerned with the paradox of indirect speech. That is, speech that does not say what its speaker actually means. This kind of speech is less efficient, and more prone to being misunderstood than direct speech. So why is it so common? PNL suggest that part of the answer lies in the phenomenon of plausible deniability. In certain social contexts, an indirect locution yields a higher payoff (or lower cost) than a direct locution. This is because the implied meaning of the indirect locution is plausibly deniable, and so blocks a higher cost outcome. This is evinced by the Briber’s Dilemma.

But this is, at best, a partial explanation for the prevalence of indirect speech. The role of plausible deniability is easily appreciated in a case like that of the Briber’s Dilemma because in a case like that, the costs and benefits of indirect speech are quite obvious. The problem is that indirect speech pervades innumerable social contexts, including contexts in which the costs and benefits are barely noticeable. Why is this? To answer that question, we need to appreciate the role of indirect speech in relationship negotiation. We’ll consider that the next day.

*To me, this is just a decision problem, not a game theory problem because it doesn’t consider the payoffs from the perspective of the traffic cop. Still, PNL refer to it as “game theoretic conundrum”.

Thursday, March 21, 2013

Should pornography be considered "speech"? (Part Two)



(Part One)

This is the second part in my series looking at pornography and the free speech principle. The series is focusing on the arguments analysed in Andrew Koppelman’s article “Is Pornography “Speech”?”. In part one, we looked at Frederick Schauer’s argument. In this post, we will look at John Finnis’s one. Both authors suggest that pornography is not covered by the FSP.

A brief recap is in order first. Koppelman claims that both Schauer and Finnis develop their arguments within a common argument-pattern. I sketched out that argument-pattern in part one. It started with the basic free speech principle (FSP):

Free Speech Principle: (Some or all types of) speech ought to be free from government censorship and regulation

The FSP is a very common trope (well, more than that) in political debates. The idea that some (perhaps all) categories of speech must be protected from government interference is a powerful one, and a variety of arguments can be amassed in its favour.

The key move made by the likes of Schauer and Finnis is to claim that in order for speech to be covered by the FSP, it must do certain things. Specifically, it must contain ideas or engage the faculty of reason. Thus, both authors hold that the following Coverage Principle can be derived from the FSP:

Coverage Principle: The FSP only covers speech that communicates ideas and/or engages the faculty of reason.

It is this coverage principle that forms the basis of the common argument pattern, which runs as follows:


  • (1) The FSP only covers speech that communicates ideas and/or engages the faculty of reason. (Coverage Principle)
  • (2) Pornography neither communicates ideas nor engages the faculty of reason. (Factual claim)
  • (3) Therefore, pornography is not covered by the FSP.


In the remainder of this post we’ll look at how John Finnis defends this argument-pattern. It should perhaps be noted that the Finnis article which is the subject of discussion here dates from 1967 (LINK). I’m not sure if Finnis has changed his views on the matter since then, but I doubt they would have changed greatly given the general character of his beliefs on matters of sexual morality.


1. Finnis on Reason and Passion in Symbolic Communication
The first thing we need to do is to see how Finnis defends the key premises of the argument-pattern, starting with premise (1). As you will recall from the previous post, premise one only makes sense in light of the more general defence of the FSP. More precisely, premise (1) needs to be derived from the general defence of the FSP. It is only once we know why the defence of free speech is important that we can know what kinds of communications or artistic productions fall within the remit of the FSP.

So why does Finnis think free speech is important? I’m going purely by the representation of his views in Koppelman’s article, and according to that representation Finnis offers a fairly typical instrumentalist defence of the role of free speech in a democratic. Quoting from James Madison, Finnis comments that one of the main roles of government is to aim at some control over people’s passions. That is to say, to prevent them acting on a purely emotional basis; and to allow reason to rein in the excessive and deleterious manifestations of the passions.

Building upon this, Finnis claims that value of free speech lies in its ability to create independent rational (not emotional) critics of the government. Thus, any expressions that engage the faculty of reason, in preference to the emotions and passions, deserves protection from government interference. For it is those kinds of expressions we need in order to keep the peace and to keep the government honest.

Finnis thinks this allows him to derive the coverage principle because:

“…to the extent that expressions derive from the passion end of the reason-passion continuum, the rationale for that freedom [i.e. freedom of speech] disappears.

This comment might seem a little cryptic, but file it away in your mind for now, paying particular attention to its reference to the “reason-passion continuum”, because we’ll be coming back to it later.

One thing that is clear from Finnis’s discussion is that he does not think this defence of the FSP only covers so-called “political speech”. On the contrary, Finnis thinks there is a robust and defensible space for purely artistic forms of expression under the FSP. He just argues that art always expresses emotions in a detached, reason-enriching way. In other words, artistic expression doesn’t merely provoke or enrage; it symbolically represents emotions thereby enabling us to critically reflect upon them. This view of artistic expression is what allows him to defend premise (1) without falling into the trap of excluding all non-political forms of speech from the FSP.

Granting this, we proceed to consider premise (2). As you might imagine, Finnis’s claim is that pornography falls outside the scope of the FSP because its “expressions” engage the passions, without providing space for the detached critical reflection in other forms of art. Perhaps surprisingly, Finnis thinks that erotic feelings can legitimately be symbolised in works of art. He merely argues that in pornography the feelings themselves swamp or overwhelm their symbolisation.

Finnis envisions the stereotypical consumer of pornography as being one who desires a state of emotional arousal (most commonly orgasm) to the exclusion of any high-level aesthetic reflection. Somewhat amusingly, Koppelman gives this stereotypical consumer a name and highlights the role he (or she!) plays in the argument:

Grimly Purposive Masturbator (GPM): The stereotypical consumer of pornography is one who wants to achieve orgasm and does not care how he achieves it. The pornography is a mere aid to their sexual satisfaction.

I must confess to being a little puzzled as to how discussion of the GPM helps us to defend premise (2). But, being charitable, I guess the claim could be that it is the fact that the GPM is the typical consumer of pornography that provides evidence for premise (2). Or, it could be that it is the fact that the producer of pornography has the GPM in mind when he creates pornography that provides evidence for premise (2). Bringing these two observations together we can say that: if the GPM is both the intended user and the typical user we have good evidence to support the claim that pornography does not engage the faculty of reason.


2. Evaluating Finnis’s Argument
So what should we make of all this? Let’s look at the defence of both premises in turn. The first thing I would say is that the instrumentalist, Madisonian defence of the FSP isn’t entirely persuasive to me. While speech definitely has a role to play in ensuring an intelligent and critical electorate, I think it has other, intrinsic and instrumental benefits too.

For starters, it is the medium through which knowledge and culture are shared, and through which important actions are performed. Not all of these things are primarily reason-based, but they are valuable nonetheless. Separate from this, I see no reason to think that certain kinds of emotions should not be valued and protected by society. Indeed, and on this very point, Andrew Altman has argued that there might be a “right to be turned on”, a right that could be protected through pornographic speech.

In addition to this, I think one has to bear in mind Schauer’s sceptical defence of the FSP from part one. As you’ll recall, one classic defence of the FSP is that the failure of governments to exercise powers of censorship in a sensible and non-abusive manner in the past gives us good reason not to hand such powers over to them again. If we accept this rationale for the FSP, we should be pretty sceptical about the government’s ability to reliably distinguish the materials that primarily engage the passions from those that primarily engage reason. Thus, the underlying defence of the FSP seems dubious to me.

But even if I accepted the Madisonian line, thought that only speech which affected the capacity for reason could be covered by the FSP, and that the government could reliably distinguish such materials, I would be pretty sceptical of Finnis’s argument. The problem is that I simply don’t buy his notion of a reason-passion continuum. To me, this word “continuum” creates the impression that there can be forms of communication that fail to impact upon the emotions, and vice versa. This doesn’t seem to be the case to me. Reason and passion are intimately bound together. Every interpretation and every act of attention involves some input from the passions.

This, of course, brings us to premise (2) and the claim that pornographic speech overwhelms the faculty of reason. This also seems pretty weak to me. In line with what I just said, I think every form of expression can engage the faculty of reason. Certainly, there’s no reason to think that pornography cannot do this. This is something that several feminist critics of pornography seem to accept. As discussed previously on this blog, MacKinnon (among others) argues that pornography involves some extremely powerful, and politically important, symbolic encoding. Specifically, it encodes the permissibility conditions for heterosexual relations. The very fact that these kinds of arguments can be made illustrates that aesthetic reflection about the nature and value of the symbolism in pornography is possible. This runs contrary to Finnis’s claim. Furthermore, as Koppelman point out, many great works of art play to and encourage erotic responses (he mentions the play Equus as an example). But this doesn’t preclude aesthetic reflection.

These two responses to the defence of premise (2) highlight perhaps the main flaw in the whole argument. Namely, that the focus is misplaced. The fact the likely consumer of pornography (the GPM), or the likely producer of pornography, have certain intentions and goals in mind, does not support the claim that pornography cannot engage the faculty of reason. If there is one lesson from the arts and humanities, it is that all human artifacts are fit objects for deep aesthetic reflection. Pornography is no different in this regard.


3. Conclusion
To sum up, neither Finnis nor Schauer’s arguments are persuasive. In the first instance, their defences of the FSP are either dubious or fail to provide support for the coverage principle. In the second instance, their claims that pornography fails to communicate ideas or engage the faculty of reason are weak. Pornography can do both of these things, irrespective of the characteristics of the typical user.

None of this, however, should be taken to imply that pornography is a valuable cultural product, worthy of respect and protection by our laws. There may be other reasons to object to or restrict pornographic speech. It’s just that the claim that it is not “speech” or does not fall under the auspices of the FSP is not one of them.

Sunday, March 17, 2013

Should pornography be considered "speech"? (Part One)



This post considers whether or not pornography should be covered by the free speech principle (FSP). According to this principle, all (or most) forms of speech should be free from government censorship and regulation. But this raises the question: which types of symbolic productions are covered by the FSP? And is pornography among them?

This isn’t the first time I’ve talked about pornography and free speech. Back in December (2012 for those of you reading this in the distant future), I looked at an argument claiming that pornography qua speech act subjugated and silenced women. That argument interested me largely for its unique deployment of speech act theory, which is among my research interests. The arguments I look at in this post and the next hold no particular fascination for me. I just happened to read a paper one day and thought it might be worth blogging about.

The paper in question was Andrew Koppelman’s “Is Pornography “Speech”?”, which appeared in the journal Legal Theory several years ago. In it, Koppelman looks first at a general argument pattern claiming that pornography is not covered by the FSP, and then at two specific instantiations of that pattern. I’ll follow suit in this post. Ultimately, Koppelman disagrees with these arguments. I think I do too.

A couple of caveats at the outset. For those of you who have read Koppelman’s paper — or who feel inclined to do so after reading this post — be aware that I take some liberties with his original discussion. To be precise, I wasn’t too impressed by the way in which he laid out the general argument pattern, nor with how he explained its connection to the two specific versions that he critiques. I try to improve upon this aspect of his discussion here.

In addition to this interpretive caveat, there is a second, arguably more important, caveat. This has to do with the limited nature of the arguments being discussed. The focus is purely on whether pornography should be covered (definition below) by the FSP. There is no general argument for or against pornography on display. Certainly, if one reaches the conclusion that pornography is covered by the FSP, it becomes more difficult to argue in favour of its regulation or censorship, but that is a long way from saying that it is impossible to do so. Similarly, although concluding that pornography is not covered by the FSP may make it easier to argue for is restriction, it does not automatically lead to that conclusion.

This should be borne in mind at all times.


1. The General Argument Pattern
Koppelman says that there is a general argument pattern, used by several theorists, which holds that pornography is not covered by the FSP because it does not involve the communication of ideas. I don’t like Koppelman’s formulation of this argument pattern, so I’m going to lay out my own version of it.

I start with the free speech principle itself, which underlies all aspects of this debate. I define that principle as follows:

Free Speech Principle: (Some or all types of)* speech ought to be free from government censorship and regulation 

I appreciate that this definition is far from perfect. There are so many views about what should and should not be included within the freedom of speech that it would be difficult to come up with a definition that was agreeable to all. Indeed, I’m not even sure that I agree with this formulation. For example, while I think any version of the FSP worth talking about would hold that government censorship of speech was off limits, I’m much less sure about regulation. I think there might be certain forms of regulation (or regulation in certain contexts) that is permissible.

As it happens, those worries are not too important for the discussion here. What is important here is to keep in mind two facts. First, the FSP can be defended with a variety of different arguments. Typically, these arguments point to the value of free speech from an individual or societal perspective. And second, those who argue that pornography should not be covered by the FSP think that the arguments they use to defend the FSP entail the following Coverage Principle:

Coverage Principle: The FSP only covers speech that communicates ideas and/or engages the faculty of reason.

How a defence of the FSP could entail this principle is a matter of some dispute. Indeed, it is ultimately the major bone of contention between Koppelman and the set of views he challenges. Still, I think it is clear that those in opposition do hold to this entailment relation. (Note: this is where I differ from Koppelman in setting out the argument pattern. He doesn’t explicitly define these two principles, or identify how it is that the purported defences of the former supposedly entails the latter.)

What is meant by the use of the word “covers” in this principle? Roughly, a type of speech is covered by the FSP if there is a prima facie case for its non-regulation and non-censorship. This is to be contrasted with the notion of “protection”. A type of speech is protected by the FSP if there is more than a prima facie case for its non-regulation and non-censorship. The question of coverage by the FSP is a threshold question: it identifies possible candidates for protection. The critics of pornography being considered here think that pornographic speech fails to meet this threshold.

That brings us, at last, to the argument pattern:


  • (1) The FSP only covers speech that communicates ideas and/or engages the faculty of reason. (Coverage Principle)
  • (2) Pornography neither communicates ideas nor engages the faculty of reason. (Factual claim)
  • (3) Therefore, pornography is not covered by the FSP.


Our job is to fill in this general pattern with the views of some real life philosophers and legal theorists, lest we be accused of constructing a straw man. This we shall do by looking at the arguments of Frederick Schauer and John Finnis. I’ll only look at Schauer’s argument in the remainder of this post. Finnis will be covered in part two.


2. Schauer’s Argument
Frederick Schauer is a legal philosopher whose views I have considered before on this blog. I know him primarily from his writings on evidence law and legal positivism. I must say I knew nothing about his work on free speech before reading Koppelman’s piece. Then again, my ignorance is boundless. Let’s see how Schauer would defend the premises of the argument outlined above.

To start with, we need to consider how he defends the FSP. As I said, it is the defence of that principle that supposedly entails the Coverage Principle. Interestingly, Schauer is somewhat critical of most defences of the FSP. Nevertheless, he thinks that are two plausible ways to defend it:

Autonomy of thought: We each have a right to form our own beliefs, to decide for ourselves what the objects of our thought should be. This entails a right to receive information, unhindered by government interference.
Scepticism of Government: Grant for the sake of argument that some speech can be permissibly restricted, even then we should be deeply sceptical about the ability of governments to reliably and accurately identify this type of speech. History suggests that governments are prone to abuse such a power.

I have a great deal of sympathy with both arguments, but the question is whether they support the Coverage Principle. I’m less sympathetic in this regard.

The scepticism argument would certainly seem to be an uncomfortable bedfellow of the Coverage Principle. Even assuming that there is a type of speech that fails to communicate ideas — an assumption we shall challenge in part two — why should we trust the government to reliably distinguish that type of material from that which does in fact communicate important ideas. James Joyce’s Ulysses is undeniably a great and important work of art. Nevertheless, it was censored for obscenity in its day. Governments could make similar categorisation mistakes in the future.

On the other hand, the autonomy of thought argument may support something like the Coverage Principle, though a lot would turn on how we understood the necessary conditions for independent thought. The problem, as I see it, is that autonomy of thought probably entails other rights too, (the right to education maybe?) and whose to say that those rights wouldn’t support the non-regulation of pornography.

This is a little too speculative, however, so let’s just assume that the autonomy of thought argument works. Then we run into the factual issue: does pornographic speech fail to communicate ideas? Schauer suggests that a certain type of pornography might:

Imagine a motion picture of ten minutes’ duration whose entire content consists of a close-up colour depiction of the sexual organs of a male and a female who are engaged in sexual intercourse…The film is shown to paying customers who, observing the film either reach orgasm instantly or are led to masturbate while the film is being shown. 
(Schauer, Free Speech: A Philosophical Enquiry, 1982, p. 181)

I feel sorry for these customers who “reach orgasm instantly” but leaving that to the side, Schauer’s point is that this kind of pornography communicates no ideas. It is simply a sexual surrogate, like a vibrating toy or a blow-up doll, designed purely to get people off, not to make them think. The argument would appear to be something like this:


  • (4) Pornographic speech is purely a sexual surrogate, i.e. a something that provokes or leads to orgasm.
  • (5) If pornographic speech is purely a sexual surrogate it cannot communicate ideas.
  • (6) Therefore, pornographic speech does not communicate ideas.


There are problems with both premises of this argument. As Koppelman points out, premise (4) confuses the effect of pornographic speech on some people with its content and form. This is a point we return to in part two. Beyond that, it’s not clear that premise (5) is really true. In order to produce a sexual response, pornography does need to engage the mind. As Koppelman puts it, pornography provides a kind of fantasy that is decoded by the mind and the body to produce the sexual response. It gives us the idea of a certain kind of sexual activity, and that can be used to produce an orgasm (if the viewer/consumer so wishes), but it doesn’t do so with the intermediating effects on the faculty of reason.

So it’s not clear that Schauer’s argument works.

This seems like a good place to end the discussion for today. In the next post we’ll look at Finnis’s argument, which is slightly more complex in how it develops and defends the Coverage Principle.

*The bracketed phrase needs to be included to make sense of the views of those in favour of restriction and regulation.

Saturday, March 16, 2013

Nagel on the Burden of Enhancement (Part Two)



(Part One)

It has oft been observed that people are uneasy about the prospect of advanced enhancement technologies. But what is the cause of this unease? Is there any rational basis to it? I’m currently trying to work my way through a variety of arguments to this effect. At the moment, I’m looking at Saskia Nagel’s article “Too Much of Good Thing? Enhancement and the Burden of self determination”, which appeared a couple of years back in the journal Neuroethics.

In this article, Nagel presents two main arguments. The first — which I call the Well-Being Argument (WBA) — suggests that there might be an inverse relationship between the level of choice and the level of well-being, and since enhancement increases the amount of choice it could have a negative impact on well-being. The second — which I call the Social Responsibility Argument (SRA) — suggests that the growth of enhancement technologies may lead to increased social pressure to enhance, which has negative implications from the perspective of both social equality and well-being.

I looked at the WBA in part one and questioned some of its premises. In the remainder of this post, I do the same for the SRA.


2. The Social Responsibility Argument
Let me start with a story.

Jake is an epileptic. His condition is serious, but manageable. Although he is extremely prone to seizures, if he takes his prescribed medication he can usually avoid them. Suppose Jake is about to undertake a long journey by car. He knows that if he doesn’t take his medication in advance of the trip, there is a high probability that he will have a seizure and that this may lead to an accident. Despite this, he decides not to take his medication, thinking it will be too much hassle to get it out of his bag. So he drives off. Two hours into the journey he has seizure, which causes him to swerve the car off the road at which point he collides with a pedestrian. The pedestrian dies.

Is Jake responsible for the pedestrian’s death? I don’t know how you feel about it, but my feeling is that he is. His awareness of the risk posed by his condition, coupled with the fact that he deliberately ran that risk, seem sufficient to make him responsible for the pedestrian’s death (deterministic considerations being ignored for present purposes). But now contrast the judgment about this case with another case. Suppose this time Jake has the same condition, but there is no medication capable of controlling it. Suppose, further, that he is forced to take control of a car when the driver has a heart attack. Nevertheless, he still has a seizure and causes a fatal accident. Is he responsible this time? I don’t think so, but what’s the salient difference between the cases? Well, in the former, Jake has both the means and the opportunity to avoid (or, at least, minimise the risk of) the accident. Thus, he has the medication, and he has the decision-point prior to entering the car. In the latter case, he has neither of these things. There is no controlling medication, and he has to take control of the car in an emergency.

By suggesting that the growth of enhancement technologies creates more and more scenarios like that of the first Jake case, Nagel uses the difference in attitudes between it and the second case to fashion an argument against enhancement. The argument has two stages, the first of which looks something like this:


  • (1) If we have both the means and the opportunity to avoid a particular kind of social risk, then we are likely to be held responsible for the realisation of that risk (if and when it arises). 
  • (2) The growth of enhancement technologies increases both the means and the opportunities we have for avoiding social risk. 
  • (3) Therefore, the growth of enhancement technologies increases the number of occasions on which we are likely to be held responsible for the realisation of social risk.


By itself this is a pretty interesting argument, and I will say more about its key premises and concepts below, but hopefully you can see why it needs a second stage. The problem is that this first stage is largely descriptive-predictive. It holds that increased enhancement leads to increased responsibility for social risk, but this in itself isn’t a bad thing. Additional premises are needed to turn this into an anti-enhancement argument.

What might those premises look like? Unfortunately, there is no clear indication from Nagel’s article. This is largely my fault, since I am the one insisting on the formal reconstruction of what she has to say. Still, I think Nagel would be keen to push this argument back in the direction of well-being related concerns. In other words, she would like to argue that the increased likelihood of being held responsible creates a social expectation that we will take responsibility for those risks, which in turn imposes burdens and anxieties on the individual. So:


  • (4) An increased likelihood of being held responsible for social risk will lead to increased an social expectation that we will take responsibility for those risks.
  • (5) An increased social expectation for taking responsibility creates anxieties and imposes burdens, both of which are inimical to individual well-being.
  • (6) Therefore, the growth of enhancement technologies is likely to be inimical to individual well-being.


Although this is certainly an argument worth making, I think the first stage of the SRA could be pushed in an alternative, equally worthy direction. That direction has to do social equality. My thinking is that an increased social expectation to use enhancement may lead to the social risk burden being distributed in an unequal way. In other words, those who are most vulnerable and most in need of social protection may face a higher level of expectation. The argument would look like this (the first premise is the same as in the previous argument):


  • (4) An increased likelihood of being held responsible for social risk will lead to an increased social expectation to take responsibility for social risk.
  • (7) This increased social expectation is likely to lead to the social risk burden being distributed in an inegalitarian manner.
  • (8) Therefore, the growth of enhancement technologies is likely to lead to an inegalitarian distribution of the social risk burden.


To the extent that inegalitarian social distributions are deemed morally circumspect, this would also count as an anti-enhancement argument.


2. Assessing the Social Responsibility Argument
The fact that the SRA can take different forms creates difficulties when it comes to its assessment. Ideally, one should cover all the possibilities, but in the interests of time and space I will only cover one. Thus, I will focus on the first stage, which is common to both forms, and the social equality version of the second stage. I justify this on the grounds that I already spoke about well-being in part one.

So let’s look at the first stage of the argument. Are the key premises of this stage supported? Well, premise (1) seems pretty reasonable to me. It can be supported in two ways. First, one could simply rely on thought experiments like the one I outlined involving Jake the epileptic. If there is widespread agreement about responsibility in such cases, then it seems that we are likely to be held responsible in such situations. But in addition to this intuition mongering, we can also point to some actual legal cases in which this seems to be the case. The current law on criminal responsibility (in England and Wales at least) suggests that you can indeed be responsible for the downstream consequences of failing to minimise some social risk. For example, if you continue to drive while feeling drowsy you can be held responsible for causing death by dangerous driving, even if you are actually asleep at the time of the offence. This leads me to believe that premise (1) is sound.

Premise (2) is a bit less clear to me. It may well be that certain forms of enhancement minimise social risk, but it’s not obviously true. The link between enhancement and the minimisation of social risk may be extremely indirect in many cases. Alternatively, enhancement may itself increase social risk. For example, an enhanced memory or imagination may have a debilitating effect on one’s ability to concentrate on a particular task. This could be risky. Still, if we stick with the driving while drowsy example for a minute, I think it is fair to say that a cognitive enhancer like modafinil might minimise the social risk in such a case. Since modafinil keeps you alert and awake for longer periods of time, it is one way of minimising the social risk associated with driving in such conditions.

But it’s not the only thing that will minimise the social risk. Choosing not to drive is another way of doing this. So, in this case, although enhancement may give us the means and opportunity to avoid social risk, it won’t be the only thing doing this, nor even the most effective. This has implications for the argument as a whole as it highlights a gap between the premises and the conclusion. Specifically, increasing the number of means and opportunities for minimising social risk does not, in and of itself, increase the likelihood of our being held responsible in cases where the risk is realised. Why not? Because the judgment of responsibility probably only requires one plausible and accessible way of minimising the risk to have been neglected. That said, an increased number of opportunities for avoiding the risk certainly won’t decrease the likelihood of being held responsible.

That brings us onto the second stage of the argument, the one focusing on the distribution of the social risk burden. Before I begin to assess this stage, it’s worth briefly defining what I mean by “social risk burden”. In every society, there is cooperation between multiple persons. This typically creates a social surplus. Thus, for example, if we cooperate to create a local police service, we improve the security and safety for most people in our community. That improved level of security and safety would be a social surplus. But to attain that social surplus, people have to accept certain duties and responsibilities (burdens). For instance, the police officers need to be paid, so people need to contribute to their payment. If they do not, they risk losing the social surplus. This gives us the definition: the social risk burden is simply the bundle of duties of responsibilities imposed upon people in order to secure the social surplus (or “to minimise the risk of losing that surplus”).

Social surpluses and burdens are distributed across the members of society. Oftentimes this distribution can be unequal. Some people get more of the surplus, some people get more of the burden. This might be morally problematic, depending upon one’s theory of distributive justice. The key claim in the second stage of the argument comes in premise (7). It says that the increased social expectation to take responsibility (that comes with the growth of enhancement technologies) is likely to be inegalitarian. For present purposes, I define “inegalitarian” in terms of strict egalitarianism. In other words, a distribution is inegalitarian if people do not have exactly equal shares of surplus/burden. I know that this is problematic.

So now we come to the crux of it: is there any reason to think that enhancement would result in an inegalitarian distribution of the risk burden? Superficially, yes. Some people create more risks than others. If they are expected to take responsibility for avoiding those risks, and if enhancement does make available means and opportunities for minimising those risks, then we might expect the burden to fall on them. For example, if a surgeon could minimise the risks of surgery going wrong by taking cognitive enhancing drugs, we might impose a burden on them to take those drugs. But since surgeons are involved in more risky activities than most (or, at least, more activities that are themselves intrinsically high risk), a greater burden is likely to fall on them than on the rest of us.

But is this morally problematic? Maybe not. For example, one could argue that surgeons are well-compensated for their high risk work. Thus, when you work out the total distribution of surpluses and burdens, they aren’t really being unfairly treated. Their high remuneration and social status offsets the burden. Still, there are others who might be unfairly treated. Take Jake the epileptic once again. His physical condition makes him a high risk individual. If we impose a burden on him to avoid the risks associated with his condition, he can function as an ordinary citizen and we can get along with our day-to-day lives. But there is no benefit to him over and above the benefits to ordinary citizens. In other words, he gets the burden without the compensating reward. That might be morally problematic.
The saving grace here could be the fact that enhancement technologies create their own compensating rewards. In other words, in addition to minimising certain social risks, enhancement technologies might, to borrow John Harris’s phrase make us “more of everything we want to be”.


3. Conclusion
To sum up, Nagel cautions us about the effects of human enhancement in two ways. First, she argues that by increasing both the level of choice and the number of opportunities for regret, enhancement may impact negatively on our well-being. This was the basis of the WBA, which I examined in part one. Second, she argues that enhancement might increase the social expectation to take responsibility for things. This could impact negatively on our individual well-being, but also on the distribution of social surpluses and burdens. This was the basis of the SRA, which I considered in this post.

I think the SRA, as it pertains to social equality, is an interesting argument. It might be worth pursuing its various implications in more depth. I have sketched some of the potential avenues of discussion above, but certainly not all of them. The problem being posed, as I see it, is that people might be forced to enhance in order to minimise social risks without getting any compensating reward. The question is whether enhancement always creates its own compensating reward. These are things worth thinking about.

Friday, March 8, 2013

Nagel on the Burden of Enhancement (Part One)



A couple of weeks back, I looked at David Owens’s article “Disenchantment”. In this article, Owens argues that the ability to manipulate and control all aspects of human life — which is, arguably, what is promised to us by enhancement technologies — would lead to disenchantment. Those of you who read my analysis of Owens’s article will know that I wasn’t too impressed by his arguments. Since then I’ve been wondering whether there might be a better critique of enhancement, one which touches upon similar themes, but which is defended by more rigorous arguments.

I think I may have found one. Saskia Nagel’s article “Too Much of a Good Thing? Enhancement and the Burden of Self-Determination” presents arguments that are broadly similar to Owens’s, but in a slightly more persuasive way. Like Owens, her concern is with the increased choice that enhancement affords us. Like Owens, she thinks that this increased level of choice may have a disorienting and discomfiting effect on our well being. But unlike Owens, Nagel supports her arguments with copious references to decision theory and psychology. The result is a more satisfying set of arguments.

As I read her, Nagel there two distinct arguments underlying Nagel’s thesis. The first — which I shall call the Well-Being Argument Against Enhancement (WBA for short) — suggests that there might be an inverse relationship between choice and subjective well-being. This runs contrary to the presumptions of certain liberal theories. The second — which I shall call the Social Responsibility Argument Against Enhancement (SRA for short) — highlights the burdens and responsibilities that might be imposed on people as a result of enhancement, which has further negative effects for individual well-being.

In this post, and the next, I want to look at both arguments, considering the structure and the defence of the key premises in some detail. I kick things off in this post by looking at the WBA.


1. The Well-Being Argument Against Enhancement
There is an assumption in liberal thought — perhaps with its roots in the work of John Stuart Mill — that says: the more choice the better. The reasoning behind this is straightforward: An individual’s well-being is determined largely by whether they satisfy their preferences (i.e. whether they get what they want); there is a huge diversity of preferences out there to satisfy; individuals are themselves the best judge of which actions (or combinations thereof) will satisfy those preferences; so by giving them more options, we make it more likely that they will satisfy their preferences, which will in turn increase the overall level of well-being.

Nagel’s first argument — the WBA — takes issue with this line of reasoning. The WBA holds that far from it being the case that more choice is better, more choice may actually be inimical to well-being. It may be true, up to a point, that more choice means increased well-being, but beyond that point there is an inverse relationship between choice and well-being. Indeed, beyond that point choice has a paralysing and disorienting effect on our lives. Enhancement technologies exacerbate this problem by dramatically increasing the number of choices we have to make about our lives.

That would give us the following argument:


  • (1) Although more choice might lead to increased well-being up to a certain point, beyond that point more choice actually reduces the level of individual well-being.
  • (2) Enhancement technologies will increase the number of choices we have to make about our lives beyond the optimal point.
  • (3) Therefore, enhancement technologies will reduce the level of individual well-being.


Before assessing the two premises of this argument, I need to say a word or two about how it relates to what Nagel says in her article. First off, as per usual on this blog, nowhere in her article does Nagel actually present an argument in these formal terms. This is strictly my reconstruction of what she has to say. Second, it’s probably not entirely fair to suggest that Nagel really thinks that enhancement technologies “will reduce the level of individual well-being.” Parts of what she says could be read as supporting that conclusion, but I suspect it’s fairer to say that she thinks enhancement technologies might have this effect. My reconstruction of the argument uses the more definitive expression “will” because that just makes it more interesting and fun to engage with. A cautiously hedged argument is more philosophically and academically acceptable, but it risks setting the burden of proof too low since, of course, enhancement “might” have that effect. The possibilities here are vast. What’s important is what is likely to happen.

With that interpretive caveat out of the way, I can proceed to evaluate the key premises of the WBA, starting with (1).


2. Assessing the Argument
Nagel defends the first premise of her argument with a variety of observations, some based on decision theory and the psychology of decision-making, others based on the psychology of regret. Before looking at those observations, however, we need to be clear about what premise (1) does not say. Premise (1) does not deny the liberal contention that more choice increases well being. What premise (1) is saying is that there is an optimal level of choice, beyond which more choice actually has a negative impact on well-being. Think of an inverse-U (or N) shaped graph, with the number of choices available to an agent displayed along one axis, and the level of well-being experienced by that agent displayed along the other. Although initially the level of well-being increases along with the increase in choice, after a certain point the level of well-being plateaus and eventually decreases. The key practical question for the evaluation of the WBA is: have we reached the plateau already, will enhancement technologies push us over the edge, or will they simply add to the ongoing decline?

Still, this is simply to clarify the nature of the claim being made by premise (1). Is there any reason to think that this claim is true? Here, Nagel relies heavily on the work of Barry Schwartz and his book The Paradox of Choice. Based largely on studies of consumer behaviour, though backed-up by other cognitive psychology studies such as those by Kahneman and Tversky on decision heuristics, Schwartz’s work suggests that more choice increases the level of anxiety among decision-makers. If we imagine that every decision we make can be formally modelled on a decision tree, and that decisions are made by selecting the best branch of the tree, then adding more choices increases the number of branches and increases the complexity of the decision problem that needs to be solved. This leads to the decision-maker getting “stuck” in the decision problem, paralysed by the uncertainty of the various options.

Of course, in reality decision-makers don’t really follow the strict axioms of decision theory. Instead, they adopt heuristics for quick and dirty decision-making under uncertainty. But even then, Schwartz and Nagel argue, adding more choices upsets their usual heuristics, and increases their level of decision-making anxiety. This has a negative impact on well-being for the simple reason that they find it difficult to pick options that actually will have a positive impact.

Regret is an additional factor here. With more choices come more opportunities for regret. We often wonder about what might have been, and I know that in my own life I tend to regret the roads I have not taken, particularly if the roads I have taken don’t seem to be all that great. This can have an obvious and negative impact on well-being. Regret arises after the decision is made, so it’s not directly linked to the decision-paralysis problem, but it is rather a separate reason for thinking that more choice is not necessarily better.

So we have a two-pronged defence of premise one:


  • (4) Increasing the number choices makes the decision problems that need to be solved by each individual decision-maker more and more complex. This can lead to decision-paralysis. 
  • (5) Increasing the number of choices increases the opportunities for regret, and regret impacts negatively on well-being.


These look like plausible reasons to support premise (1), but a couple of cautionary words are in order. First, even my minimal wikipedia-based research into Schwartz’s thesis suggests that the evidence for it is not clearcut. Second, to reiterate the point made above, we need to recognise that even if premise (1) is true, the claim it makes may be pretty modest. I’m perfectly willing to accept that there may be a point at which we have far too many choices to make, but what matters is exactly where the threshold lies. Which brings us to…

…The defence of premise (2). Is there any reason to think that it’s true? Here, Nagel is considerably sketchier than she is in relation to premise (1). She seems to think it obvious that enhancement increases the number of choices we have, and in fact she adopts a very similar thought experiment to that adopted by David Owens when defending his claims about enhancement. She asks to imagine that there are a variety of drugs on the market (“MoodUp, ProSocial, AntiGrief, LiftAnxiety, AgileDayandNight, Prothymos, 7-24-Flourishing”). If effective, such drugs would obviously increase the number of choices we have to make in our daily lives. We would be forced into new decision-problems: Do we want to be perked up or relaxed? Do we want to be outgoing and social, or more introverted and focused?

I enjoy imaginative exercises of this sort, but they fall short of making the case for premise (2). Even if we are willing to tolerate speculation about future enhancement technologies, we still need to know whether the increased number of choices will either (a) push us into the negative part of inverse-U on the well-being/choice graph; or (b) push us further into the negative part of the well-being/choice graph. The only thing that Nagel really says that might support these position is that the kinds of choices we make with enhancement technologies would be qualitatively different from everyday consumer-like choices. Thus, with enhancement we actively choose the kinds of emotional states we wish to have, rather than just the type of cake we want to eat, or car we want to drive.

But again, this doesn’t get us to premise (2). And even if it did I’m not sure that it would be successful. I’ve heard people like Owens and Nagel suggest that the choices one could make with enhancement technologies would be qualitatively distinct from the kinds of choices we currently make, but I’m not convinced. It seems to me like we frequently make choices that regulate and affect our mood, or that change the type of person we want to be. For example, when I go running or cycling, I do so because I know it will improve my mood, and will strengthen certain character traits. The only possible difference with a cognitively altering drug is that the manipulation of character and mood might be more direct and less effortful, but I see no reason to think that this by itself would be enough to increase decision-anxiety and reduce well-being.

So I have to conclude that premise (2) is not defended. This obviously detracts from the force of the argument.



3. Conclusion

To sum up, one of Nagel’s arguments against enhancement is the well-being argument (WBA). According to this, there can be an inverse-relationship between the number of choices we have and our level of well-being. More precisely, if we have lots of choice, we face more complex and difficult decision problems, and more opportunities for regret. Both of these things impact negatively on our well-being. But while this may well be true, there is no good reason (yet) to think that enhancement technologies in and of themselves will increase the level of choice to the point where the relationship between it and well-being is inverted.

That brings us to the end of my discussion of the WBA. In part two, I will look at Nagel’s second argument against enhancement: the social responsibility argument (SRA).

Thursday, March 7, 2013

A New Critique of Penal Substitution (Part Two)



(Part One)

This is the second, and final part, of my brief series of posts on the penal substitution theory (PST) of the atonement. The series is looking at the arguments presented in Brent Kyle’s recent article “Punishing and Atoning: A New Critique of Penal Substitution”.

As noted in part one, the atonement is pretty central to Christian thought. If there is one thing that could unite Christians — and I’m not sure that there is — the belief that Jesus’s life, death and/or resurrection somehow managed to solve (or ameliorate) the problem of sin looks like as good a candidate as any. The PST claims that this happened when Jesus was crucified. For it was through his crucifixion that he received punishment for the sins of mankind. The basic argument for the PST looks something like this:


  • (1) In order to solve the problem of sin it was necessary/optional for God to punish the sinful (normative premise)
  • (2) Through the crucifixion God punished Jesus, and Jesus voluntarily received this punishment on behalf of the sinful (factual premise)
  • (3) Therefore, the problem of sin was resolved through Jesus’s death on the cross (PST)


Kyle’s critique is aimed at the second premise of this argument. He adopts what, in part one, I called the “necessary condition” critique. Proponents of this critique try to locate the necessary conditions for punishment, and argue that these conditions do not hold in the case of Jesus’s crucifixion. There have been many such critiques over the years. In this post, we’ll see which necessary conditions Kyle finds absent in the case of Jesus.

To that end, the remainder of the post is structured as follows. First, I lay out the essential elements of Kyle’s critique. Second, I consider Kyle’s defence of his critique from a variety of objections. And third, I look at how Kyle claim that another theory of the atonement — Anselm’s satisfaction theory — avoids his critique. I think this claim is probably correct, but trivial since Anselm’s theory is itself vulnerable to very similar types of criticism.


1. The Belief-Responsibility Condition
Locating the necessary and sufficient conditions for different concepts is somewhat of a mug’s game. A particular concept — say the concept of “chair” — is put before us; then a necessary condition for the existence of a chair is proposed — e.g. it must have four legs; at which point a counterexample is introduced which defeats that necessary condition — e.g. a stool with only three legs; and so a new set of necessary conditions is proposed; and so on. This process can cycle on indefinitely with ever more precise accounts of the necessary conditions of chairhood being proposed, and ever more complex counterexamples being introduced to defeat those accounts.

I say this is “somewhat of a mug’s game” because I’m not at all convinced that conceptual analysis of this sort is a fruitful exercise. Still, it is the game that Kyle and other like-minded critics of the PST wish to play, and for the time being we need to play along.

Consider the following example:

Smacking the Child: A father comes home from work one day tired and frustrated. He decides to vent his anger on his child by smacking her repeatedly, even though he knows that she did nothing wrong.
Question: Did the father punish the child?

Kyle argues, plausibly enough, that the father did not punish the child. To be sure, the father performed an act that constitutes a widely recognised form of punishment, but for all that he still failed to punish the child. This argument works because of the distinction between the primary and secondary sense of the word “punishment”, which were discussed in part one.

Assuming we accept Kyle’s contention that the father did not really punish the child, the question then becomes “why not?” Which necessary conditions for punishment were not met in this particular example?

As a first pass, Kyle suggests that the most obvious thing that is missing is that the child has not performed any sort of moral transgression. But, he argues, this is not in itself a necessary condition for punishment. Suppose the child had done something wrong earlier in the day, but the father knows nothing about it. Would it then be fair to say that the father punished the child? Kyle suggests that it would not.

So what is necessary for a successful instance of punishment (in the primary sense)? Kyle’s offers the following:

Belief-responsibility condition: In order for S to successfully punish P, S must believe that P is (at least partly) responsible for some offence.

A couple things about this condition. First, note that believing that P is responsible for the offence is not the same thing as saying that P committed the offence. P could, for instance, be an army commander who would have command responsibility for actions performed by his or her troops. This is presumably going to be of some comfort to a defender of the PST since I imagine they don’t think that Jesus committed any sinful acts.

Second, Kyle supports this belief-responsibility condition in an interesting way. Primarily, he supports it by reference to thought experiments such as the “smacking the child” one. But he also supports it by arguing that it is preferable to the stronger “knowledge-responsibility” condition. He does so on the grounds that a knowledge-responsibility condition would not allow for the possibility of punishing the innocent, which he thinks is logically possible.

Now, as it happens, this point is not essential to Kyle’s critique. His argument would work even if the stronger, knowledge-based condition were imposed. Nevertheless, I find it interesting because I think it arguably blurs the lines between the primary and secondary senses of punishment. It seems like one could punish the innocent in the secondary sense, but I’m not sure that one could punish the innocent in the primary. Certainly, punishment of the innocent is a classic moral objection to consequentialist theories of punishment: the very fact that they seem to allow for this possibility is thought to show that they are unacceptable. Perhaps it does the same for any proposed account of the necessary conditions for punishment.


2. Kyle’s Argument
Let’s leave some of these tricky conceptual points to the side for now and focus instead on Kyle’s actual objection to the PST. Taking advantage of the belief-responsibility condition for punishment, Kyle is able to make the following argument against premise (2) of the PST argument.


  • (4) In order S to successfully punish P, S must believe that P was (at least partly) responsible for some offence.
  • (5) God could not possibly have believed that Jesus was responsible for an offence.
  • (6) Therefore, God did not successfully punish Jesus.


Premise (5) is defended largely through a priori claims about the nature of God and Jesus. God, being omniscient, cannot have any false beliefs. Jesus, being perfectly innocent, could not have been responsible for human sin. Therefore, God could not have believed that Jesus was responsible for human sin. In addition to this, there are theological consideration, arising from the doctrine of the trinity, that support premise (5). If God and Jesus are different parts of the same being, then God would have to have believed that he himself was responsible for human sin. This seems absurd (though note Kyle’s argument does not rely on this understanding of the trinity — even if God and Jesus were completely separable the problem of belief would stand). This then is the cornerstone of Kyle’s critique.

Another point Kyle makes in defending this argument is that objections to it typically run foul of the distinction between the primary and secondary senses of “punishment”. Indeed, this is perhaps the main point that Kyle makes in his article. The PST, he contends, only seems plausible because people confuse the two senses of punishment. While it is true to say that Jesus underwent a recognised form of punishment (secondary sense) when he was crucified, it does not thereby follow that he was punished (primary sense). The latter is what is needed for a defence of the PST, not the former.


3. Objections and Replies
Kyle examines three major objections to his argument. The first is what might be called the “transference of responsibility” objection. According to this objection, God could have honestly believed that Jesus was responsible because the guilt of humans was transferred onto Jesus at the time of the crucifixion. Certainly, this notion of transference is central to the PST (it is the “penal substitution theory” after all), but this doesn’t give it a free pass. We still have to ask: is it plausible to say that responsibility was transferred to Jesus?

Kyle suggests that there are two forms of transference that might be at play here: literal and figurative. If there is literal transference, then it is actually true that Jesus is responsible for all human sin at the time of the crucifixion. There are two problems with this. First, it seems conceptually impossible for someone to transfer their responsibility to another. If I break your car window with my errant tee-shot in golf, my friend may choose to pay for the repair on my behalf, but that doesn’t make him responsible for breaking the window. Second, there is a theological problem with supposing that Christ is literally responsible for all human sin, namely: Jesus’s perfect innocence is often singled out as the key trait that makes him a fit subject to atone for human sin. If we accept these replies, then we are left with the possibility of figurative transference, according to which Jesus is punished as a scapegoat or sacrificial lamb. Kyle argues that figurative transference still runs afoul of the belief-responsibility condition. It would be like the father smacking the child when he really wants to smack his boss who he genuinely believes to have done something wrong. The father may use the child as a scapegoat, but he doesn’t honestly believe in that case that the child is responsible.

The second objection to Kyle’s argument has been foreshadowed already (to some extent). It is simply to provide counterexamples to his account of punishment. In other words, to point to cases in which somebody is successfully punished, despite the fact that they were not believed to be responsible for an offence. The strategy here is akin to that in my early example of someone objecting to the four-legged account of chairs by pointing to a three-legged stool. It is an exercise in intuition-mongering: seeing whether intuitions about necessary conditions are really shared. It is possible that some people’s views of punishment do not demand belief in responsibility. For example, consequentialist theories might view responsibility as something that is a throwback to a retributive view, and not really compatible with their theory at all.

Kyle’s response here is to fall back, once more, on the distinction between the two senses of punishment. So, he argues that people only think the father is punishing the child because they conflate the primary and secondary senses of “punishment”. To overcome this problem, he argues for a replacement test which enables us to see whether the intuition about the belief-responsibility condition really is shared by participants to the debate. The replacement test holds that if, for whatever counterexample you propose, you replace a widely-recognised form of punishment with another form of harsh treatment which is equivalently harmful, and your view of the case remains the same, then you are tapping intuitions related to the primary sense of punishment. If not, then you are conflating the two senses. Thus, in the case smacking the child, if you replace the smacking with equivalent harsh treatment (say mocking or ridicule), you must ask yourself: is it still punishment? Kyle seems to suggest that it is not. But I have to say, I’m not convinced that the replacement test helps in this instance. I think someone who originally thought that smacking the child was a genuine case of punishment would still think this after the replacement test. Thus, I’m not sure that I buy the replacement test.

The final response to Kyle’s argument is to consider an alternative necessary condition for punishment, which is intuitively as well-supported as Kyle’s one, but which does away with the belief-responsibility condition. One such account is Boonin’s “reprobative condition” account. According to this, punishment necessarily involves disapproval of the general type of offence for which the punishment is imposed. This condition is compatible with the possibility of someone being punished even if they are not believed to be responsible for an offence. Thus, Christ’s death on the cross could be a genuine instance of punishment because it involves the expressed disapproval of human sin, even if Christ is not himself believed to be responsible for that sin.

On the face of it, this is not a hugely promising objection. Suppose the reprobative condition is necessary for punishment, it does not then follow that the belief-responsibility is not. They both could be. In addition to this, as Kyle argues, the reprobative condition may not be necessary for punishment. Imagine a town in which the mayor is arrested for the possession of marijuana. The local judge is a former beatnik who approves of marijuana use, and always gives out lenient sentences in these cases. Still, holding the mayor to a higher standard, the judge decides to give him a moderate sentence. In this instance, it seems like the mayor is genuinely punished, but there is no disapproval of the type of offence. Thus, the reprobative condition looks to flawed (possibly an easier critique is to say that the reprobative condition leads to moral absurdities, but I’ll ignore that here).

If all this is right, and there are no further objections, Kyle’s critique of the PST looks to be in pretty good shape.


4. Conclusion: The Anselmian Satisfaction Theory

With his critique of the PST in place, Kyle concludes his article by looking at a rival theory of the atonement, the Anselmian Satisfaction theory. I don’t know much about this theory, so I’ll just try to faithfully reproduce what Kyle says. According to him, the Anselmian theory rests on a disjunctive normative claim:

Satisfaction Principle:In order to solve the problem of sin, it was necessary/optional for God to either (a) punish the sinful or (b) receive compensation for sin.

Now, if the preceding critique works, then it is not possible for Jesus’s death to count as punishment. But that leaves open the possibility of Jesus’s death compensating God for sin. Kyle argues that this might be the case. Indeed, Jesus is supposedly the ideal candidate for this. Ordinary humans cannot compensate God for their sin because they already owe him everything. They have nothing left to give. Jesus, on the other hand, is innocent and partially divine. He can compensate God for sin.

The major advantage of the Anselmian view from Kyle’s perspective is that it avoids the problems raised by the belief-responsibility condition. In order to compensate someone, you don’t yourself have to be responsible for any wrongdoing. Although I would certainly concede this point, I’m pretty sure that the Anselmian view runs into lots of other problems. What is the mechanism of compensation here? Why is death needed for compensation? If God and Jesus are the same being, then how can God compensate himself by killing himself? I can’t compensate myself for the injuries I suffer through someone else’s negligence, can I?

It seems the debate would have to continue.