Contracts under Coercion

Should one Keep an Agreement with a Robber?

Abstract: It is nearly universally accepted, both in law and in morals, that agreements entered into under coercion are not binding. or e!ample, you are not obligated to "eep, and the law will not en#orce, a contract you underta"e to save your li#e at gunpoint, no matter how bene#icial the agreement would be to you under the circumstances. $hat is the theory behind this invalidation o# coerced contracts? %ost e!planations see coercion as vitiating the voluntariness o# the agreement, in either a psychological or a moral sense, and conclude that coerced agreements cannot be understood as true e!pressions o# the victim&s will and so should not be en#orceable. 'he main alternative to this is a deterrence approach, according to which coercive contracts should not be en#orced because o# the e##ect on encouraging the use o# coercion. In this Article I begin by arguing that the appeal to involuntariness to e!plain the invalidity o# coerced contracts is not compelling, and that despite the truth o# its empirical claim, the deterrence approach is e(ually problematic. In light o# the di##iculty encountered by traditional accounts to e!plain the invalidity o# coerced contracts, I then revisit the common wisdom that agreements entered into under coercion are nonbinding. As it turns out, a more compelling argument #or the conventional wisdom can be discovered, ironically, through a care#ul study o# the only philosopher who re)ects the conventional wisdom about coerced agreements, namely 'homas *obbes. +sing *obbes& discussion as a point o# departure, I o##er a *obbesian,style argument #or the invalidation o# contracts entered into under coercion. $hile the *obbesian perspective #or which I argue ultimately de#ends the conventional wisdom about the en#orceability

o# coerced contracts, it does so in a way that is less sweeping and more narrowly drawn than traditional theories have proposed.

+pdated: -./0.12 3age 1 o# 94 I. I4'R56+C'I54: '*7 3R5897% Imagine you are )ogging along a deserted street in the evening when an arm hoo"s around your nec" #rom behind and a deep voice growls, :;our money or your li#e.< ;ou #eel something hard press into your bac", which you ta"e to be a gun. ;ou e!plain you were out #or a )og with no money=not even an A'% card. ;ou have plenty o# money in the ban" and are willing to return the ne!t day with >0,111 cash, provided that your assailant releases you now. *e doubts your sincerity, but you e!plain, :'his deal is in my interest. A#ter all, the alternative is considerably worse #or me. I# you believe me you will release me, and i# you do not, you will "ill me, assuming that you are rational, since no murder to your credit and a good deal o# cash is superior to no money and a high li"elihood o# arrest #or a serious #elony.< 'here#ore, I have every reason to convince you o# my sincerity to abide by the agreement i# there is any way #or me to do this. ;our assailant considers this argument. :So you see,< you continue,< :we will both be better o## and neither o# us worse o## under the deal I propose than i# you proceed to "ill me now.< Should the robber believe you? Assume #or the moment that the robber is rational, that you are rational, and that each o# you has "nowledge o# the other&s rationality. In this case, it seems unli"ely you will be able to convince the robber that you regard it as in your interest to "eep your promise i# it is not. 'he :common "nowledge o# rationality< that holds between you e##ectively ma"es +pdated: -./0.12 3age 2 o# 94 blu##ing promises highly unli"ely to succeed. rom this it #ollows that you will be able to convince the robber you will actually return with the money only i# it is, indeed, rational #or you to return with the money. So the (uestion now is:

would it be? It would be possible to make it rational #or you to return with the money i# there were some sort o# penalty you could set up to ensure your #uture per#ormance. I#, #or e!ample, he had some way o# #ollowing up should you #ail to appear the ne!t day, matters would be di##erent. 8ut alas, he cannot veri#y any in#ormation you might give him about yoursel#, as you have le#t all personal identi#ication at home. 'o your mutual #rustration, the situation seems hopeless. 8ut #inally an idea dawns on you which you (uic"ly propose to your assailant: you o##er to write a contract in which you will commit to returning with the promised sum the ne!t day in e!change #or your immediate release. ;ou e!plain that you are a pro#essor o# contract law and are reasonably adept at dra#ting such documents, and you assure him that you will spare no pains to ma"e sure the contract is correctly written. Satis#action o# the statue o# #rauds, unambiguous terms, a li(uidated damages provision, and a binding arbitration clause=nothing will be le#t out? 8ut to your dismay, your assailant turns out to be none other than . . . a law student? And he now points out that none o# these measures could possibly help. *e reminds you o# a little impediment to your plan "nown as the duress de#ense,@ which will be a bar to en#orcement o# this "ind o# contract in +pdated: -./0.12 3age 3 o# 94 every court in the land. *e coc"s his gun and prepares to shoot. :$ait?< you shout. :$hat i# we agree to suspend the duress de#ense as part o# our contract?< ;ou e!plain that since in this case both parties would pre#er to eliminate the de#ense, allowing the suspension is surely a winAwin proposition. urthermore, you argue, the duress de#ense is #or the bene#it o# the victim, and here you=the victim=are proposing yoursel# to eliminate it? $ouldn&t any sensible and compassionate )udge honor an agreement between the

parties to suspend the de#ense under these circumstances? ;our assailant, however, is once again s"eptical. Any clause suspending the duress de#ense would itself be written under duress, he points out, given that it is written at the very moment at which he is threatening to shoot. It might have been di##erent i# the agreement to eliminate the duress de#ense had been entered into at a prior time, be#ore you were threatened with #orce. 8ut even then, he says, courts would be reluctant. 'he duress de#ense, he reminds you, is not a mere default rule, but a mandatory rule, / and #or good reason: contracts under coercion are not supposed to be binding under any circumstances, #or whatever reason. Individuals are not allowed to e!empt themselves #rom this rule )ust because there are particular occasions on which the rule wor"s out badly #or them. ;ou are almost certainly out o# luc", he e!plains. As the above discussion ma"es clear, you and your assailant #ace the #ollowing parado!ical situation. ;ou and he would both li"e to enter into an +pdated: -./0.12 3age 4 o# 94 en#orceable agreement in which he relin(uishes his hold on you now, in e!change #or your promise to return the ne!t day with the money. 'he agreement, considered at the present moment, would bene#it both, relative to a noAagreement baseline. 8ut no matter how #ervently you both insist that your agreement is mutually bene#icial, no matter how much you insist that there is a :meeting o# the minds<B between you, no matter how sincere your intentions and how great the loss o# disallowing such agreements, no court in any )urisdiction would en#orce it. %oreover, this result is not the least bit controversial. 'he proposition that contracts entered into under duress are unen#orceable is as solid a proposition o# blac" letter law as ever there was.C 9i"e their legal counterparts, philosophers are similarly unanimous in the view that commitments made under coercion entail no

moral obligation.0 Can the con#idence with which courts and commentators hold this proposition be )usti#ied? I# so, how? II. 7DIS'I4E ACC5+4'S %ost scholars writing on coercion begin by identi#ying a set o# core cases, which they thin" count uncontroversially as instances o# coercion, and then proceed by o##ering some sort o# theoretical account o# them. 3hilosophers, #or the most part, are engaged in a :de#initional< e!ercise: they see" to establish necessary and su##icient conditions #or an act o# coercion, usually a coercive speech act, and continue by #irst elaborating their accounts o# these cases and then +pdated: -./0.12 3age 5 o# 94 applying them to a set o# more marginal cases.F %ost legal scholars, by contrast, do not attempt to supply an account o# the notion o# coercion itsel#, but instead limit their in(uiry to tracing the implications o# identi#ying an act as the product o# coercion.G 'heir interest, then, is in e!plaining why the #act that an agreement or an act is coerced deprives it o# legal validity. 'hey are engaged in what I call an :implicative< account. - In this Article I ta"e mysel# to be ma"ing a contribution to the second body o# literature, with a particular #ocus on coerced agreements, as opposed to criminal o##enses.2 I as" a#resh whether contracts under coercion should be en#orceable, and i# not, why not. I argue that the usual accounts o##ered in the )urisprudential Hi.e. legalI literature o# why coercive contracts are unen#orceable #ail. 8ecause the philosophical and the legal (uestions are not entirely distinct, however, I should ma"e some initial remar"s about the various accounts o# coercion, as well as about the theory o# contract invalidation each may suggest or with which each would naturally be associated. I begin by suggesting that although the common assumption that coerced contracts are invalid is correct, the reasons #or this are #ar less obvious than both philosophical and legal scholars have previously thought. 'he reason #or this is

simple, yet it has been consistently overloo"ed by both philosophers and lawyers: coerced agreements are usually #ully chosen by the parties, and this re#lects the #act that both sides see themselves as advantaged under the agreement at the time they enter into it. or this simple, yet unnoticed reason, it turns out to be much +pdated: -./0.12 3age 6 o# 94 more di##icult to e!plain why coerced agreements are universally held to be invalid than one might suppose. %y aim in this paper is to attempt to highlight the di##iculties o# the tas", as well as to attempt to reconstruct a di##erent approach to the problem than has generally been given. 'he better account o# the logic o# invalidation, I suggest, can be #ound in an unli"ely place, namely in the writings o# the only philosopher who in#amously actually re)ected that logic altogether, 'homas *obbes. As we will see, *obbes has shoc"ed many generations o# scholars with his response to this problem by arguing that contracts with robbers are actually obligatory. 8ut I argue that delving into the *obbesian edi#ice a bit deeper gives us a basis #or teasing a power#ul argument #or invalidation out o# the *obbesian contractarian #ramewor". 'hat *obbes himsel# #or the most part #ailed to notice this should not provide a basis #or re)ecting it, or ironically, #or #ailing to credit *obbes with its discovery. 9et us now turn to the basic accounts o# coercion in the philosophical and the legal literature. 'he two approaches to coercion that predominate among scholars o# both sorts, whether their pro)ects are :de#initional< or :implicative,< are a deontological approach on the one hand, and a conse(uentialist on the other. Eiven the divide between the de#initional and implicative accounts, the #urther split between deontological and conse(uentialist approaches should in theory give us #our categories: two de#initional accounts, one deontological and one +pdated: -./0.12 3age 7 o# 94

conse(uentialist, and two implicative accounts, one deontological and one conse(uentialist account.@1 A. 7!isting Accounts 'o complicate matters #urther, deontological accounts, whether de#initional or implicative, seem to come in several varieties, depending on the #oundation on which the deontological claim rests. In particular, two versions o# the deontological account appear to predominate: :psychological< accounts and what might call :moraliJed< accounts.@@ 5n the psychological version o# the deontological account, coercion vitiates the agent&s ability to choose, so that his consent to the relevant agreement, while intentionally o##ered, does not re#lect his true will. Adherents o# the psychological approach, whether philosophical or legal, o#ten e!press this thought by saying that the agent&s :will was overborne,<@/ that the agent&s consent was o##ered :against his will,< or that the individual was :deprived o# his #ree will< or :mental #reedom,<@B and thus could not truly e!ercise choice. Alternatively, the point is sometimes put by saying that the agreement to the contract was :involuntarily< obtained.@C %any legal scholars ta"e the same view o# duress in the criminal law, e!plaining the law&s willingness to e!cuse an individual who commits a crime under duress by saying that he :did not act voluntarily< or that he was :not responsible #or his actions.<@0 3aul Robinson, #or e!ample, writes that :KtLhe +pdated: -./0.12 3age 8 o# 94 e!cusing condition in duress is the impairment o# the actor&s ability to control his conduct.<@F And Eeorge letcher says that :KeL!cuses apply on behal# o# morally involuntary responses to dangerM they ac"nowledge that when individuals merely react rather than choose to do wrong, they cannot #airly be held accountable.<@G 5n the psychological version o# the deontological claim coerced actions bear some resemblance to re#le!ive responses, in that they shortAcircuit the agent&s rational #aculties.

'he moraliJed version o# the deontological account #ocuses instead on the permissibility o# the coercive threat and the victim&s entitlement to resist the threat. 'he most prominent version o# the moraliJed account is that o##ered by Alan $ertheimer. $ertheimer distinguishes between what he calls the :proposal prong< and the :choice prong,< namely the re(uirements that the threat be wrong#ul, on the one hand, and that the agent&s choice be impaired, on the other.@'he #ormer re(uirement is understood to be a normative re(uirement, while the second is at least initially presented in psychological terms. $hile $ertheimer generally #ollows what he ta"es to be the prevailing approach in the law, by combining both choice and proposal elements, he does not thin" the choice prong can carry much weight.@2 *ence, $ertheimer ultimately advances a moraliJed account #ocused on the proposal prong. According to $ertheimer, the important element in coercion is the combination o# the wrong#ulness o# the coercer&s threatening proposal with the victim&s entitlement to resist the threat./1 5n the +pdated: -./0.12 3age 9 o# 94 choice side, $ertheimer calls :the principle that coercion undermines voluntariness< :uncontroversial,</@ but he insists on a moraliJed account o# voluntariness itsel#, according to which an agent acts voluntarily )ust in case he acts #rom secondAorder principles to which he adheres.// $ertheimer maintains that a threat coerces )ust in case it wrong#ully lowers its recipient&s moral baseline in such a way that it re(uires him to act according to secondAorder principles he does not truly accept, and which he is entitled to re)ect./B In this counterintuitive sense, he thin"s, the coerced agent acts :involuntarily.</C 'hus ultimately the choice prong is also normative, and the pure psychological position one normally #inds in an analysis o# involuntariness drops out. An alternative version o# the moraliJed account is that o##ered by Robert

4oJic". According to 4oJic", coercion re(uires the coercer to have threatened a conse(uence that each party "nows would leave the victim worse o## than he would be i# the coercer did not bring about that conse(uence./0 'here are #urther conditions, but they are not relevant #or my purposes. 'he core o# the account lies in 4oJic"&s analysis o# the notion o# a :threat.< Something is a threat i# the conse(uences would ma"e the recipient worse o## than he would be in the natural or e!pected course o# events./F 'hus a central conclusion o# 4oJic"&s is that threats can be coercive but o##ers cannot, and in this way 4oJic" is able to dispose o# a number o# marginal cases that one might otherwise have included among the core cases o# coercion. +pdated: -./0.12 3age 10 o# 94 Although the #oregoing deontological accounts are not particularly #ocused on the implications o# identi#ying a contract as coercive, it is generally assumed that the coercive nature o# an agreement, whether merely moral or contractual, is a reason #or its invalidation. 8ut while this point is assumed, the basis #or it is never #ully spelled out. or proponents o# a HtrulyI psychological approach, the tacit assumption appears to be that coercive contracts must be invalidated because they do not re#lect the voluntary agreement o# the victim. 'he precise relationship between involuntariness and invalidation, however, is never #ully articulated, and the assumption is that the invalidity o# coercive contracts #ollows seamlessly #rom involuntariness, since involuntariness implies lac" o# consent. In the second, more moraliJed account, the same conclusion would be drawn #rom a slightly di##erent premise, namely that the wrongfulness o# the coercive method ma"es the victim&s agreement :morally involuntary.< *ere the invalidation does not #ollow (uite as automatically as it does under the psychological account. 8ut the :moral involuntariness< o# coercive contracts

might easily be thought to entail the invalidity o# the agreement in a way that parallels this same move in psychological accounts: involuntariness entails invalidation because it de#eats true consent. And it de#eats consent, such theorists thin", because the consent is wrong#ully obtained #rom the victim. 'here is o# course something o# a worry about circularity on such accounts, a worry that does +pdated: -./0.12 3age 11 o# 94 not pertain to the same e!tent to purely psychological accounts. 'he moraliJed accounts seem to start #rom the premise that the pressure applied in cases properly labeled :coercive< deprives the victim o# consent because it wrong#ully lowers her baseline. 8ut what test do such accounts use to establish whether the lowering o# the baseline was in #act wrong#ul? 'his point is o#ten le#t open,/G but it is very di##icult to speci#y wrong#ulness on the coercer&s part other than by saying that it leads to morally unacceptable concessions on the part o# the victim. And this is (uite close to saying that pressure applied by the coercer is coercive i# it leaves the victim worse o## in a way we are inclined to )udge as coercive. It is not clear, then, that baseline accounts can supply the independent account o# coercion we re(uire to ma"e the baseline approach nonAcircular. 8y contrast with the #oregoing deontological approaches, the conse(uentialist approach to coerced contracts does not proceed by o##ering an account o# coercion, but instead directly addresses the (uestion o# the rami#ications o# coercion #or contract law. Conse(uentialists e!plain the invalidation o# coercive contracts not by something intrinsic to the nature o# coercion, by rather by the signi#icant social harm that would come #rom en#orcing such agreements. In particular, they point to the obvious #act that i# we en#orce coercive contracts, all things being e(ual, we would increase incentives to use coercion, and hence will increase the total amount o# coercion in society.

%ost conse(uentialist approaches to coercion in the legal literature hail +pdated: -./0.12 3age 12 o# 94 #rom the law and economics tradition. Richard 3osner, #or e!ample, e!plains the unen#orceability o# contracts under duress by saying, :the en#orcement o# such o##ers would lower the net social product by channeling resources into the ma"ing o# threats and into e##orts to protect against them.</- 5ther legal economists describe coercion along similar lines. Robert Cooter and 'homas +len, #or e!ample, worry that en#orcing coerced contracts would create an incentive #or :destructive acts.</2 Coercion is destructive in that it involves a :threat to destroy e!isting value.<B1 4onAcoercive o##ers, by contrast, are threats to :bloc" the creation o# additional value.<B@ inally, according to Cooter and +len, i# coerced contracts were en#orced, individuals will waste resources on protection against coercion.B/ 'he legal economist thus regards it as crucial that the state re#use to endorse contracts produced by illegitimate means, since this would create power#ul incentives #or individuals to engage in unproductive activity. Eiven that legal economists share the rational choice theorist&s view that rational agents are utilityAma!imiJing, the legal economist will be per#ectly prepared to allow that an individual who enters a coercive contract with the structure o# our robber e!ample is behaving rationally. +nli"e the deontologist, he would not see" support #or the invalidation o# such contracts by denying that the agents who enter into them are e!ercising their rationality #aculties. 8ut the legal economist also denies the relevance o# personal wel#are ma!imiJation in the #ace o# considerations o# social +pdated: -./0.12 3age 13 o# 94 wel#are. Eiven the personal advantages to the victim o# entering into a coercive contract under these conditions, it is a shame that we cannot allow the victim to ma"e use o# the institution o# contract law to #ree himsel# #rom a li#eA threatening

situation. 8ut general policies cannot be made by particular cases, and the damage o# such a general policy would be #ar too great. 8. 'he Rational Contractarian Approach 9i"e the legal economist, my primary #ocus in what #ollows will be the implications o# identi#ying an agreement as coercive, rather than the criteria #or ma"ing that identi#ication in the #irst place, as is the general #ocus in most o# the philosophical literature. $hat this means is that I shall #ocus on core cases o# coercive agreements=cases that must be analyJed as instances o# coercion on anyone&s theory=and see" to understand the basis #or re)ecting the legal validity o# such contracts. 'his is once again in contrast with the usual philosophical pro)ect that see"s to develop a comprehensive account o# coercion by developing a de#inition o# coercion and then attempting to e!tend that de#inition to marginal cases. 'he nuances o# subtle variations in de#inition are there#ore mostly irrelevant #or our purposes, since our problem will be the #ar more basic one o# whether and why coercive contracts under any reasonable definition should be invalidated, in light o# the rather strong pre#erences o# both coercer and victim to the contrary. +pdated: -./0.12 3age 14 o# 94 In addressing the (uestion o# the implications o# coercion #or the en#orceability o# contracts, I adopt what I call a :rational contractarian< perspective, namely a perspective #rom which legal policies and rules must be )usti#ied to the individuals whose conduct falls under their sway. 'he approach is contractarian inso#ar as it re(uires us to be able to posit, in at least a hypothetical sense, the agreement o# such persons to the rule or policy in order #or its authority over them to be )usti#ied. It is :rationalistic< inso#ar as the approach it will ta"e to this (uestion will be a #unction o# the perceived sel#Ainterest o# the parties: rational agents would agree to a rule or policy only inso#ar as they regard it, broadly spea"ing, as in their interest. It is to be contrasted in this respect with

:normative contractarian< approaches,BG such as a Rawlsian or a Scanlonian might endorse, because agreement to a social rule or policy would not be predicated on normative #actors such as the representation o# human beings as :#ree and e(ual persons,<B- or what human beings could not :reasonably re)ect.<B2 5n a rationalistic approach, the (uestion to as" is whether you and your assailant in the "ind o# coercion scenario with which I began would regard either o# the above rationales=the deontological or the conse(uentialist=as ade(uate grounds #or re)ecting your contract. %ost saliently, would you, the victim, regard either e!planation as a satis#actory basis #or denying your urgent plea to ma"e your contract with your assailant en#orceable? rom the victim&s standpoint, the #oregoing accounts will either appear #lawed as accounts o# coercion in their own +pdated: -./0.12 3age 15 o# 94 right or will seem to provide inade(uate grounds #or re#using to en#orce the contract. Consider #irst the choice version o# the deontological approach. rom the victim&s point o# view, the argument will seem curiously bac"wards. $hat basis is there, he might thin", #or accusing a person who pre#ers to stri"e a monetary deal rather than su##er immediate death o# irrationality? 'o suggest that a person would only ma"e such a commitment because his rational processes had been shortAcircuited or his will overborne seems absurd. 'he victim in such a case would #eel that nothing could be #arther #rom the truth, and #urthermore that the person who choose not to ma"e such a commitment rather than lose his li#e is the one whose rationality ought to be (uestioned. Indeed, any victim who was able to negotiate such a deal #or hersel# is surely to be commended #or her clear thin"ing rather than dismissed #or lac" o# rationality. And this suggests that contrary to the reigning orthodo!y on this (uestion, the victim&s action in a case o# this sort is

both chosen and voluntary, despite the #act that it is not an action she would choose #or its own sa"e.C1 'he rational choice perspective on voluntariness, according to which an action is voluntary )ust in case the agent per#orms it #or the sa"e o# something he hopes to attain, is actually not the product o# modern egoism as some philosophers would choose to portray it. It is in #act the teleological approach to intentional action that begins with Aristotle, and that authors as little interested in egoism as 7liJabeth Anscombe have e!pounded. C@ +pdated: -./0.12 3age 16 o# 94 5nce one considers it care#ully, the #oregoing point is di##icult to challenge. Although he wants to preserve the connection between coercion and involuntary action, $ertheimer too recogniJes that coercion cannot impair voluntariness in the way that re#le! reactions do, or even in the more attenuated way in which physical torture might. In standard cases o# what he calls :volitional coercion,< namely coercion that operates through threats rather than physical #orce, the victim&s deliberative #aculties are intact.C/ Indeed, we might say that volitional coercion presupposes intact rational #aculties, as the threat would be wholly deprived o# e##ectiveness i# the victim lac"ed the capacity #or rational decision ma"ing. As 'homas Schelling long ago suggested, it is in theory possible to render onesel# immune to coercion simply by ma"ing onesel# irrational, #or e!ample, with the aid o# an irrationality pill.CB 'he "ind o# coercion I have been discussing thus achieves its aim by ma"ing use o# rationality rather than obviating rationality: it involves a piling on o# reasons on the side o# one o# the courses o# action available to the agent, rather than delusion, drugging, or impairment o# rational #aculties.CC 'he con#used suggestion o# many philosophers and lawyers that coerced actions are involuntary is nevertheless understandable. irst and #oremost, they come to this conclusion via the apparently irre#utable conviction that coercive

agreements must be invalid. 9ac"ing any clear normative e!planation #or why this must be, they see" a psychological basis #or their intuitions. In this way o# +pdated: -./0.12 3age 17 o# 94 thin"ing, however, they are engaging in bac"ward reasoning, as they are in#erring the psychology #rom the normative conclusion it is supposed to support. urthermore, this mista"en move is probably #urther encouraged by a "ind o# slide #rom :nonvolitional< coercion, namely coercion based on physical #orce, to volitional coercion. or where nonvolitional coercion is concerned, it is per#ectly correct that choice, decision ma"ing, and re#lection have no role to play in producing the actions o# the agent. In such cases, the agent becomes a mere instrument o# someone else&s will. ;et volitional coercion is signi#icantly di##erent, inso#ar as an agent still selects a course o# action because he sees it as in some way better #or himsel#.C0 Rational choice theorists, then, seem correct in their assumption that bad choices are not to be e(uated with nonA choices, and a power#ul incentive to do one thing rather than another does not ma"e the agent&s subse(uent actions in any meaning#ul sense involuntary. or the #oregoing reasons, the moraliJed versions o# the deontological approach seem at #irst blush more promising. I will consider $ertheimer&s and 4oJic"&s accounts together, since they have a common structure. In both accounts, the assailant has impermissibly lowered the baseline o# the victim by putting her into a situation in which she is #orced to ma"e a highly undesirable choice: a choice to avoid a situation that is below a certain moral baseline or a choice to avoid a situation that diverges #rom the :natural or e!pected course o# events, < which sets a nonAmoral baseline.CF 8oth $ertheimer and 4oJic" +pdated: -./0.12 3age 18 o# 94 recogniJe moral and nonAmoral baselines. %oral baselines enter the discussion to address the problematic case o# the slave who is threatened with a beating unless

he does something #or his owner. 'he comple!ity o# the case stems #rom the #act that the slave can e!pect a beating tomorrow no matter what, so there would be no divergence #rom the natural and e!pected course o# events Hthe nonA moral baselineI. +sing a nonAmoral baseline in this case would suggest that it is not an instance o# coercion, but our sense that the threat is highly coercive suggests that a moral baseline might be more appropriate.CG 4oJic" and $ertheimer diverge, however, in how they thin" we should determine which baseline is appropriate to use in which circumstances. 4oJic" believes that the appropriate baseline is the baseline that the potential victim himsel# would prefer were used in assessing his sub)ection to coercion.C- $ertheimer believes that the appropriate baseline is determined by context and our overall moral theory.C2 At this point, both views #all into the problem that I have been stressing. Regarding 4oJic", i# the victim&s pre#erences matter in choosing the appropriate baseline, the victim&s pre#erences should also matter in the selection o# the correct moral and legal response to the coerced contract And in $ertheimer&s case, i# our overall moral theory is what determines the choice o# baselines, a moral theory that re(uired that a proposed rule be )usti#ied to a victim might have the result o# suspending a duress de#ense in contract law rather than supporting it. Conse(uently, considerations that go into identi#ying the proper baseline seem +pdated: -./0.12 3age 19 o# 94 ultimately to suggest that identi#ying a case as an instance o# coercion does not automatically generate the implications o# such an identi#ication. 'hus $ertheimer and 4oJic", whatever their di##erences, would presumably both want to re)ect the suggestion that the victim is made better o## by entering into a contract with the assailant. 'he sense o# :better o##< in this case is an arti#icial product o# the assailant&s wrong#ul manipulation o# the victim&s options, and this gives us a basis #or depriving the assailant o# the gains #rom his

coercive tactics. Since coercion impermissibly lowers the victim&s baseline by restricting her options in a way that she strongly dispre#ers, we cannot thin" o# the agreement as consented to, even hypothetically, by the victim. 8ut i# we ta"e up the victim&s standpoint once more, we can see that the #oregoing moral versions o# the deontological argument will be no more compelling than the psychological version. :'rue,< the victim will say, :I did not want to be placed in this position, and would o# course rather not have had to #ace the choice between my money and my li#e. And it is also true that the means my assailant used to place me in this position was wholly illegitimate: using #orce to limit my options and thereby improve his own position is the clearest e!ample o# an improper means #or gaining an advantage over another imaginable. 8ut none o# that is particularly relevant to me now,< she might say, :#or #ocusing on the way I came to be in this position does not now bene#it me one bit. 'hose #eatures o# my situation are in e##ect :sun" costs.< %y only interest now is how I can save +pdated: -./0.12 3age 20 o# 94 my li#e, and the best way to do that is to enter into a binding agreement that commits me to return with money the ne!t day. So although I agree that it is e!tremely un#air o# my assailant to have put me in a position where I had to resort to such an agreement, it seems doubly un#air to now deprive me o# the only means I have to save my li#e under such circumstances.< 'he victim has a point. I# the law #ocuses on the wrong#ulness o# the assailant&s conduct to the e!clusion o# her wel#are, it will be in e##ect conspiring to lower the victim&s baseline even #urther than the assailant would have lowered it i# le#t to his own devices. 'he moraliJed version o# the deontological approach thus seems no more promising #rom the rational contractarian perspective than the psychological version was, as it leaves the victim&s plea entirely unanswered.

It might be tempting to suppose that the conse(uentialist approach to this problem would best capture the rational choice perspective, since deterrence is an important concern #or rational agents in the conte!t o# social agreement. 8ut a moment&s re#lection will ma"e clear that the appeal to deterrence also #ails to capture the viewpoint o# rational agents who are to be sub)ect to the rules governing coercive contracts, and so #ails to respect the most basic principle o# contractarian approaches, namely that we can represent the resulting social arrangement as contractual because we can in#er agreement #rom the #act o# mutual bene#it. 'hus, #or e!ample, the economic version o# the conse(uentialist solution says that we must re)ect coerced contracts because en#orcing them would +pdated: -./0.12 3age 21 o# 94 produce overall lower social utility than re)ecting them, despite the #act that #rom the standpoint o# the individual parties, en#orcing the contract may be 3areto optimal. %ore generally, a conse(uentialist might say that neither party would want to live in a world in which anyone can ma"e e!tensive use o# their superior threat advantage and receive legal blessing #or doing so. 'he lives o# all o# us would be less secure i# such agreements were rendered en#orceable. Conse(uentialists there#ore assert that the plight o# particular victims must there#ore yield to the sound policy o# deterrence. It is a shame that this policy will have its casualties, but there is no other reasonable way to guard against the specter o# massive increases in coercive demands that #ailing to adopt this policy would produce. 5nce again, #rom a rational contractarian perspective, there is a power#ul basis #or re)ecting this rationale as a reason #or re#using to en#orce coercive contracts. Addressing the situation #rom the victim&s perspective, it will be hard to see enhancements to overall social utility as an ade(uate basis #or )usti#ying the

adoption o# a policy that will harm her. She will rightly see hersel# as a sacri#icial lamb: :It is bad enough that I was put in this aw#ul position by my assailant,< she might say, :but now I am being condemned to death #or the sa"e o# a social policy that will benefit everyone else but me! Indeed, even my coercer will bene#it #rom this policy, since he will live to tell the tale and so to en)oy the #uture bene#its o# this deterrence, but I, whose li#e must now end, will not.< And so once again, +pdated: -./0.12 3age 22 o# 94 #rom a rational contractarian perspective there is reason to conclude that the deterrence account is seriously lac"ing in necessary comple!ity, because it #ails to supply a )usti#ication that the victim in this scenario would have reason to see as compelling. 'his line o# attac" raises a #undamental (uestion about what would count as an ade(uate )usti#ication #or the policy o# nonAen#orcement o# coercive contracts. I have suggested that such a policy cannot be )usti#ied to the victim who must su##er its conse(uences, and thus #ails the rational contractarian re(uirement that the parties who must live under the rule see something #or themselves in it i# its authority over them is to be legitimate. 'o summariJe the di##iculty, the problem in this case is that the di##use personal bene#it each may receive #rom an enhancement to overall social utility may not stri"e rational agents as ade(uate bene#it to counterbalance whatever personal loss o# utility the rule may entail. In this conte!t, the (uestion #rom a contractarian perspective is not whether the coercer has impermissibly lowered the victim&s baseline below what it would have been without his intervention, or whether social utility would, in the aggregate Hor even on averageI be enhanced by re#using to en#orce coerced contracts. 'he (uestion is rather whether individuals selecting the terms of their own legal system would choose to include a duress defense that invalidated

coercive contracts or whether they would choose instead to eliminate that defense +pdated: -./0.12 3age 23 o# 94 and allow such contracts to be enforced. 'he relevant (uestion, in short, is precisely the one raised, but not answered, in the imaginary conversation between the victim and the robber: i# the parties were to sit down together long in advance o# any need to enter into a coercive agreement, would they choose to retain the duress de#ense or would they choose to eliminate it? 4ote that economists sometimes pay lip service to the contractarian approach by suggesting that it is evidence #or the truth o# the economic conclusion. 3osner, in his remar"s about coerced contracts (uoted above, continues by saying: :$e "now that this class o# :contracts< is nonA optimal because e! ante=that is, be#ore the threat is made=i# you as"ed the 8s o# this world whether they would consider themselves better o## i# e!tortion #lourished, they would say no.<0@ 8ut it is #ar #rom clear that 3osner is correct in this regard. In e##ect, despite the respect#ul remar"s made about the contractarian thought e!periment, legal economists do not ta"e seriously the individualistic perspective that the contractarian approach in #act re(uires, and so in the end would not be prepared to accept the upshot o# any actual or even putative agreement among rational agents.0/ *ow would we go about determining what the parties to a potential e! ante :coercion< agreement would decide? 'he (uestion is whether the parties would consider themselves better o## living with the added deterrent bene#its that ma"ing coercive contracts unen#orceable would provide, or whether they would +pdated: -./0.12 3age 24 o# 94 pre#er to protect their ability to buy their way out o# coercive situations in case they #ound themselves sub)ect to such coercion. rom an e! ante perspective, the answer is #ar #rom clear. Eiven that the parties are aware prior to the moment o#

contracting that they may need to save their lives by ma"ing an en#orceable promise to a robber, and given the intensity o# their desire to ma"e binding promises under such circumstances Hwhich they can anticipateI, the balance o# personal utilities could weigh in #avor o# allowing such contracts, and so the parties might select a regime that en#orced coercive contracts, even in the absence o# the immediate need to enter one. 'he rational course o# action would depend on the increased li"elihood o# coercion, the e##ectiveness o# the promises one could ma"e under such circumstances, and so on. or the rational choice theorist, the answer to the (uestion o# coercive contracts thus appears to be a highly contingent one.0B 'o ma"e the suggestion that the parties might opt #or protecting their ability to buy their way out o# coercive situations more plausible, consider the conditions that obtain in some countries today o# #re(uent "idnappings #or ransom. In such countries, "idnapping is usually economically motivated and victims are nearly always released i# the ransom money is provided. 8ecause law en#orcement is wea" and the chances o# prosecution low, the victims are virtually certain to be "illed i# the ransom is not provided. Clearly there are signi#icant utilitarian advantages to the position that "idnappers should not be rewarded with +pdated: -./0.12 3age 25 o# 94 ransom money. I# such a principle were consistently adopted, there would be #ewer "idnappings in that country. It is e(ually clear, however, that once one is the victim o# such a "idnapping, one would desperately hope that one&s country did not #ollow a noAcooperationAwithA"idnappers policy. $hat this "ind o# case reminds us, then, is that there is a potentially signi#icant gap that obtains between personal utility and social utility calculations. 'here appears to be no particular reason to assume that the ma!imiJing solution #rom the standpoint o# social utility

will be the same solution that rational agents see"ing to ma!imiJe their own utility would adopt, i# unanimous or close to unanimous agreement is re(uired. or this reason, the utilitarian and the contractarian solutions to social problems will o#ten diverge.0C In this case, the utilitarian solution would li"ely support a noA cooperationwithA "idnappers policy.00 8ut what would the rational contractarian solution suggest? 'he relevant (uestion, once again, is what a rational agent would say about the decision to adopt the noAcooperation policy i# as"ed prior to being "idnapped, "nowing there is a certain li"elihood o# #inding himsel# a "idnapping victim with no other way. 5n the one hand he must consider the e##ects on #uture "idnapping victims o# entering into bargains with "idnappers, as well as contemplate the li"elihood that he would be among them. $ith only these two #actors to consider, it would not be surprising i# he reached the same conclusion as the utilitarian social planner, given that the loss o# deterrence might be severe, +pdated: -./0.12 3age 26 o# 94 while the chances o# being "idnapped #or a single individual might remain low. 8ut when, on the other hand, he considers the gravity o# the potential evil, and considers the importance o# preserving his ability to buy his way out o# a "idnapping situation he may reach a very di##erent conclusion. In light o# these considerations, it is #ar #rom obvious that #rom an e! ante perspective one would side with the more abstract social policy considerations over one&s personal pre#erences as one imagines them to be e! post, even though the individual is also the bene#iciary o# the social policy. A #urther argument, however, may trans#orm the nature o# the perceived bene#it and provide a de#initive reason #or the rational agent to #avor protecting his ability to save himsel# #rom dire, but unli"ely situations, and hence to re)ect the conse(uentialist approach. In many situations it seems possible to achieve the

deterrent e##ect one see"s #rom a policy o# nonAcooperation or nonA en#orcement by simply increasing the penalties #or the initial coercion itsel#. Assuming that penalties can always be #urther increased, there is no reason to thin" that the deterrent bene#it on which the utilitarian argument depends must come #rom invalidating the contract Has though invalidation were the only available way to create disincentivesI. In cases in which independent punitive measures are possible, the individual utilities #rom allowing coercive contracts can be assessed (uite separately #rom the deterrent bene#its #rom invalidating coercive contracts. 'here would then be little individual bene#it to banning coercive contracts or +pdated: -./0.12 3age 27 o# 94 re#using to cooperate with "idnappers, and, on the other side, signi#icant individual security #rom retaining the ability to enter into such contracts or engage in such negotiations. or the #oregoing reasons, the deontological and the conse(uentialist arguments against en#orcing coercive contracts have signi#icant wea"nesses, particularly #rom a rational contractarian perspective. $hile the ob)ections to these approaches may be compelling, developing an a##irmative alternative approach to the problem is (uite another matter. I will be aided in this tas" by considering the remar"s o# a thin"er who adopts the rational choice perspective, namely those o# 'homas *obbes. It is perhaps not surprising that *obbes is also the only philosopher to have written on this sub)ect who thin"s that, at least in some circumstances, contracts under coercion should actually be en#orceable. I will use *obbes& remar"s on this topic to consider whether the contractarian approach is inevitably committed to this counterAintuitive position, or whether the more common sense conclusion that coercive contracts are unen#orceable might yet be )usti#ied in rational choice terms. III. *5887S A46 '*7 R5887R A. 4atural and Civil Coercion

Against what is surely the nearly universal position on coerced contracts, +pdated: -./0.12 3age 28 o# 94 'homas *obbes stridently de#ends the claim that agreements entered into under duress or coercion are binding. *is #irst de#ense o# this position appears somewhat con#usedly in De Cive, where he writes as #ollows: It is a usual (uestion, whether compacts e!torted #rom us through #ear, do oblige or not. or e!ample, i#, to redeem my li#e #rom the power o# a robber, I promise to pay him @11l. the ne!t day, and that I will do no act whereby to apprehend and bring him to )ustice: whether I am tied to "eep promise or not. 8ut though such a promise must sometimes be )udged to be o# no e##ect yet it is not to be accounted so because it proceedeth #rom #ear. or then it would #ollow, that those promises which reduced men to a civil li#e, and by which laws were made, might li"ewise be o# none e##ectN0F *obbes& position comes substantially into #ocus when he ta"es up this topic in DIO o# Leviathan. *e #irst addresses this (uestion in the conte!t o# agreements entered into in a state o# nature to #ree onesel# #rom the grip o# an enemy: Covenants entered into by #ear, in the condition o# mere nature, are obligatory. or e!ample, i# I covenant to pay a ransom, or service #or my li#e, to an enemy, I am bound by it. or it is a contract wherein one receiveth the bene#it o# li#eM the other is to receive money, or service, #or itM and conse(uently, where no other law Has +pdated: -./0.12 3age 29 o# 94 in the condition o# mere natureI #orbiddeth the per#ormance, the covenant is valid.0G 'hat is, *obbes& reasoning is that the enemy and his victim e!change items o# value, and each receives a bene#it #rom the other in so doing. or *obbes, the bene#it received ma"es the commitment #ully voluntary, as we have already seen. 'here is thus no psychological basis #or invalidating the contract. *e e!plains: 'he cause o# #ear which ma"eth such a covenant invalid must be always something arising a#ter the covenant made Has some new #act or other sign o# the will not to per#ormI, else it cannot ma"e the covenant void. or that which could not hinder a man #rom promising, ought not to be admitted as a hindrance o# per#orming.0It is not the #ear itsel# that invalidates the agreement, but the cause o# the #ear, as long as that cause has to do with new circumstances not apparent at the time the

contract was made. Signi#icantly changed circumstances or acts o# Eod, as they are typically called, might ma"e a contract invalid, but only i# unavailable to the parties at the time o# contracting. *obbes& basic thought is that circumstances available to the parties at the time o# the agreement do not impair the parties& obligations i# they did not impair the parties& ability to e!change promises in the #irst place.02 And this implies that promises made as part o# an e!change always obligate the promisor, as long as they are voluntarily made, in *obbes& sense.F1 +pdated: -./0.12 3age 30 o# 94 *obbes ta"es the same position with regard to the De Cive e!ample, namely the case o# agreements made with robbers in civil society. And as *obbes also points out in the De Cive discussion o# this point, the sovereign may have good reasons #or #orbidding the robber&s conduct separately, which would provide a conventional limitation on one&s natural obligations: KILn commonwealths, i# I be #orced to redeem mysel# #rom a thie# by promising him money, I am bound to pay it, till the civil law discharge me. or whatsoever I may law#ully do without obligation, the same I may law#ully covenant to do through #earM and what I law#ully covenant, I cannot law#ully brea".F@ *obbes thus thin"s an agreement entered into with a robber in civil society, made to save one&s li#e is binding, as long as the law does not #orbid such contracts. 8ut since he also thin"s the civil law should be a translation o# the natural law into civil #orm, we can suppose that he probably thin"s the law should en#orce such contracts. At any rate, no civil government has a duty to invalidate contracts o# this sort #or moral reasons, *obbes #irmly believes. *obbes& account o# why and when coercive contracts are binding in civil society is largely the same as the one he o##ers #or the state o# nature. In spea"ing o# :whatsoever I may law#ully do without obligation,< *obbes is suggesting that a person can bind himsel# to anything, as long as he is #ree to do or not to do it. And this suggests that agreements entered into out o# #ear are binding, as long as +pdated: -./0.12 3age 31 o# 94

the promised actions are not otherwise the sub)ect o# obligations. 'hus acts that are already obligatory, one cannot now ma"e more obligatory by promising once more to per#orm them, and acts that are the sub)ect o# prohibitions, such as illegal acts, one cannot bind onesel# to per#orm. 5nce again, the relevant logic is that what does not impair the promise should not impair the contract. In this case, the agent is #ree=not impeded=#rom giving a robber money i# he wishes. *e is not already obligated to do so, nor is he #orbidden #rom doing so. And what he was #ree to promise he is also #ree to impose on himsel# as an obligation. 9est one suppose that *obbes is not serious about these remar"s about the robber, the same idea is recapitulated in his discussion o# sovereignty by ac(uisition, according to which :the sovereign power is ac(uired by #orce.<F/ 'he sovereign power in such cases is created when men :#or #ear o# death or bonds do authoriJe all the actions o# that man or assembly that hath their lives and liberty in his power.<FB *obbes goes on to e!plain in this passage that the only di##erence between sovereignty by institution and sovereignty by ac(uisition is that in the #ormer case men choose their sovereign :#or #ear o# one another,< whereas in the latter case, it is #or #ear :o# him whom they institute.<FC As *obbes points out, the two sorts o# sovereignty are ultimately not too di##erent, since in both cases :they do it #or #ear,< which he thin"s is worth noting as against people who erroneously :hold all such covenants as proceed #rom #ear o# death or violence void.<F0 urther, *obbes argues, i# #ear did indeed render all agreements void, :no +pdated: -./0.12 3age 32 o# 94 man in any "ind o# commonwealth could be obliged to obedience,<FF i.e. no man would be obligated to obey the law. or e!ample, *obbes says elsewhere, a man might pay a debt only because he #ears imprisonment.FG 8ut he nevertheless pays the debt #reely, and he was at liberty not to pay it. *obbes is suggesting,

reductioAstyle, that coercion could not negate #reedom because we would then have the absurdity that we could no longer see anyone as obeying the law o# his own #ree will. Coercion, li"e law more generally, merely changes the payo##s #or rational agents. It does not deprive those agents o# their ability to choose. *obbes& argument on this topic is stri"ing, not )ust because o# the counterintuitive nature o# its conclusion, but also because it seems to be at odds with one o# the central premises o# his political philosophy, namely the idea that in a state o# nature, there are no obligations o# any "ind. As *obbes #amously says in Chapter DIII o# Leviathan in describing the state o# nature: :'he notions o# right and wrong, )ustice and in)ustice have there no place. $here there is no common power, there is no lawM where no law, no in)ustice.<F- It #ollows #rom the Chapter DIII remar"s that no contract in nature can be binding, as there is no power to en#orce any agreement. *ow, then, does *obbes reconcile his claim that contracts with robbers are binding in a state o# nature with his e(ually clear statements to the e##ect that contracts in a state o# nature are not in general binding? %oreover, i# the general appeal to the invalidity o# contracts were not +pdated: -./0.12 3age 33 o# 94 compelling enough, *obbes ma"es a more speci#ic point that should compound any con#usion about the case o# the robber. *e says, :he which per#ormeth #irst does but betray himsel# to his enemy, contrary to the right Hhe can never abandonI o# de#ending his li#e and means o# living.<F2 rom this one would have e!pected the #ollowing line o# reasoning: since the robber, as #irst per#ormer, merely :betrays himsel#< to his victim, and cannot there#ore e!pect promises made to him to be "ept, his victim has no obligation to "eep the promise, since it would not have been rational #or the robber to rely on him in the #irst place. 'his would seem to be the obvious conclusion #rom a rational choice perspective, and it

would have perhaps been a sensible thing #or *obbes to argue. 8ut curiously, it is not the view he too". Instead, he appears to be saying that although contracts in nature are generally not binding, contracts with robbers are. 8ut how could any sane person de#end such a claim? Are we sure *obbes really believed it? 8. Contracts and Covenants 'he e!planation may lie in a distinction *obbes ma"es between two "inds o# contracts, which I call :#irstAper#ormer< and :#uture per#ormance< contracts,G1 as well as in the argument that *obbes uses against the in#amous #ool, whose challenge to the third law o# nature provides a pivotal insight into *obbes& political philosophy more generally. Considered care#ully, it appears that *obbes is not in #act committed to the blan"et suggestion that contracts in a state o# nature +pdated: -./0.12 3age 34 o# 94 are invalid. *obbes actually thin"s that some contracts are rational to "eep, even in the absence o# en#orcement or moral obligations. 'hese contracts he calls :covenants,< in order to distinguish them #rom ordinary contracts, which he calls :contracts.< 'he #ormer is a contract in which one party per#orms at the time the contract is made and the other party commits to per#orming some time in the #uture,G@ whereas the latter is one in which both parties commit to per#orming at some time in the #uture.G/ It will there#ore turn out that the contract with the robber is a species o# #irstAper#ormer contract, or :covenant,< and as such provides an e!ception to the general rule that contracts in nature are not binding. %any commentators have overloo"ed the distinction *obbes was trying to draw between covenants and contracts. Among the commentators that have #ocused on the distinction, the usual view is that a covenant is )ust an ordinary e!change o# promises #or #uture per#ormance o# some sort. *oward $arrender, #or e!ample, says that a covenant is when :one or both parties are to per#orm in

the #uture and are to be trusted in the meantime.<GB Similarly, Eregory Kav"a says that :Contracts in which one or more parties are called on to per#orm their parts at some time a#ter the contract is made are covenants.<GC And 6avid Eauthier says that :A covenant is that species o# contract in which at least one o# the parties Pis to per#orm in time to come.&<G0 'here are good reasons, however, not to read the notion o# a covenant in this openAended way. irst, in the passage where *obbes introduces the notion o# covenant, he +pdated: -./0.12 3age 35 o# 94 clearly has in mind a case in which one party per#orms now in e!change #or a promise on the other party&s part to per#orm in the #uture. *e writes: :Again, one o# the contractors may deliver the thing contracted #or on his part, and leave the other to per#orm his part at some determinate time a#ter Hand in the meantime be trustedIM and then the contract on his part is called 3AC' or C5O74A4'.<GF And second, the #oregoing interpretation leaves commentators with no way o# e!plaining the distinction between covenants and ordinary contracts, despite the #act that they notice *obbes drawing such a distinction. 'hey cannot, #or e!ample, say that covenants are that species o# contract in which at least one party must per#orm in the #uture, leaving the larger class o# contracts to include agreements in which both parties per#orm at the time the contract is made. or *obbes says a contract re(uires a :mutual trans#erring o# right,< and #or this reason, he does not thin" o# barter as a contractual e!change.GG *obbes gives us no other hint about how to draw the distinction between covenants and ordinary contracts, and it is hard to see any reasonable alternative. 4ow there is a distinction in modern law that roughly corresponds to the distinction between #utureAper#ormance and #irstAper#ormer contracts, and this is the distinction between soAcalled :bilateral< and :unilateral< contracts. 8ilateral contracts are contracts in which parties e!change promises, and both

per#ormances are to occur in the #uture.G- +nilateral contracts are contracts in which one party e!changes a current per#ormance #or another party&s promise.G2 +pdated: -./0.12 3age 36 o# 94 'he distinction between bilateral and unilateral contracts is very close to the distinction *obbes is trying to draw between contracts and covenants, but does not e!actly trac" it, since what *obbes is concerned to condemn is being a first performer, in any type o# contract in which it might be re(uired.-1 And this suggests that it would be rational to act on a bilateral contract i# one&s per#ormance were re(uired second, rather than #irst, as long as the other party had already per#ormed. 5n the other side, *obbes would presumably be prepared to re)ect the rationality o# entering into a unilateral contract, i# the terms o# the contract re(uired one to be the #irst per#ormer. $e cannot, then, simpli#y and say that *obbes regards bilateral contracts as void but unilateral contracts as valid in a state o# nature, since the distinction he traces is that between being a #irst per#ormer versus being a second per#ormer, a distinction that cuts at right angles the distinction between bilateral and unilateral contracts. $hat, then, are the implications o# *obbes& view o# being a #irst per#ormer #or the validity o# contracts in nature? irst consider bilateral contracts. I# it is irrational to be a #irst per#ormer, then it would be rational to re)ect any e!change o# promises in which both parties are to per#orm in the #uture, as it would be irrational to commit to #irst per#ormance, and naQve to commit to second per#ormance Hsince the #irst per#ormer will almost certainly never per#ormI. So a mere e!change o# promises cannot be a valid contract in nature. 'hus the condemnation *obbes o##ers o# the +pdated: -./0.12 3age 37 o# 94 rationality o# being a #irst per#ormer will invalidate the rationality o# bilateral contracts in nature across the board. 'he #act that *obbes allows that being a

second per#ormer can be rational, then, does not help to establish the rationality o# bilateral contracts. $hat about unilateral contracts? Arguably, on *obbes& assumptions, it is rational to o##er to e!change a promise #or another&s per#ormance. According to blac" letter law, o##ers #or unilateral contracts are deemed accepted only upon per#ormance.-@ 'hus i# A o##ers to pay 8 >011 i# 8 climbs to the top o# a #lagpole, 8 is not deemed to have accepted until he reaches the top, and correspondingly A is not bound to pay until 8&s per#ormance is complete. 'hus A&s obligation would commence only once he received the bene#it he was trying to achieve. And this ma"es second per#ormance very di##erent in a unilateral contract than in a bilateral contract, given that the second per#ormer can promise to per#orm in such a contract without ris". rom this it would appear to #ollow that it is rational to be the o##eror or promisor in a unilateral contract by *obbes& lights, but not to accept another&s o##er #or a unilateral contract by per#orming #irst. *obbes& de#ense o# the rationality o# "eeping #irstAper#ormer contracts appears in what might seem an unli"ely place, namely in the response he o##ers to the #ool&s challenge to the third law o# nature. 'he third law o# nature, to recall, is presented as a blan"et in)unction to "eep covenants: +pdated: -./0.12 3age 38 o# 94 rom that law o# nature by which we are obliged to trans#er to another such rights as, being retained, hinder the peace o# man"ind, there #olloweth a third, which is this that men per#orm their covenants made, without which covenants are in vain, and but empty words, and the right o# all men to all things remaining, we are still in the condition o# war.-/ 4otice that the third law o# nature, as stated, applies to all contracts, and not only to #irstAper#ormer contracts. %y suggestion in the preceding pages, however, has been that *obbes intends it to apply only to #irstAper#ormer contracts. $hile e!ploring *obbes& remar"s in these wellA"nown passages may seem something o#

a digression, understanding his argument on this point turns out to be crucial #or analyJing his de#ense o# contracts with robbers, and in particular #or understanding how such contracts are to be distinguished #rom the mass o# ordinary contracts in nature. 'he ool argues that there can be no general rule o# reason that advises men to "eep their covenants, since it can only be rational to "eep covenants when it would leave one better o## to do so.-B 5n this view, the rationality o# "eeping covenants is entirely an empirical matter: there simply is no a priori way to rule out the possibility o# cases in which it is rational to brea" covenants. 4ow *obbes allows that the ool is generally correct about contracts in nature. 8ut he insists that the ool&s position does not apply to second per#ormance on a #irst+pdated: -./0.12 3age 39 o# 94 per#ormer contract: or the (uestion is not o# promises mutual where there is no security o# per#ormance on either side Has when there is no civil power erected over the parties promisingI, #or such promises are no covenants, but either where one o# the parties has per#ormed already, or where there is a power to ma"e him per#orm, there is the (uestion whether it be against reason, that is, against the bene#it o# the other to per#orm or not. And I say it is not against reason.-C Consistent with the argument about bilateral and unilateral contracts I considered earlier, in this passage *obbes is associating cases in which one party has already per#ormed with cases in which there is a common power over the parties to ma"e them per#orm. 'hat is, #irstAper#ormer contracts are treated as on a par with contracts entered into in civil society. And such contracts, *obbes e!plains, are not sub)ect to the general invalidity o# contracts in nature, and hence are e!empt #rom the ool&s challenge. *obbes gives a rather elusive e!planation #or this surprising claim. irst, he suggests that antisocial acts that normally tend to one&s own destruction cannot be rationally vindicated )ust because they turn out to be advantageous on

particular occasions. 'he #act that one can get away with violating a contract every once in a while does not mean that habitual contract brea"ing is a rational +pdated: -./0.12 3age 40 o# 94 thing to do. *obbes then continues in this vein but adds a #urther argument: *e, there#ore, that brea"eth his covenant, and conse(uently declareth that he thin"s he may with reason do so, cannot be received into any society that unite themselves #or peace and de#ence but by the error o# them that receive himM nor when he is received, be retained in it without seeing the danger o# their errorM which errors a man cannot reasonably rec"on upon as the means o# his security . . . and so as all men that contribute not to his destruction #orbear him only out o# ignorance o# what is good #or themselves.-0 $hat is *obbes& argument here? 5ne way to understand it is to see it as a #orm o# a reductio. Suppose it were rational to be a covenantAbrea"er in an e!pected bene#it sense, because, let us suppose, the chances o# detection were su##iciently low, and the bene#its o# being a #ree rider su##iciently high, that it seemed rational to issue a deceit#ul promise. Assume that others have the same access to the rationality o# one&s conduct that one does onesel#=in other words, that i# a certain course o# action is rational, and is "nown to the actor to be so, then others will understand it to be rational as well. 'his assumption is an o##shoot o# what game theorists call :common "nowledge o# rationality.< -F It #ollows that i# it were rational to covenant to per#orm and nevertheless de#ect, others would "now it was rational to do so. 'he result would be that they would +pdated: -./0.12 3age 41 o# 94 not e!pose themselves by covenanting with you, and you would end up e!cluded #rom cooperative arrangements. Against the bac"ground o# the common "nowledge assumption, the #irst per#ormer would not have per#ormed in a unilateral contract unless it were rational #or the second per#ormer to "eep his promise and per#orm in turn. I# the #irst per#ormer has per#ormed, then, it #ollows that he has ta"en the second per#ormer to be a rational agent, and "nowing

himsel# to be such, the second per#ormer has every reason to per#orm. I# I am correct in my reading o# *obbes& response to the ool, *obbes is here endorsing the conditional strategy "nown to game theorists as :titA#orAtat.< 'itA#orAtat is the strategy o# responding to one&s #ellow player by always repeating what the other player does.-G An ob)ection will arise at this point, however, that titA#orAtat is unnecessary, and indeed irrational, in a situation in which common "nowledge o# rationality obtains. or i# one "nows with certainty what the other player believes to be rational and one also "nows that action to be rational, then one can per#ectly predict the other person&s conduct. In this case, cooperation will o#ten be rational, at least in situations in which there are mutual gains #rom cooperation. 8ut titA#orAtat would be an e!cellent strategy #or a situation in which there is less than per#ect certainty about the e!istence o# common "nowledge, and hence less than per#ect certainty about whether one is dealing with a #ellow cooperator. Arguably this is precisely the situation that obtains i# parties in a state o# nature are mostly, but somewhat inconsistently rational. I# one is uncertain +pdated: -./0.12 3age 42 o# 94 whether the other party is rational, it ma"es sense to re#rain #rom engaging him until one has evidence about the state o# his rationality. 'he problem with a bare e!change o# promises, then, is that one cannot be certain at the time one must render one&s commitment that the other party is rational and is prepared to behave cooperatively. 5ne would thus be obliged to commit onesel# to per#ormance in the absence o# "nowing the other party&s HtrueI intentions. 'his is true even i# one is contemplating being a second per#ormer in a #uture per#ormance contract. or in such a case, at the time ones commitment is made, one does not "now whether the other party is a cooperator It is true that one can always ma"e one&s

commitment conditional on the other party&s per#ormance, and thus establish that one&s obligation to per#orm on the contract will cease i# the #irst per#ormer #ails to act. ;et a conditional agreement is either a #orm o# #irst per#ormer contract H:I will pay you >@11 i# you climb to the top o# the #lagpole<I or a promise to do something in e!change #or the other&s promise. And *obbes& point is that in the absence o# "nowledge o# the other person&s rationality, it cannot be rational to ma"e a commitment based on the receipt o# a mere promise on the other party&s part. Clearly being the promise, or second per#ormer, in a unilateral contract stands on a di##erent #ooting. In this case, one can be assured that the #irst per#ormer is rational before one is obliged to ma"e a commitment. 5r rather, one&s commitment is conditional on the #act o# his actual cooperation, rather than +pdated: -./0.12 3age 43 o# 94 a mere promise o# cooperation. And thus the promisor in a unilateral contract is e!posed to no ris" by ma"ing a promise that is conditional on the #act o# the other party&s per#ormance. 'his, then, is the basis #or distinguishing #irst per#ormer #rom second per#ormer contracts in nature, and hence #or *obbes& willingness to e!empt second per#ormance on a unilateral contract #rom the general invalidity o# contracts in nature. 'he argument *obbes ma"es to the #ool can essentially be applied directly to the robber case to e!plain his views on this e!ample: i# it was rational #or you to renege on the promise to bring the money, the robber would "now it was rational #or you to do so, and would not have trusted you by releasing you. So the robber&s willingness to release you #rom his grip, in e!change #or your promise to return with the money, suggests that it is rational #or you to return with the money, in accordance with a plan that leaves both o# you better o## than you

would otherwise be.-- And notice, by the way, that the contract with the robber is a unilateral contract: the robber trades a present per#ormance on his part= releasing his grip on you=#or your promise to return in the #uture with the money. Since you are not the #irst per#ormer, it is rational #or you to "eep your end o# the bargain, given that there is no danger that you will not receive the bene#it under it that you were due. $hat about the arguments #rom baselines I considered in connection with $ertheimer&s and 4oJic"&s accounts? A baseline theorist would want to say that +pdated: -./0.12 3age 44 o# 94 the victim&s promise is invalid because it was the product o# an arti#icially lowered baseline that the coercer himsel# produced. 8ut #or *obbes, it is the archetype o# the rational e!change in a state o# nature that a person is able to reap the #ruits o# his domination over another when both begin in a position o# e(uality. $hatever concessions the threatening agent can e!tract #rom the other are #air game, and a stronger or more able agent should not be deprived o# the bene#its o# his superior abilities merely because they wor" to the disadvantage o# the wea"er party. or *obbes, then, the coercive, unilateral contract re#lects the returns to the per#ormer o# his superior position in the state o# nature. Since coercion supplies a right#ul basis #or gaining concessions #rom another in a state o# nature, and since it is o#ten rational to surrender power in the #ace o# coercive tactics in nature, the agreement with the latter in e##ect supplies the paradigm o# the binding agreement in nature. I# the other voluntarily agrees to some mutually bene#icial course o# action in order to mitigate the e##ects o# that e!ercise o# coercive #orce, the #irst per#ormer is entitled to demand the other&s #idelity to that agreement, since he, the #irst per#ormer, has cashed in his superior position to secure it. C. 7!ploring the *obbesian ramewor"

All that has )ust been said about *obbes and coercive contracts will no doubt only serve to hasten the speed with which most readers will be inclined to +pdated: -./0.12 3age 45 o# 94 dismiss a rational choice approach to such a topic, at least i# such approaches re(uire us to adopt a *obbesian stance. or modern readers=li"e *obbes& contemporaries=will re)ect a political philosophy that legitimiJes coercers ta"ing advantage o# their superior threat advantage. Indeed, it is usually regarded as one o# the #undamental advances o# modern contract doctrine that it recogniJes ine(uality in bargaining position as a basis #or voiding a contract. 'hus #ar #rom allowing individuals to claim a right#ul entitlement to the #ruits o# a superior bargaining position, modern legal scholars, as well as contemporary philosophers, will see the argument #rom domination as an additional reason to treat such contracts as void. urthermore, they will assert, *obbes& commitment to the #oregoing argument supplies all the more reason to (uestion his political philosophy across the board. 'hus *obbes& account o# contracts under coercion seems a microcosm through which to (uestion the entire *obbesian edi#ice. Still, let us consider care#ully whether this response is warranted. In particular, two #eatures o# *obbes& analysis give us reason to suspend )udgment and investigate #urther. irst, remember the plight o# the victim and her ardent wish to be able to suspend the duress de#ense in order to enter into a valid :coercive< contract. In de#ending coercive contracts, *obbes is not only granting the coercer permission to ta"e advantage o# his superior bargaining power. *e is also accommodating the victim&s pre#erences in a way that needs to be ta"en seriously #rom a rational choice perspective. Contemplating this aspect o# such +pdated: -./0.12 3age 46 o# 94 situations, even $ertheimer is #orced to admit: KIL# 8 would pre#er that A ma"e a #reedomAenhancing proposal, given a plausible understanding o# 8&s present situation, and i#, at

least prospectively, 8 would pre#er to be bound by his agreement, it is arguable that there should be at least a reasonably strong presumption that the agreement is not coerced and ought to be en#orced.-2 5# course $ertheimer puts the point by saying that in such cases there should be a presumption that the agreement was not coerced. 5# course we might as well say that the agreement was valid, though coercive, rather than use its validity to deny that agreements with these #eatures are coercive. A more serious problem with $ertheimer&s admission here, however, is that he #ails to notice that 8&s pre#erence to be bound by the agreement is not an occasional, outlying #eature o# coercive agreements. It is the hallmar" o# such agreements, at least those that result #rom soAcalled volitional coercion. 'he general structure I have identi#ied in the gunman scenario is a #ully generaliJable aspect o# coercive contracts: e! post the victim will always pre#er such contracts to be en#orceable. 7! ante he may or may not, and thus $ertheimer&s gloss to that e##ect does not strengthen his account. 8ut i# even a deontologically oriented theorist li"e $ertheimer can #eel the #orce o# the victim&s e! post wishes and their relevance #or e! ante decision ma"ing, we have reason to consider approaches that give such pre#erences their +pdated: -./0.12 3age 47 o# 94 due. 5# course this point does not, by itsel#, provide anything li"e an allAthingsconsidered argument in favor o# en#orcing coercive contracts. It is only one consideration that might be raised in support o# such an argument. $hat we should hope to #ind in the end is a way o# respecting the victim&s pre#erences, at the same time that we are able to bring *obbesian legal philosophy into synch with contemporary intuitions o# )ustice, and thus #ind a way o# )usti#ying, rather than embracing, the invalidity o# coercive contracts in the #ace o# the rational choice arguments. Second, recall that *obbes& views on this (uestion are meant to apply in

the absence of any law to the contrary. 'hus the entitlement created by coercive power is a natural entitlement only, one that persists in civil society only in the absence o# sovereign interdiction. $hat, e!actly, a *obbesian has to say about coercive contracts in civil society, and in particular what the possible )usti#ications #or #orbidding coercive contracts might be, re(uires more care#ul e!ploration. *obbes le#t us something o# a blac" bo! in that regard, and we need to e!tract the *obbesian position on coercion in civil society #rom other aspects o# his political and legal philosophy. It is to this tas" that I now turn. IO. '5$AR6S A C54'RAC'ARIA4 A33R5AC* '5 C57RCIO7 C54'RAC'S In the imaginary discussion with the robber, I argued that any e##ort to +pdated: -./0.12 3age 48 o# 94 ma"e a coercive agreement en#orceable by contractually suspending the duress de#ense is sub)ect to a #atal #law: since the agreement to suspend the ban on duress would itsel# be entered into under duress, there would be no basis #or e!empting it #rom the general rule that voids contracts under coercion. I started to consider whether matters might be di##erent i# the parties had an opportunity to address such an agreement prior to the coercive interaction. 9et us now return to that (uestion. $ould parties who are presently unaware o# whether they will be victims o# coercion, or whether they will want to e!ercise coercive power over others, decide to ma"e coercive contracts en#orceable? 5r would they choose to ban them, as current policy suggests we do? 3reviously I as"ed this (uestion in the isolated twoAparty case, and we imagined that you and the robber would answer it by prior agreement, without the bene#it o# "nowledge as to what your #uture positions might be. $e might now imagine that we have a group o# individuals choosing rules #or the establishment o# a legal system, #aced with a choice among di##erent possible regimes: would rational agents select a regime in which coercion is generally permitted, where gunmen are allowed to e!tract bene#its #rom others by #orce, where coercive

agreements are upheld, coercive trans#ers o# property respected, coercively sought se!ual #avors permitted, and so on? 5r would they pre#er a regime in which coercion is re)ected in all o# its uses and mani#estations=where the state sought to punish those who attempted to gain #rom the private threat o# #orce, and where +pdated: -./0.12 3age 49 o# 94 it deprived those who did resort to coercion o# the bene#it o# those e##orts, despite the #act that they were merely e!ploiting natural advantages? It is not hard to see that even on *obbesian premises, the regime that permitted coercion would be ruled out, as it would constitute a nearA certain reversion to a state o# war. And this gives us a basis #or thin"ing that *obbes can be rescued #rom the perils o# moral per#idy. *e can simply say that although agreements with robbers are binding under natural principles, the sovereign has good reason to #orbid the use o# coercion in civil society. 'his would s(uare with *obbes& remar" that where :the thing promised is contrary to the laws,<21 #or e!ample where :#orce is prohibited,< the actions used to induce agreement are impermissible, and so the agreement would be void.2@ $ill this (uic" and dirty solution prove an ade(uate route bac" to common sense #or a *obbesian? In my view it will not, at least not without #urther elucidation. *obbes& implicit argument here appears to be that i# the sovereign has good reason to illegaliJe coercion, it #ollows that he has good reason to bar coercive contracts. 8ut the suggestion depends upon *obbes& con#lation o# two (uite di##erent (uestions that must, at least in the #irst instance, be "ept apart: the (uestion o# the illegality o# the content o# a promise, versus the (uestion o# the illegality o# the actions used to induce the promise.2/ In principle, we might have varying responses to these di##erent types or sources o# illegality. 5ne can imagine a situation in which the conditions creating the coercive e##ect are not +pdated: -./0.12 3age 50 o# 94

illegal HI threaten to brea" o## my #riendship with you unless you promise to do suchAandAsuchI, but the content o# the promise is Hyou must promise to do a drug run #or meI. Alternatively, the conditions creating the coercion might be illegal HI threaten to "ill you unless you do suchAandAsuchI, but the demanded action is not Hyou must ma"e a deposit into my ban" accountI. And i# this is true, we can see that *obbes has not thoroughly considered the implications o# his remar"s about coercion #or actual legal regimes. or it might be possible to #orbid the use o# coercion but nevertheless en#orce contracts #lowing #rom that coercion. 'he #oregoing conclusion is bad news #or *obbes. or it implies that rational agents are not in #act #orced to choose between a regime in which coercion is permitted in every conte!t and one in which it is prohibited across the board. 'here may be many more possibilities #rom which *obbesian agents can choose. And i# they have a greater array o# options, and i# the rational incentives are as I suggested, then *obbesian agents might well pre#er a regime that en#orced at least some coercive contracts. 'he simple argument I presented above about reversion to a state o# nature would then be unavailable to a *obbesian. In particular, there is an intermediate regime that might prove attractive to *obbesian agents: illegaliJe the initial use o# coercion, attempt to obtain all re(uired deterrent e##icacy #rom that interdiction, yet nevertheless uphold agreements based on the use o# illegal coercion. Such an approach is at least a logical possibility. An analogy might be the IRS&s position on income #rom +pdated: -./0.12 3age 51 o# 94 illegal activities: it is illegal to ma"e money by smuggling drugs, but the duty to pay ta!es on money so earned still e!ists, and it is a separate o##ense to #ail to pay ta!es on such earnings. Another e!ample might be the relationship between trespassing and adverse possession: it may be illegal to trespass on someone&s

land, but i# one does so consistently Hopen and notoriouslyI #or twenty years, one ac(uires an easement to the land where one #ormally trespassed. It would clearly be possible, though not necessarily desirable, #or a state to illegaliJe the use o# coercion and still en#orce contracts that emerge #rom illegal methods o# bargaining.2B 6oes it seem li"ely that the individuals attempting to ma!imiJe their own utility in the conte!t o# bargaining over an initial social contract would end up #avoring a mi!ed regime o# this sort=the regime that outlaws coercion but en#orces coercive contracts once coercion is nevertheless illegally used? Such a conclusion seems possible on *obbesian premises. 5utlawing the use o# coercion is a per#ectly sensible thing #or a sovereign to do by *obbes& lights, as it would discourage the use o# coercion and thereby ma"e individuals more secure. And outlawing only the initial use o# coercion preserves the victim&s ability to buy her way out o# coercive encounters, which pre#erences we are bound to respect. In particular, i# the primary coercion can be separately outlawed, there seems no )usti#ication #or invalidating :voluntary< agreements that represent its downstream e##ects i# the only reason to do so is to #urther discourage instances o# primary +pdated: -./0.12 3age 52 o# 94 coercion. $e can now connect this point with my original suggestion about the pre#erences o# the victim. 4otice that invalidating coercive contracts #orces the victim o# coercion to pay #or some o# the costs o# #uture deterrence, as opposed to keeping all deterrent costs firmly on the perpetrator. 'his is because the victim is now deprived o# the chance to save hersel# by stri"ing a deal with the robber, in order to deter #uture coercers #rom committing similar acts. 8ut arguably it would be more e(uitable, as between coercer who is in the wrong and victim who has done nothing wrong, to put all the burden o# deterrence on the coercer and sharply

increase his penalty #or the initial use o# coercion, thus e!empting the victim #rom having to pay with her li#e #or the bene#it o# overall social utility. And as arguably i# such an arrangement would be more e(uitable, it would be more li"ely to be chosen by the parties as their pre#erred social arrangement, given that they do not "now their positions ex ante. Someone might ob)ect to this conclusion on the grounds that the coercer also gains i# the agreement with his victim is en#orceable. 8ut that bene#it can be overcome by simply #urther increasing the punishment #or the initial coercion, since there is no particular di##iculty with piling on ever increasing penalties #or the initial coercive act.2C So to be sure, i# coercive contracts are en#orced, the penalties #or coercion will need to be increased. 8ut this hardly seems an ob)ection, given all that has been said. It was then this mi!ed response that +pdated: -./0.12 3age 53 o# 94 *obbes did not, but should have, considered given other aspects o# his philosophy. *obbes may now, however, be bo!ed into a corner. I argued previously that *obbes is not as out o# synch with common moral intuitions as one might have supposed upon #irst hearing his suggestion that the agreement with the robber is binding in a state o# nature, #or it is clear that he is prepared to invalidate coercive contracts i# there is reason to illegaliJe coercion generally in civil society. I then noted, however, that *obbes appears to have missed the #act that coercion can be separately punished, and so it is in theory possible to en#orce a contract that results #rom an illegal use o# coercion. And in noting this possibility, *obbes is deprived o# an argument #or the common sense solution, which invalidates a contract i# the means o# obtaining it are themselves illegal. 6espite the #act that en#orcing the coercive contract in many situations has its attractions

#rom the standpoint o# the victim, then, it will seem to many a reductio o# the *obbesian approach to legal (uestions i# it ends up endorsing contracts with robbers entered into under duress. 6oes *obbes have a di##erent way out, other than that we considered above? 9et me step bac" #rom the gunman scenario #or the moment and return to this e!ample a bit later. I will be in a better position to address it i# I #irst consider a situation in which en#orcing a coercive, or at least (uasiAcoercive contract is less ob)ectionable than it appears to be in the gunman case. I turn instead to the +pdated: -./0.12 3age 54 o# 94 related, but milder cases o# contract renegotiation. O. %5R7 39A+SI897 A339ICA'I54S: 7D395I'A'I54 A46 '*7 3R7A7DIS'I4E 6+'; R+97 'he soAcalled :preAe!isting duty rule< in contract doctrine maintains that a promise or per#ormance is not valid consideration #or a contract i# the party o##ering it already had a preAe!isting duty to do the thing promised or per#ormed.20 'hus I cannot o##er my boss my willingness to use best e##orts in e!change #or increased salary i# using such e##orts was already among the conditions o# my employment. 4or can I demand more pay in e!change #or a )ob I have already contracted to do, i# doing that )ob was already an obligation I had ta"en on #or lesser pay under a previous contract. In such cases, standard doctrine has it, the new promise or per#ormance is invalid because it is barred by the e!istence o# previous consideration.2F 'he preAe!isting duty doctrine has traditionally been e!plained in terms o# the doctrine o# consideration: a commitment to do something one already has a duty to do cannot supply the consideration #or a new obligation on the other party&s part because there is nothing new to be e!changed. 'he doctrine o# consideration is an e!pression o# the re(uirement o# e!change. 3arties cannot bind themselves to a promise unless their promise is o##ered in e!change #or the

promise Hor per#ormanceI o# another, the bene#it o# which the party o##ering the +pdated: -./0.12 3age 55 o# 94 promise hoped to obtain. 'he e!planation most #re(uently o##ered #or the preAe!isting duty rule, however, does not have to do with consideration but rather with e!ploitation.2G 'he concern is that without the rule, one party would be in a position to e!ploit the other midway through per#ormance and demand increased wages or other bene#its. In the absence o# the rule, the other party would be #orced to agree to the increase, as he would #ind himsel# in a particularly di##icult situation i# deprived o# all means to induce the other to complete his wor". It would seem, then, that having a preAe!isting duty rule helps to avoid a :holdAup< problem that would occur in many contractual relations. 'he real rationale #or the rule there#ore, has to do with coercion, rather than consideration.2Contracts scholars, however, have noted that this rule can wor" to the disadvantage o# both parties.22 'hus in the #amous Alaska ackers case, #ishermen threatened to wal" away #rom their employment at a cannery in the middle o# a #ishing season, unless they received higher wages than originally agreed.@11 In response to the threat o# losing its wor"#orce midseason the cannery promised to pay the higher wages, and the #ishermen remained on the )ob. 8ut when the cannery #ailed to pay on the renegotiated contract at the end o# the season, the #ishermen sued, and the court ruled that the cannery&s subse(uent promise was barred by the prior duty the #ishermen had to wor" #or their original wages. 'he renegotiated contract was #ound to be barred by the preA e!isting duty +pdated: -./0.12 3age 56 o# 94 rule. Suppose, now, that both sides had been more aware o# the preA e!isting duty rule than it appears they were. $ould either side have been pleased to "now at the time o# renegotiation that any new agreement they reached would li"ely be

barred in this way? As in the robbery situation, it is clear that both parties would have pre#erred not to have a preAe!isting duty rule at the time o# renegotiation: the #ishermen clearly would have pre#erred that, since they were hoping to obtain a commitment #rom the cannery to pay higher wages. 8ut the cannery too=li"e the victim in the robber case=would also have pre#erred to be able to renegotiate the contract, given that in the absence o# credible renegotiation, the #ishermen would have wal"ed o## the )ob. So )ust as in the actual holdAup situation, the #ishermen and the cannery might both pre#er not to be barred #rom renegotiating a binding contract at the moment o# renegotiation. 4otice that this answer depends on several conditions being satis#ied. irst, as 5ren 8arAEill and 5mri 8enAShahar have emphasiJed, the threat to breach the contract must be a credible one.@1@ I# the #ishermen had threatened to breach the contract and the cannery had reason to doubt their sincerity, the cannery might have pre#erred to have the support o# the preAe!isting duty rule in re#using to accede to their demands. Second, there must be the legal e(uivalent o# the common knowledge of rationality, which we might re#er to as the common knowledge of legality. +nder this condition, both parties have "nowledge o# the +pdated: -./0.12 3age 57 o# 94 prevailing legal rules, and each has "nowledge o# each other&s "nowledge o# such rules. +nder these conditions, in the #ace o# a sincere threat to breach the contract, the cannery would pre#er not to have a preAe!isting duty rule than to have it, as long as it can be assumed to be clear to both that the renegotiated contract will be binding. It is, a#ter all, that advance "nowledge on which the cannery relies, and any uncertainty on the #ishermen&s part will undermine the cannery&s ability to convince the #ishermen o# the sincerity o# its promise to pay. 5# course, the #act that the parties at the moment of renegotiation would

want to abandon the preAe!isting duty rule is neither surprising nor conclusive, given what we have seen in the case o# the robber. 'he relevant (uestion instead is whether they would have pre#erred to leave open the possibility o# contract renegotiation at the time o# contracting to allow #or revision, even in the absence o# changed circumstances. A rational conclusion would be that neither side would wish to ta"e the ris" o# being sub)ect to the other party&s e!ploitation, and would want to guard against it. 8ut consider the ris"s on the other side: there is the considerable danger o# being on the receiving end o# a credible breach o# contract and having no means o# convincing the other party to ma"e good on its promise.@1/ So the ris"s associated with not being able to engage in contract renegotiation may in the #inal analysis seem greater than the ris" o# e!ploitation, and both sides are li"ely to choose not to include the preAe!isting duty rule in their initial contract. +pdated: -./0.12 3age 58 o# 94 In the case o# contract renegotiation in the #ace o# a sincere threat o# breach, there is a ma!imally strong argument #or en#orcing the contract, despite the #act that the contract emerges #rom coercive, or e!ploitative tactics. 8ecause o# cases li"e this, legal scholars are increasingly calling #or the elimination o# the preAe!isting duty rule.@1B Such scholars recogniJe the ris" that those with greater bargaining power may try to e!ploit their position in the absence o# the rule. ;et most scholars who propose eliminating the preAe!isting duty rule predicate their suggestion on the e!istence o# another important bac"ground legal rule: the de#ense o# duress. $e could there#ore eliminate the preAe!isting duty rule as long as we retain the duress de#ense, or so they say.@1C 'he suggestion o# such scholars thus seems to be that we should give e##ect to coercively obtained contracts in a renegotiation situation, as long as the coercion is not too e!tensive. 5nce the

coercion rises to the level o# true duress, we should continue to punish the coercer by invalidating the resulting contract. $hile the intuition is a sound one, there is a problem with this approach, as I have already implicitly shown: the duress de#ense is sub)ect to the same argument that compels the above scholars to abandon the preA e!isting duty rule: both parties might see themselves as better o## and neither worse o## under particular agreements the de#ense #orbids. 6uress is in this regard indistinguishable #rom the preAe!isting duty rule, and the two appear to be separated only by the degree o# coercion at issue. Rust as the coercion e!ercised +pdated: -./0.12 3age 59 o# 94 by the #ishermen against the cannery does not eliminate the cannery&s desire to enter into a binding renegotiated contract #or higher wages, so the robber&s coercive tactics against his victim do not eliminate the victim&s desire to bind himsel# to an en#orceable agreement to pay the robber to release him. 'he coercion on which the duress de#ense is predicated, in other words, does not eliminate the pre#erence #or binding commitments. 5n the contrary= it appears to increase it. So while it seems in many circumstances as though the renegotiation should be allowed, the argument #or eliminating the preAe!isting duty rule is sub)ect to the reductio that it would result in en#orcing deals between robbers and their coerced victims. $hat we need to de#end the common sense solution, then, is a way o# distinguishing the gunman situation #rom the renegotiation conte!t. 'hat is, we need a way o# preserving the duress de#ense in the #ace o# a sensible re)ection o# the preAe!isting duty rule. 9et us return to the suggestion o# the mi!ed regime considered above. Recall that this suggestion seemed attractive #rom the standpoint o# rational choice considerations, since it would achieve the deterrent bene#its hoped #or by invalidating coerced contracts, at the same time that it would relieve the victim

#rom bearing some o# the costs o# that deterrent e##icacy. It is also clear, however, that the suggestion is li"ely unacceptable as an actual, administrative response to the gunman scenario, and so its availability merely served in my argument to detract #rom, rather than enhance, the attractiveness o# the *obbesian approach. +pdated: -./0.12 3age 60 o# 94 'he same solution, however, is vastly more appealing in the conte!t o# the pree!isting duty rule. 'hat is, we could (uite sensibly en#orce the renegotiated contract between the cannery and the #ishermen, yet sanction the #ishermen independently #or any improper or coercive tactics they might have used in gaining the agreement o# the cannery. *ow would this "ind o# solution be implemented? A court would #irst have to determine whether the #ishermen had indeed behaved coercively in threatening to wal" o## the )ob. I# they were responding to a real economic necessity=#or e!ample, they were underpaid and discovered they could bene#it signi#icantly by ta"ing employment elsewhere=it is less li"ely that the threat to breach the contract would be seen as coercive. I# the threat to breach the contract were a blu## or otherwise poorly motivated, a court might #ind coercion, and set a penalty. 'he siJe o# the penalty would be a re#lection o# the degree o# coercion the court believed was involved. 'hus although the renegotiated contract would be allowed to stand, since it was agreed to voluntarily, the #ishermen&s award might be reduced by the amount o# coercion they used to get the cannery to o##er that agreement. In the event o# e!treme coercion, there would be no return to the #ishermen, and might indeed be a #urther penalty Hwhether civil or criminalI e!tracted as punishment #or improper behavior. 4ow let us compare this bi#urcated regime in the renegotiation conte!t to +pdated: -./0.12 3age 61 o# 94 the same approach in the gunman scenario. 'he problem is that since it is

possible to punish the gunman separately #or his coercion=we have a distinct crime, armed robbery, that is separately prohibited and so can be separately punished=why would we not be willing to say that the deal with the gunman should always stand, and that the general deterrence we re(uire could in this instance be separately obtained by e!tensive punishment o# the original coercion?@10 Indeed, at least one pair o# legal scholars thin"s no appropriate distinction can be drawn. 5ren 8arAEill and 5mri 8enAShahar have written that whether a coerced agreement should be en#orced should turn on the credibility o# the threat o# the coercing party.@1F A threat is credible i# the coercer&s payo## #or carrying out the threat e!ceeds the payo## #rom not carrying the threat out.@1G 'he idea, li"e that pointed out above, is that the possibility o# an e! post remedy induces the coercer to carry out the threat since their payo## would be greater. $hen the coercer is blu##ing, however, an e! post remedy could not induce the blu##ing coercer to carry out the threat, given that they presumably saw threate!ecution as not worthwhile in the #irst place.@1- Conse(uently, 8arAEill and 8enA Shahar argue that whether an e! post remedy should be available #or a coerced party should turn only on whether the threat was credible or a mere blu##. 'hat 8arAEill and 8enAShahar reach such counterAintuitive results is obviously problematic #or their account. $hat is perhaps more problematic #or them, however, is that the criterion o# credible threat does not distinguish the +pdated: -./0.12 3age 62 o# 94 serious case o# the gunman #rom the less serious case o# the Alaska ackers.@12 'his is because in all cases the credibility criterion is a :descriptive understanding o# the threatening partySs incentives,< which determine whether the :ideal outcome< is attainable.@@1 Credible threats, as well as blu##s, are e(ually possible in all cases o# coercion, regardless o# severity. It loo"s as though the more severe the coercion the more li"ely we will be

able to separate it #rom the resulting contract, and the more appropriate separate punitive measures would be, and hence the more )usti#ied we would be in en#orcing the coerced contract? 8ut this is the reverse o# what intuition would tell us, since we are more willing to en#orce the resulting contract in a mild case o# coercion li"e Alaska ackers than in a deal negotiated at gunpoint. In order to re)ect this logic, let us consider whether there is not a way o# distinguishing the logic o# the gunmanAtype scenario #rom the preAe!isting dutyAtype case, such that we can reach the more intuitive result, namely to employ the bi#urcated approach in the latter case but not in the #ormer. 'here is one natural way to distinguish the cases, which I highlight only to re)ect. 5ne might be tempted to distinguish the cases in terms o# the severity o# the threat. 'hus, the #ishermen in Alas"a 3ac"ers are threatening to breach an antecedent contract between themselves and the cannery, where the relevant contract is one made between the :aggressor< and the :victim< at an earlier moment in time. 8y contrast, the robber threatens what would appear to be a +pdated: -./0.12 3age 63 o# 94 much greater wrong than mere breach o# contract: he threatens to "ill the victim. Since "illing is HusuallyI worse than the ordinary violation o# a civil employment contract, we would ideally li"e to distinguish these two threats in terms o# degree of wrongfulness, and say that the contract is disallowed in the gunman case because the threat #rom which it results is worse. 8ut degree of wrongfulness would appear to return us to the deontological approach I re)ected on rational choice grounds at the outset. $e must there#ore #ind a di##erent way. Is there another basis #or distinguishing the two types o# threats? 5ur problem, as is now clear, is that the two cases are structurally highly similar, indeed more similar than might #irst be thought. or the wrong#ulness o#

un)usti#ied "illing in a *obbesian system is also a #unction o# contract violation, albeit one made not with the victim alone, but with all o# society. In a *obbesian system, the way to e!plain the wrong#ulness o# "illing is that unprovo"ed or nonde#ensive intentional "illing violates the very terms o# the social contract itsel#. It will there#ore not wor" simply to say that "illing is worse than contract violation and to distinguish the cases that way, since the wrong#ulness o# "illing is itsel# inseparable #rom the notion o# contract violation. OI. '*7 S5CIA9 C54'RAC' A46 C57RCI54 'he observation that the robber&s threatened action violates the terms o# the social contract should now allow us to articulate a crucial distinction between +pdated: -./0.12 3age 64 o# 94 the Alaska ackers case and the case o# the deal with the robber. $hile it may be open to the victim to waive the terms o# a contract that he has made with the coercer in the preAe!isting duty conte!t, the victim is not #ree to waive the contract violation where all other members o# society are also party to the agreement. or in doing so, the victim would undermine not only his own entitlement to insist on #idelity to mutually bene#icial e!changes, but the right o# everyone else to do the same. 'hus what the robber is threatening to do, should the victim #ail to agree to the o##ered contract, is to violate the de#ensive conditions o# a contract entered into with all other members o# society, and this is something the victim is not entitled to do. $aiving de#ensive rights o# the social contract would render the position o# all insecure by eliminating the protections they have chosen #or their agreement, and that it is not the prerogative o# any given individual. 'his o# course e!plains why matters would be di##erent in a state o# nature. 'here the robber&s threat is not :wrong#ul,< since there is no social contract the robber is threatening to violate. And this is )ust to say that the robber can use the

bene#it o# his threat advantage in a state o# nature, but not in civil society. 'he point o# the social contract a#ter all, is to rule out precisely that sort o# threat advantage. An ob)ection might be raised that the robber is only threatening to breach the contract=he is not actually breaching the contract, unless the victim #ails to +pdated: -./0.12 3age 65 o# 94 comply with his re(uest. $hile there are di##erent positions among philosophers about the relationship between threats and actions in #urtherance o# such threats, it is clear #rom the literature that it is a per#ectly reasonable position to suppose that threatening to violate a contract is the same wrong#ul act as actually violating it. Admittedly, threatening to do what is wrong#ul might be its own "ind o# wrong, distinguishable #rom the wrong o# the threatened action. It might, #or e!ample, be a wrong o# a lesser order, related o# course to the wrong that is itsel# threatened, but still a wrong o# a di##erent, and probably lesser, signi#icance. 8ut I have not argued that threatening to violate the contract is itsel# a violation o# the social contract, and it is indeed unclear whether such a wrong should be ran"ed as among the wrongs that the parties would be concerned to protect themselves against in the basic de#ensive social covenant. So the #ocus should probably not be on the potential wrong#ulness o# the threat itsel#, but rather on the wrong#ulness o# the action being threatened. And the (uestion about that action must be whether, by entering into an agreement to pay the robber #or his suspension o# violence, the victim is e##ectively releasing the robber #rom his commitment to the social contract. or this is what I have claimed the victim may not do. It is reasonably clear that he is. irst, there is the relatively straight#orward #act that the robber will surely not enter the agreement unless it contains an immunity #rom prosecution in addition to the promise to pay. or i# +pdated: -./0.12 3age 66 o# 94

the robber is to be prosecuted, he will almost certainly #ail to reap the gains he was hoping #or under the agreement. And to as" #or immunity #rom prosecution is e##ectively to as" #or an e!emption #rom the conse(uences o# violating the social contract, the possibility o# which we can understand the robber has already re)ected. 7ven the victim does not have the power to grant immunity #rom prosecution. Criminal prosecution is a public #unction, a #act that underscores the idea that it is a wrong o# a di##erent sort that is being threatened here. 'he wrong o# "illing the victim is not only a wrong to the victimM it is a wrong to society as a whole, and there#ore society retains the right to prosecute it. 'his right=the right o# prosecution under the social contract=is one the victim is powerless to waive. 'hus the rational choice analysis o# contracts under coercion suggests leaving the invalidation o# contracts intact in cases in which the initial contract is entered into under coercion, but allowing the contract to stand where the coercion is employed in the conte!t o# contract renegotiation. 'his approach would honor the #act that individuals acting under voluntary coercion are not literally deprived o# choice. 'hey actually do e!ercise choice, and there are surely many cases in which their wel#are may depend on their ability to enter into the very bargain they would later li"e to invalidate. 8ut because this suggestion would have untoward e##ects on the social contract in cases o# true coercively induced initial agreements, we invalidate the contract in all such cases, despite the immediate bene#its to the victim #rom bargaining with gunmen. +pdated: -./0.12 3age 67 o# 94 OII. A +R'*7R '7S': 397A 8AREAI4I4E AS C54'RAC' 9et us turn now to a di##erent application o# the bi#urcated approach to coercive contracts, one that appears in the law en#orcement setting. %uch has been written in recent years about the voluntariness o# agreements reached as part o# a plea bargain. And indeed many legal scholars have ob)ected to plea bargains

on the ground that such agreements are the product o# coercion. @@@ 'hey maintain that i# we were applying ordinary contract principles to these cases, plea bargaining, agreements would rightly be void. or e!ample, a prosecutor who threatens a criminal de#endant with prosecution #or capital murder to gain the de#endant&s agreement to a twenty year sentence is thought to be li"e the gunman in the initial e!ample: the de#endant must e##ectively agree to plead guilty and accept the sentence under pain o# worse alternatives. 'he choice he gives the de#endant is said to be no choice at allM hence the resulting plea cannot be voluntary, it is argued. Rust as I have shown in the case o# ordinary agreements and contracts, this description o# the agreement between prosecutor and de#endant re(uires some care#ul parsing. 'he plea bargain, unless physically #orced or agreed to under torture, is surely #ully volitional in the sense discussed earlier. And lawyers do not have an alternative de#inition o# voluntariness to put in its place. According to Supreme Court )urisprudence on this (uestion, #or e!ample. a plea is :involuntary< when :the de#endant&s will is overborne,<@@/ when the de#endant is +pdated: -./0.12 3age 68 o# 94 not in :possession o# his mental #reedom,<@@B when :the accused was deprived o# his #ree choice to admit, deny, or re#use to answer,<@@C when :the choice is between the roc" and the whirlpool,<@@0 and so #orth. 8ut it seems substantially more li"ely that the relevant #eature o# such situations that courts are trying to capture has nothing to do with lac" o# voluntariness, as I argued earlier in the case o# agreements o# other sorts. 'he reason that the )urisprudence o# plea bargaining has developed this con#used notion o# voluntariness is clear enough. Courts, li"e lawyers and philosophers, ma"e the mista"en assumption that unless we are prepared to say

that a person gave his assent to an agreement involuntarily! we will have no basis #or voiding the agreement. 'hus the move is made to posit a psychological condition under which the de#endant is not responsible #or his choices, and then to label the agreement made in that condition :coercive,< in order to attac" its en#orceability. 8ut the #laws in this line o# reasoning should by now be clear, given what we have seen in the case o# private contracts. irst, there is no reason to suppose that we must treat an agreement as involuntary in order to call it coercive. Second, the )urisprudence o# plea bargaining assumes that unless a plea is dubbed :coercive,< we would have no basis #or re)ecting it. 8ut as in ordinary contracts, neither o# these assumptions is )usti#ied. Indeed, in some cases, we may well wish to say that the de#endant was coerced into accepting a plea bargain, without either regarding his assent as involuntarily o##ered or +pdated: -./0.12 3age 69 o# 94 invalidating the plea. In particular, suppose we generally regard pleas as acceptable, but we wish to invalidate them when they are produced by certain tactics or methods on the part o# prosecutors. And we believe they should be invalidated in such cases, despite the #act that we regard them as #ully voluntary. In this case, voiding the agreement is unli"ely to be an e##ective deterrent o# coercive tactics on the prosecutor&s part, because the prosecutor will #ind himsel# in precisely the situation he would have been in had he never initiated negotiations with the de#endant: he will have the choice to bring the case to trial or to drop it, or perhaps to e!plore new pleas that are li"ely to be upheld. 8ut he is not personally penaliJed #or his improper behavior, at least in most cases. 'he bi#urcated approach to coercion might be particularly appropriate under this set o# circumstances: a de#endant who was coerced into a plea agreement could be held

to that agreement, i# his guilt was reasonably supported by additional evidence, other than the plea. At the same time, however, the prosecutor could be separately sanctioned #or the use o# coercive tactics. $hile there are o# course criminal sanctions available #or prosecutors and police o##icers who go far beyond the limits o# their o##ices, the threshold #or sanctioning police and prosecutors #or improper techni(ues #or achieving pleas is set (uite high as matters presently stand. 'he bi#urcated approach would recommend setting the threshold substantially lower, but sanctioning prosecutorial misconduct directly.@@F +pdated: -./0.12 3age 70 o# 94 In conclusion, it is worth noting that there is a crucial, substantive reason to be concerned about the practice o# plea bargaining that has nothing to do with voluntariness or coercion, and this is the threat that plea bargaining poses to the presumption o# innocence. $hat is worrisome is that de#endants are induced by the practice o# plea bargaining to abandon their constitutional right to prove their innocence by the #ear o# what will happen i# they are unable to succeed at trial. 6etractors o# the practice once again have a tendency to say that de#endants are :coerced< into giving up their right to a )ury trial. 8ut o##ers o# plea agreements need not be thought any more coercive than many attractive o##ers might be. urthermore, we can ob)ect to the inducement the plea provides without calling the resulting agreement :coercive,< or, alternatively, without see"ing to invalidate it #or this reason. 'he ob)ection may simply be that we do not want de#endants giving up their right to trial by )ury for the wrong reasons. In particular, we do not want innocent de#endants to be induced to give up that right, since their reasons #or giving up the right surely do not advance the values the criminal )ustice system see"s to endorse. I will not e!plore this line o# argument #urther on the present occasion, as

it is not directly relevant to my immediate concerns. $hat should be clear is that in the plea bargaining conte!t, as in other areas o# the law, an important, substantive ethical and political concern has been ob#uscated by the language o# involuntariness and its mista"en association with coercion, as well as by +pdated: -./0.12 3age 71 o# 94 con#usion about the implications o# labeling an agreement or an action as coercive. 'he motivation #or the ob#uscation is the di##iculty itsel# o# reaching desired ethical conclusions with straight#orward moral reasoning. 5nce we recogniJe that the mista"en psychologiJing o# coercion is partially motivated by ethical concerns, it should help us to disentangle the ethical #rom the psychological and start a#resh. OIII. C54C9+SI54 In this Article, I have sought to e!pose a tension between the perspective o# a victim in cases o# coerced contracts and the usual )usti#ications that are o##ered #or invalidating such contracts. In particular, I have suggested that the victim in such cases would regard re#using to en#orce such contracts as a "ind o# double insult to her wellAbeing: #irst the coercer lowers her baseline wel#are by #orcing her to choose between her li#e and her economic wellAbeing, and then the state lowers it even #urther, by re#using to allow the victim to ma"e the binding commitment that would enable her to ma"e the best o# her remaining options. I have also argued that an ade(uate )usti#ication #or a social policy o# this sort would re(uire that the policy be )usti#ied to the victim, in terms that the victim can endorse and regard as #air to her interests. Such )usti#ications would be contractarian in nature, and would di##er substantially #rom the types o# )usti#ications typically on o##er, which tend to be moraliJed or conse(uentialist. +pdated: -./0.12 3age 72 o# 94 In the attempt to discover a )usti#ication #or barring coercive contracts, I

noted that such contracts strongly resemble a situation involving the use o# the preAe!isting duty rule, in which parties attempt to renegotiate contracts under threat o# breach #rom the other party. 5#ten they discover that they cannot engage in such renegotiation because it would be barred by the other side&s preAe!isting duty to meet his contractual commitments. In such cases, intuition strongly supports allowing the parties to engage in binding renegotiation, despite the pree!isting duty rule. And this intuition is strongly supported by the contractarian thought that it is un#air to bar the victim to buy her way out o# a coercive situation by ma"ing binding agreements under the pressure o# some adverse threat. I# the contractarian argument supports en#orcing agreements reached under threat o# breach o# contract, despite the :coercive< aspect o# such agreements, why would it not support en#orcement o# contracts made with gunmen? As I noted, at least some contracts scholars in the economic tradition have reached the conclusion that the two situations are identical in this respect. And it is true that #rom an economic perspective, there is little to distinguish the cases. 8ut intuition spea"s strongly in #avor o# trying to draw such a distinction, and as it happens, such a distinction is supported by contractarian re#lection. I have argued that it ma"es sense to thin" that the rights that are being waived in the preAe!isting duty case are rights the victim entirely controls: the norm being violated in that case is the contractual commitment that the +pdated: -./0.12 3age 73 o# 94 perpetrator previously made to the victim. In the gunman hypothetical, however, the norms being violated are general bac"ground norms having to do with the collective commitment to the inviolability o# human li#e. 'hese the victim is powerless to waive. I there#ore conclude that a contractarian approach can plausibly distinguish cases o# coercion #rom one another, and that in some

instances, the #act that the contract is entered into under duress, or (uasiAduress, does not entail its invalidation. In others, however, the e!istence o# duress does have that implication, and there the traditional wisdom that coercive contracts are invalid is precisely right. As we have seen, however, this is not #or the reasons that traditional analysis would once have provided. +pdated: -./0.12 3age 74 o# 94 4otes @ 'he "estatement #$econd% of Contracts T @G0 de#ines duress as #ollows: I# a partySs mani#estation o# assent is induced by an improper threat by the other party that leaves the victim no reasonable alternative, the contract is voidable by the victim. American 9aw Institute, "estatement #$econd% of Contracts, T @G0 Hemphasis addedI. Compare this with the &odel enal Code&s de#inition, establishing duress as a de#ense to a crime: It is an a##irmative de#ense that the actor engaged in the conduct charged to constitute an o##ense because he was coerced to do so by the use of! or a threat to use! unlawful force against his person or the person o# another, that a person of reasonable firmness in his situation would have been unable to resist. American 9aw Institute, &odel enal Code, T /.12 Hemphasis addedI. 'he criminal law is concerned with coercion through unlawful #orce rather than an improper threat. 'he class o# improper threats is clearly larger than the class o# unlaw#ul threats. %oreover, the &odel enal Code&s #ocus on the ob)ective standard o# the :person o# reasonable #irmness< is distinct #rom contract&s criteria o# :no reasonable alternatives.< It is unclear, however, whether there is divergence between the two criteria. 3resumably, one way to assess the +pdated: -./0.12 3age 75 o# 94 reasonableness o# one&s alternatives is to thin" about whether a person o# reasonable #irmness would have no alternative. / 6e#ault rules are rules that courts will use to interpret contracts unless the parties have speci#ied to the contrary. %andatory rules, by contrast, are rules o# contract #ormation that the parties are not #ree to ignore. See, e.g., 7. A.

arnsworth, Contracts, Cth ed. HAspen, /11CI, T G.@0A.@G. B See Dickinson v. Dodds, 9.R. / Ch. 6. CFB, CG/ H@-GFI H:It must, to constitute a contract, appear that the two minds were at one . . . .<I HRames, 9.RI. C :'he rules on duress . . . allow the in)ured party to undo the transaction by avoiding it.< arnsworth, Contracts, T C.@2. 0 Charles ried, Contract as romise' A (heory of Contractual )bligation HCambridge, %A: *arvard +niversity 3ress, @2-@I, 2-A22 H:I conclude that a promise procured by a threat to do wrong to the promisor . . . is without moral #orce.<IM Alan $ertheimer, Coercion H3rinceton: 3rinceton +niversity 3ress, @2-GI, B H:A coerced promise is not morally binding.<I. F See, e.g., Roel einberg, *arm to $elf H4ew ;or": 5!#ord +niversity 3ress, @2-CI, @-2A/F-M *arry ran"#urt, :Coercion and %oral Responsibility,< in +ssays on ,reedom of Action, ed. '. *onderich H9ondon: Routledge and Kegan 3aul, @2GBI, FBA-FM Charles ried, Contract as romise' A (heory of Contractual )bligation HCambridge, %A: *arvard +niversity 3ress, @2-@I, 2BA2-M 6aniel +pdated: -./0.12 3age 76 o# 94 9yons, U$elcome 'hreats and Coercive 5##ers,U hilosophy 01 H@2G0I: C/0M Robert 4oJic", :Coercion,< in $ocratic u--les HCambridge, %A: *arvard +niversity 3ress, @22GIM $ertheimer, CoercionM 6avid Vimmerman, UCoercive $age 5##ers,U hilosophy . ublic Affairs @1 H@2-@I: @/@. G See, e.g., 5ren 8arAEill and 5mri 8enAShahar, UCredible Coercion,U (exas Law "eview -B H/110I: G@GM Richard A. 3osner, +conomic Analysis of Law, Cth ed. H8oston: 9ittle, 8rown, @2G/I, @@0. or parallel implicative discussions o# duress in the substance o# the criminal law, see Claire in"elstein, U6uress: A 3hilosophical Account o# the 6e#ense in 9aw,U Ari-ona Law "eview BG H@220I: /-@, and in the pleaAbargaining conte!t see 5ren 8arAEill and 5mri 8enAShahar, U3lea 8argains 5nly #or the Euilty,U /ournal of Law and +conomics C2 H/11FI: B0BM Kenneth Kipnis, :Criminal Rustice and the 4egotiated 3lea,< +thics -F H@2GFI: 2BM Stephen R. Schulho#er, U3lea 8argaining as 6isaster,U 0ale Law

/ournal @1@ H@22/I: @2G2M Robert 7. Scott and $illiam R. StuntJ, U3lea 8argaining as Contract,U 0ale Law /ournal @1- H@22/I: @2@-A/C. inally, there are :implicative< e!plorations o# the related topic o# the preAe!isting duty rule. See 5ren 8arAEill and 5mri 8enAShahar, UCredible CoercionUM Randy 7. 8arnett, UA Consent 'heory o# Contract,U Columbia Law "eview -F H@2-FI: /F2M Robert 9. *ale, U8argaining, 6uress, and 7conomic 9iberty,U Columbia Law "eview CB H@2CBI: F1B. +pdated: -./0.12 3age 77 o# 94 - 'hese categories are rough and ready. or e!ample, one legal scholar discusses the concept, or rather :concepts,< o# coercion in a way that does not seem to #it neatly into the categories I have described. %itchell 8erman thin"s that the appropriate way to e!plore coercion is to loo" at how :coercion claims< #unction in our practical reasoning and to thin" about the types o# (uestions that coercion claims answer. 5n this approach he thin"s that there are di##erent concepts o# coercion that generate di##erent implications, #or e!ample, whether a coercion claim counts as an e!cuse, in di##erent conte!ts. So, the account is de#initional in that it identi#ies di##erent "inds o# coercion but implicative in that these the de#initions are connected with speci#ic implications. %itchell 4. 8erman, U'he 4ormative unctions o# Coercion Claims,U Legal (heory - H/11/I: C. 2 Since my interest is not in o##ering an account o# coercion but in the (uestion o# the appropriate treatment o# coercive agreements, I shall not be e!cessively concerned with the :de#initional< (uestions on which most philosophers #ocus, but will e!plore them only to the e!tent re(uired to address the :implicative< (uestion. @1 It turns out, however, that not all possible combinations are e!tensively represented in the literature, and that some writers on coercion might be thought to ta"e a mi!ed approach, which may, #or e!ample, be #ocused on the de#initional

+pdated: -./0.12 3age 78 o# 94 tas" but solely #or the purpose o# its implications. 'his is the approach that I understand %itchell 8erman to be ta"ing. 8erman, :4ormative unctions.< 5ther thin"ers rely on an implicit de#inition o# coercion but #ocus on a criterion by which to identi#y not which cases are cases o# coercion, but rather which cases should be dealt with under the law. 5ren 8arAEill and 5mri 8enA Shahar use their :credibility criterion< to serve this #unction. :Credible Coercion.< inally, some views are structured to leave open the particular moral theory to which the account o# coercion needs to be tied. In $ertheimer&s account o# coercion baselines, the moral theory that sets the moral baseline theoretically can be le#t open, and perhaps be utilitarian, despite $ertheimer himsel# being a deontologist. $ertheimer, Coercion, /@/. @@ 'his is somewhat misleading because the psychological accounts are also somewhat moraliJed in the sense that they are normative, yet their normative claims rest on psychological or empirical premises rather than moral ones, and i# this is borne in mind, these categories should not con#use. @/ %iller v. enton, CGC +.S. @1C, @@F H@2-0I. @B Lyons v. )kla., B// +.S. 02F, F1@ H@2CCI. @C See $ertheimer, Coercion, - H:KCLoercion negates or compromises the voluntariness o# an act . . . .<I. $ertheimer, however, is better classi#ied as a moraliJed deontologist, as will be clear below. +pdated: -./0.12 3age 79 o# 94 @0 or a criticism o# the psychological approach, see Claire in"elstein, U6uress: A 3hilosophical Account o# the 6e#ense in 9aw,U Ari-ona Law "eview BG H@220I: /-@. @F 3aul *. Robinson, Criminal Law Defenses HSt. 3aul: $est, @2-CI, @:B0@. @G Eeorge 3. letcher, "ethinking Criminal Law H4ew ;or": 5!#ord +niversity 3ress, /111I, -@@ Hemphasis addedI. @- See $ertheimer, Coercion, B1AB@. 'o be more precise, $ertheimer characteriJes the choice prong in terms o# what would count as not a :reasonable choice,< rather than whether there was any opportunity #or choice at all. Ibid., BF.

Ibid., /1@A1/. $ertheimer believes that nothing can come o# an account that determines choice based on the desirability o# the set o# alternatives one #aces or the relative desirability Hwhat $ertheimer calls :distance<I o# the alternativeHsI when compared with other alternatives in the set. Ibid., @2/A/1@. Rather, $ertheimer regards the essential element in coercion as an impairment o# the degree to which the victim&s choice was a mani#estation o# his secondAorder pre#erences. Ibid., /1/ and B1@A1F. $ertheimer believes his twoA prong analysis trac"s the Second Restatement&s two duress provisions. Ibid., B1. See supra note @. /1 Ibid., B-ACF. +pdated: -./0.12 3age 80 o# 94 /@ Ibid., C. // Ibid., /@G H:K'Lhe structure o# coercion discourse presupposes that A and 8 have certain obligations and rights which establish a bac"ground against which A&s proposals are understood.<I /B Ibid., /@GA//@. $hile it is clear that $ertheimer&s account centers around his account o# moral baselines, he is o#ten nonAcommittal with respect to whether a moral test is the always the correct test #or determining a potential victim&s baseline. *e states: :'here is no single right answer . . . when the results o# the nonmoral and moral tests K#or baselinesL diverge. 7ach test supports a de#ensible or plausible coercion claim.< Coercion, /@/. /CI call this view :counterintuitive,< but it incorporates a #airly standard move among philosophers, namely to thin" o# acting against second order principles one holds as acting involuntarily. See, e.g., *arry E. ran"#urt, : reedom o# the $ill and the Concept o# the 3erson,< /ournal of hilosophy FH@2G@I: 0A/1. /0 See 4oJic", :Coercion,< FA//. /F Ibid., /C. /G $ertheimer, e!plicitly leaves it open: :$hether these moral re(uirements are ultimately grounded in a deontological or conse(uentialist
@2

theory, the structure o# coercion discourse presupposes that A and 8 have certain +pdated: -./0.12 3age 81 o# 94 obligations and rights which establish a bac"ground against which A&s proposals are understood.< Coercion, /@G. /- Richard A. 3osner, +conomic Analysis of Law, Cth ed. H8oston: 9ittle, 8rown, @2G/I, @@0. See also Steven Shavell, who notes that :voiding removes the socially waste#ul incentive to engineer situations o# need.< :Contractual *oldup and 9egal Intervention,< (he /ournal of Legal $tudies BF H/11GI: BBF. 9ater I discuss 5ren 8arAEill and 5mri 8enAShahar in contrast with my own view. It is worth mentioning that while their view on coerced contracts is similar in many respects to a perspective I ta"e here, it diverges in its dependence on conse(uentialist concerns. or e!ample, 8arAEill and 8enAShahar worry that en#orcing coerced contracts might encourage :investments< in soliciting harm#ul in#ormation #or purposes o# blac"mail and encourage highway gunmen to ma"e themselves more credible coercers, a consideration that would not be directly relevant in my view. UCredible Coercion,U (exas Law "eview -B H/110I: GB/ABB. *owever, their view seems to diverge #rom a completely conse(uentialist approach in that they, li"e mysel#, are primarily concerned with the wellAbeing o# the potential victims o# coercion. Ibid., G@-. /2 Robert Cooter and 'homas +len, 9aw and 7conomics H*arper Collins, @2--I, /0C. B1 Ibid., /0/. +pdated: -./0.12 3age 82 o# 94 B@ Ibid. B/ Ibid., /0C. BG See Claire in"elstein, UA Contractarian Argument against the 6eath 3enalty,U 1ew 0ork 2niversity Law "eview -@ H/11FI: @B1-, #or an overview o# the distinction. B- See Rohn Rawls, A (heory of /ustice HCambridge, %A: 8el"nap 3ress, @2G@I. B2 See 'homas Scanlon, 3hat 3e )we to +ach )ther HCambridge, %A.: *arvard +niversity 3ress, @22-I.

$ertheimer would probably concede the point, in so #ar as he thin"s the idea o# voluntariness is a normative, not an empirical concept. Coercion, /21. And #rom a normative standpoint, tal" about mental processes being shortcircuited would be out o# place. C@ So when *obbes says, :o# the voluntary acts o# every man the ob)ect is some good to himself,< Leviathan, chap. DIO, par. -, he need not be understood as endorsing psychological egoism. Instead he can be understood as #alling into line with the traditional teleological account o# intentional action, according to which an action is correctly described as :voluntary< )ust in case it is intentional, under some description, and it is intentional under a description )ust in case it is per#ormed under that description #or the sa"e o# an end. +pdated: -./0.12 3age 83 o# 94 C/ $ertheimer, Coercion, 2. 4onAvolitional coercion would be a case in which one&s actions are literally physically compelled. CB See 'homas C. Schelling, (he $trategy of Conflict HCambridge, %A: *arvard +niversity 3ress, @2F1I. CC or a discussion o# this point in the criminal law conte!t, see in"elstein, :6uress,< /-@. C0 Sometimes the line between volitional and nonAvolitional coercion may be di##icult to draw, and this is an admitted di##iculty with these categories. or e!ample, is having one&s hand physically #orced to sign a document so very di##erent #rom being :#orced< to sign a document at gunpoint? $hat about at "ni#epoint, where the point o# the "ni#e literally presses your hand in the desired direction? 6istinctions with #uJJy edges are still distinctions, however, and the basic separation between volitional and nonAvolitional coercion seems nevertheless maintainable. CF 4oJic", :Coercion,< /C. CG Ibid. C- Ibid., /-. C2 Coercion, /@/A@G. 0@ 3osner, +conomic Analysis, @@0. 0/ 8ut see Rohn C. *arsanyi, :%orality and the 'heory o# Rational +pdated: -./0.12 3age 84 o# 94 8ehaviour< in 2tilitarianism and 4eyond, eds. Amartya Sen and 8ernard
C1

$illiams H4ew ;or": Cambridge +niversity 3ress, @2-/I. Arguably a rational choice theorist li"e *arsanyi ta"es the individualistic perspective more seriously than other conse(uentialists, despite the #act that his main purpose is to de#end the principle o# utility, on the grounds that individuals in an :original position< would choose that principle as their principle o# distribution. It is beyond the scope o# the present paper to discuss his view, since he himsel# does not discuss coercion. 8ut it might be argued that a #uller economic account o# the coercion literature could, nevertheless, ma"e room #or the individual perspective by embedding the principle o# average utility in a bargaining #ramewor". It is my own view, however, and the view o# many other contractarians, that the bargaining #ramewor" does not in #act lend support to the principle o# utility. 0B or a greater appreciation o# the di##erences between economic and rational choice approaches to law, see Claire in"elstein, :9egal 'heory and the Rational Actor,< in (he )xford *andbook of "ationality, eds. A. R. %ele and 3. Rawling H4ew ;or": 5!#ord +niversity 3ress, /11CI. 0C or #urther discussion o# the point, see ibid. 8ut cf. Rohn *arsanyi supra note C-. 00 8ut see note /-. +pdated: -./0.12 3age 85 o# 94 0F 'homas *obbes, :6e Cive!< in &an and Citi-en, ed. 8ernard Eert HIndianapolis: *ac"ett, @22@I,@/2. 0G 'homas *obbes, Leviathan, chap. DIO, par. /G. 0- Ibid., chap. DIO, par. /1. 02 'o avoid cumbersome #ormulations I have been re#erring to an exchange of promises, but, as will be clear shortly, that e!change can also be :unilateral,< namely an e!change o# a promise #or a present per#ormance. F1 *obbes thin"s that promises made without e!change, or gi#ts, are not binding i# made #or the #uture H:I promise I will give you my cow tomorrow<I. 8ut he thin"s that gi#ts where the promisor trans#ers a right to another party now are binding, even i# some part o# the per#ormance remains #or later. :I hereby trans#er my cow to you now, to be delivered tomorrow,< is a binding commitment

to deliver the cow tomorrow. 'he #irst category o# promise is the one in which the element o# e!change ma"es a di##erence. See *obbes, Leviathan, chap. DIO, pars. 2A@F. F@ Ibid., chap. DIO, pars. /G. F/ Ibid., chap. DD, par. @. FB Ibid., chap. DD, par. @. FC Ibid., chap. DD, par. /. F0 Ibid. +pdated: -./0.12 3age 86 o# 94 FF Ibid. FG Ibid., chap. DDI, par. B. F- Ibid., chap. DIII, par. @B. F2 Ibid., chap. DIO, par. @-. G1 'his distinction is the distinction between unilateral and bilateral contracts. arnsworth, Contracts, T B.C. G@ Ibid., chap. DIO, par. 2 H:'he mutual trans#erring o# right is that which men call C54'RAC'.<I. G/ Ibid. H:K5Lne o# the contractors may deliver the thing contracted #or on his part, and leave the other to per#orm his part at some determinate time a#ter Hand in the meantime be trustedIM and then the contract on his part is called . . . C5O74A4' . . . .<I. GB *oward $arrender, (he olitical hilosophy of *obbes' *is (heory of )bligation H5!#ord: 5!#ord +niversity 3ress, /111I, B1. GC Eregory S. Kav"a, *obbesian &oral and olitical (heory H3rinceton: 3rinceton +niversity 3ress, @2-FI, B1C. G0 6avid 3. Eauthier, (he Logic of Leviathan' (he &oral and olitical (heory of (homas *obbes H5!#ord: Clarendon 3ress, @2F2I, C@. GF *obbes, Leviathan, chap. DIO, par. @@ Hemphasis addedI. 'he #act that *obbes says :may< here should not be ta"en to mean that a covenant may be one +pdated: -./0.12 3age 87 o# 94 in which one party must per#orm, but rather that the person may or may not end up per#orming on the covenant. GG Ibid., chap. DIO, par. 2. *e distinguishes trans#erring a right to a thing #rom trans#erring the thing itsel#. 'he suggestion is that a direct e!change o# goods or o# money #or goods would not constitute a contract, since there are no ongoing obligations a#ter the transaction has terminated. 'his is consistent with

the position that modern law ta"es on the topic o# barter. G- arnsworth, Contracts, T B.C. G2 Ibid. -1 *obbes, Leviathan, chap. DIO, par. @-. -@ arnsworth, Contracts, T B.C. -/ Ibid., chap. DO, par. @. -B Ibid., chap. DO, par. C Hdescribing the #ool as an individual who does not believe in )usticeI. -C Ibid., chap. DO, par. 0. -0 Ibid. -F I argue that *obbes was committed to the common "nowledge assumption as a #eature o# the state o# nature in Chapter 'wo o# my #orthcoming manuscript. See :Contracts and Covenants< in *obbesian Legal (heory Hunpublished manuscriptI. +pdated: -./0.12 3age 88 o# 94 -G Cristina 8icchieri, "ationality and Coordination H4ew ;or": Cambridge +niversity 3ress, /11GI. -- 'his should not be misunderstood as suggesting that the robber&s willingness causes it to be rational #or you to return with the money. It is rather that the superior position both you and the robber en)oy #rom the e!change ma"es it rational #or each o# you to con#orm your actions to the plan, provided that the other is willing to do so as well. 'hus the robber&s willingness to abide by the plan signi#ies that he is acting in con#ormity with the rational course o# action #or both o# you, and so suggests it is rational #or you to do so as well. -2 $ertheimer, Coercion, /B-. 21 *obbes, Leviathan, chap. DD, par. /. 2@ 7!ample drawn #rom the same te!t in the 9atin version. 2/ In contract law this might be "nown as the distinction between substantive and procedural unconscionability. 6uress, o# course, re#ers to the conditions under which the agreement is created. 'hus it is largely a procedural rather than a substantive notion. +pdated: -./0.12 3age 89 o# 94 2B 'here are surely cases in which it is di##icult even to ma"e conceptual sense o# this distinction. Consider attempting to distinguish the initial coercion used in rape #rom the resulting illegal act o# coerced intercourse. 5ne can hardly ma"e sense o# the suggestion that the #orce might be illegal but the resulting se!ual encounter legally protected.

'here is no upward limit to the amount o# punishment that can be imposed #or any act, as ever increasing amounts o# torture can always be added to death. 8ut there is potentially a problem with punishment bunching as punishments increase. 'his is the #amiliar problem that i# the punishment #or bi"e the#t is increased in an e##ort to more strenuously deter this sort o# crime, then the punishment #or bi"e the#t and #or auto the#t may turn out to be too close. In this case, the schedule o# punishments will end up encouraging all potential bi"e thieves to steal automobiles instead, and that would not be an advantage. So i# the penalty #or the robber&s initial act o# coercion is su##iciently high, one ris"s encouraging greater, more lucrative crimes, etc. 5n the other hand, i# there is no upward limit on punishment generally, then other punishments can in theory be ad)usted upwards as needed. 20 See, e.g., $tilk v. &yrick, @G1 7ng. Rep. @@F-, @@F- H@-12I H9.R.C.3I H7llenborough, 9I Hholding a renegotiated contract void due to lac" o# considerationI. +pdated: -./0.12 3age 90 o# 94 2F 'hough the "estatement #$econd% of Contracts says that the modi#ication will be allowed i# the change is :#air and e(uitable.< T -2HaI. 2G See Steven Shavell, :Contractual *oldup and 9egal Intervention,< /ournal of Legal $tudies BF H/11GI: BBCAB0. 2- 5# course the doctrine o# consideration itsel# may be seen as a way o# guarding against coercion, since the e!istence o# consideration suggests reciprocation, and hence #air e!change. 8ut consideration in #act need not be #air in order to count as valid, and the doctrine is better understood as a way o# distinguishing e!changes #rom bargains than as a substantive test #or the #airness o# the agreement. 22 See 5ren 8arAEill and 5mri 8enAShahar, U'he 9aw o# 6uress and the 7conomics o# Credible 'hreats,U /ournal of Legal $tudies BB H/11CI: B2@. @11 See Alaska ackers v. Domenico, @@G . 22 H2th Cir. @21/I. @1@ :'hreats "nown to be noncredible cannot and will not coerce.< 5ren 8arAEill and 5mri 8enAShahar, UCredible Coercion,U (exas Law "eview -B
2C

H/110I: G/G Hitalics omittedI. +pdated: -./0.12 3age 91 o# 94 @1/ 'his is particularly so i# the contract calls #or personal service o# some sort. %ost courts will not order speci#ic per#ormance on such a promise because o# worries about degrading servitude. See, e.g., (he Case of &ary Clark! A 3oman of Color, @ 8lac"#. @// HInd. @-@/I. 'he most an aggrieved party Hthe canneryI can hope to get is damages, and those will li"ely be highly speculative and di##icult to prove. @1B See 8arAEill and 8enAShahar, :9aw o# 6uress,< B2BA2C. @1C As Richard 3osner put in a legal decision on this point, :KtLhe sensible course would be to en#orce contract modi#ications Hat least i# writtenI regardless o# consideration and rely on the de#ense o# duress to prevent abuse.< 2.$. v. $tump *ome $pecialties &fg., 210 ./d @@@G, @@// HGth Cir. @221I. @10 See 8arAEill and 8enAShahar, :Credible Coercion,< GBCAB0 H:K'Lhe optimal policy is not to allow victims to sue #or restitution o# their robbed possessions, but rather to increase the li"elihood o# apprehending murderers and bringing them to )ustice as well as to increase the sanction #or murder.<I. It should be noted that 8arAEill and 8enAShahar hedge on this point in that they state in passing that coerced contracts might be voided in cases where threatening parties deliberately ma"e their threats more credible He.g. by buying a gunI, as they worry that not voiding these contracts would create incentives to invest in credibility enhancing means. Ibid., GB/ABC. *owever, this (uali#ication con#licts +pdated: -./0.12 3age 92 o# 94 with their initial concern #or protecting victims o# credible coercion, since the victim o# credible coercion will not care at all whether or not the threatening party deliberately invested in ma"ing his threat more credible. Conse(uently, it remains unclear where 8arAEill and 8enAShahar really stand with respect to en#orcing contracts in the gunAman scenario. It does sound, however, as though

they are prepared to bite the bullet and en#orce such agreements when they are credible. @1F Ibid., GBCAB0. @1G Ibid., G// H:A credible threat is one that the threatening party intends to carry out.<I. @1- Ibid., H:I# it is in the interest o# the threatening party not to carry out the threatened outcome, his threat is not credible.<I. @12 Ibid., G/G Happlying the criterion to the highway gunman caseI. @@1 Ibid., GBG. @@@ See, e.g., Stephen R. Schulho#er, U3lea 8argaining as 6isaster,U 0ale Law /ournal @1@ H@22/I: @2G2 H:K3Llea bargaining seriously impairs the public interest in e##ective punishment o# crime and in accurate separation o# the guilty #rom the innocent.<I. 8ut see Robert 7. Scott and $illiam R. StuntJ, U3lea 8argaining as Contract,U 0ale Law /ournal @1- H@22/I: @2@-A/C Harguing that +pdated: -./0.12 3age 93 o# 94 e!tant contract principles o# coercion provide now reason to view plea bargaining as problematicI. @@/ %iller v. enton, CGC +.S. @1C, @@F H@2-0I. @@B Lyons v. )kla., B// +.S. 02F, F1@ H@2CCI. @@C 5arity v. 1./., B-0 +.S. C2B, C2F H@2FGI. @@0 ,rost (rucking Co. v. "ailroad Commn , /G@ +.S. 0-B, 02B H@2/FI. @@F 'he same point can be made by analogy about the 7!clusionary rule: punish police misconduct directly, but allow police to use the bene#it o# evidence obtained through such misconduct. $ee *erring v. 2.$., @/2 S. Ct. F20, F22 H/112I H:K5Lur decisions establish an e!clusionary rule that, when applicable, #orbids the use o# improperly obtained evidence at trial.<I 5ne might also ma"e the point about the entrapment de#ense. $ee /acobsen v. 2.$., 01B +.S. 0C1, 0C2A 01 H:$here the Eovernment has induced an individual to brea" the law and the de#ense o# entrapment is at issue, as it was in this case, the prosecution must prove beyond reasonable doubt that the de#endant was disposed to commit the criminal act prior to #irst being approached by Eovernment agents.<I.