In their thoughtful and provocative article, Pseudo-Contract and Shared Meaning Analysis,1 Professors Robin Bradley Kar and Margaret Jane Radin argue that contract — that is, the presumptively enforceable agreement of the parties — should be limited to the “shared meaning” of the parties, and that the boilerplate surrounding that shared meaning in modern form contracts (such as the online “terms and conditions” to which one clicks “I agree”) should be treated as “pseudo-contract” with no direct claim to enforceability. They go on to specify how “shared meaning” is created and how it is to be identified, and they provide several examples of their theory in action. In this Response, we discuss their theory with special regard to its application to boilerplate stipulations providing for mandatory arbitration of disputes.
Fundamentally, Kar and Radin rest their argument on two propositions. First, normatively in contract law “the central focus of justification is on the enforcement of common terms that parties agree to when they form contracts.”2 Second, creation of this shared meaning in turn “depends on an implicit presupposition of cooperative language use to form a contract.”3 Much, but not all, boilerplate, they argue, cannot meet these standards.4
These propositions face more substantial doctrinal difficulty than the article seems to recognize. We begin with their proposition that contract law focuses on the enforcement of shared understandings both parties agree on. “Contract meaning is shared meaning.”5 This seems to call forth the old adage that there needs to be a “meeting of the minds” to form a contract. But it is hornbook law that this maxim has been overtaken — for a century now — by the “objective” theory of contracts that favors the reasonable outsider’s interpretation of what the parties said or did.6 ed. 2004). The position the article takes needs, we think, further investigation.
Kar and Radin (following, in their words, the philosopher Paul Grice) distinguish “between ‘speaker meaning’ (or what a speaker intends when he or she utters a sentence) and ‘sentence meaning’ (or what a sentence means independent of its occasion of use).”7 “Sentence meaning” appears akin to what is often in legal discussions called “plain meaning.”8 Kar and Radin reject the proposition that contractual language can be understood without providing some context within which to interpret it. They are very persuasive in their demonstration that trying to understand a sentence without assuming some “occasion of use” simply does not work. Thus, for example, if someone says to you, “This car is worth $1500 but I can give it to you for $1000,” the sentence by itself could be taken on its face to simply be two statements of fact, or maybe even a description of a partial gift, whereas once we know it is uttered by a used-car salesperson on a car lot, it seems clear that the speaker was making an offer of a deal that you might or might not accept, with binding consequences.
But Kar and Radin are too quick to assume that once one abjures acontextual readings, the only sensible goal is to locate shared meanings. Perhaps misled by Grice’s suggestion that the alternative to plain meaning is the speaker’s meaning, they fail to consider the possibility that, in context, a speaker and a listener might not have a common understanding. Choosing between the parties’ meanings need not be based simply on the words used, and indeed the best common law cases employing the “objective theory” have not been so restricted.9 These cases say that if the speaker meant X and the listener heard Y, then if, in context, the reasonable observer would also have understood the statement to mean Y, the listener’s meaning prevails.10 These cases might be justified by the need to protect reasonable reliance, or simply by the desire not to upset contracts every time the parties’ subjective understandings differed to any degree as the “meeting of the minds” theory seems to require.11
How does this “objective theory” apply to boilerplate? In a contract formed by a (now-old-fashioned) exchange of letters, the recipient’s understanding of the meaning of a letter, if also the view of the reasonable person aware of the context, would be protected. Receipt of a reply letter saying “I agree to the terms you sent me” would form a contract on the previously communicated terms. By extrapolation, one could argue that the recipient of a computer-clicked “I agree” could reasonably believe that the many pages of boilerplate text, assuming they were available, had been agreed to.
But that extrapolation ought to fail — not because there is a lack of common meaning (which the “objective theory” allows), but because (a) the reasonable person, at least in context, would know that the party who clicked “I agree” was unlikely to have read, let alone in fact have agreed to, all the boilerplate pages, and, because (b) even the recipient of “I agree” would not in fact really believe that its sender had read, let alone agreed to, the boilerplate.12 Perhaps all that needs to be said, then, is that Kar and Radin rightly, but perhaps too quietly, reject the application of the objective theory of contracts to modern boilerplate. To accept that application would be (to use their pun on Thomas Kuhn’s famous “paradigm shift”) to engage wrongly in a “paradigm slip” — that is, in the ever-more extreme circumlocutions that are required to make modern computerized boilerplate fit within even the objective theory’s view of agreement.13 As regards this type of arrangement, Kar and Radin are right to say that only “common meaning” ought to be enforced. As to the problem of upsetting too many deals, Kar and Radin sensibly say that, even if much boilerplate ought to be ignored, those provisions that are actually agreed upon ought still to be enforced.14
The connection between existing contract doctrine and Kar and Radin’s other fundamental proposition, the “presupposition of cooperative language use,” needs more extensive discussion. Kar and Radin present their approach as based on linguistics, on what they term “maxims” about how language is used, or ought to be used, to create joint meaning.15 This is not an argument about what needs to be done regarding boilerplate in order to achieve substantive fairness or to rectify social power imbalances, but rather is presented as being in furtherance of freedom of contract. Here, as elsewhere, they tend to treat the insights of linguistic philosopher Paul Grice as if they were established truths just waiting to be turned into legal rules, which may be more weight than these admittedly helpful insights can bear. Insofar as Kar and Radin connect his theories with existing doctrine, they suggest that much of what they describe as linguistic cooperation might be recharacterized under existing contract law as an extension of the doctrine of good faith.16
As the hornbook says, the received doctrine is that “American law imposes no general duty to negotiate a contract in good faith.”17 Instead, the duty of good faith toward one’s partner arises only after the parties have entered a contract. This does not mean, of course, that there are no restraints on what can be said or done during negotiation, but rather that those limits are expressed by the less restrictive doctrines of fraud, misrepresentation, and duress. By means of some very well-argued examples, Kar and Radin claim that formation of a contract depends on the cooperative use of language that goes beyond these limits. We think they are right, up to a point.
Kar and Radin have an extended discussion of this interchange: Amber, a homeowner, says to Paolo, a professional housepainter: “If you paint my house, then I will pay you $1000”; and Paolo replies: “I accept.” The authors convincingly show that the conclusion of the law, that in this conversation Paolo has promised to paint and Amber has promised to pay, depends on inferences from the background assumption that Amber and Paolo are trying to cooperate to make a contract (rather than simply stating facts or making a prediction).18 This result is an act of interpretation — based in good part on a factual judgment as to what is probable, but also to some degree on a normative determination to treat as legally irrelevant the possibility that the parties were trying to trick each other.
The same sort of analysis can be used to show that many of the recognized legal obligations of good faith in performance carry over to good faith in contract formation. Suppose that, in a negotiation for the sale of a house, the buyer agrees to the seller’s price by saying: “I’ll pay what you are asking, contingent on my getting a mortgage for ninety percent of the price at five percent interest or less.” The cases say that, once this deal is suitably reduced to writing, the buyer has promised not only to go through with the deal if he gets the mortgage, but also to make reasonable efforts to try to get a mortgage on the stated terms.19 So stated, it is an example of the obligation to perform in good faith. But it would also be fair to say that the rule relieves the seller during the negotiation of having to respond: “And do you promise to try to get that mortgage?” and from proceeding to get such an explicit commitment from the buyer. In effect, it represents a legal obligation, in context, for the buyer to be speaking cooperatively to get a sensible deal done.
Good faith in performance carries over to good faith in negotiation at least this far, which would seem to be further than the law of duress, fraud, or misrepresentation would go. But the obligation of good faith in performance stretches across a broad band of possible courses of action and sometimes requires parties already in contracts to place the interests of other parties on an equal (perhaps even superior) basis to their own.20 As the obligation increases, it seems based less on a judgment as to what the parties actually meant, and more on a judgment as to what, in the circumstances, is a fair result, all parties’ interests considered. If one accepts the idea that parties are allowed to be more self-interested during negotiations than they are entitled to be after a deal is closed, it is harder to say that these further results also carry over to the process of contract formation. It is especially hard to say so if the basis for asserting the claim to good faith in contract formation — for asserting a requirement of adherence to the cooperative use of language — is meant to be purely linguistic, purely based on what is needed to create shared meaning.
Kar and Radin make use of the carryover from good faith in performance or enforcement to good faith in contract formation to support some of their more particular specifications of what is involved in the cooperative use of language endeavoring to reach a contract.21 There, seemingly as part of their effort to conform the law to what they perceive as Grice’s linguistic rules, they seem at times to go beyond the fair reach of the argument. For example, they say that “[p]arties who use language cooperatively to contract” have to obey “The Contractual Maxim of Relation,” which includes committing themselves to a contract only when, among other things, “they mean to confer legal standing on the other party to be able [to] demand compliance, [and] to grant that party the power to invoke the formal power of the state in cases of noncompliance.”22 The degree of caring for the interests of the opposing party here called for seems to us to rest on considerations of substantive justice that go beyond what an interpretation of cooperative behavior in the context of self-interested bargaining can sustain. How adequate proposed remedies are — whether, for example, having only the power to terminate, after notice, all further dealings is a sufficient remedy — is something, as a matter of general contract law applied to an ordinary negotiation, one would leave to each side to decide for itself. Disobeying the supposed maxim would not, it seems to us, prevent the production of a shared meaning. (Concealment, or deception as to the effect of proposed language, would, of course, be a different matter.)
Kar and Radin also suggest a possible nondoctrinal method for ascertaining the degree to which particular contract terms do, or don’t, present shared meanings reached through proper cooperation. They state:
We suggest that courts imagine that all of the written and digital text exchanged during contract formation is converted into oral form and takes place in a face-to-face conversation between the relevant parties. With respect to any disputed boilerplate text, courts can then ask the following question:
Could this boilerplate text have plausibly contributed to an oral conversation that contributes terms to a contract consistent with the presupposition that both parties were observing the cooperative norms that govern language use to form a contract?23
They are clear that they are not claiming that oral contracting is an obligatory ideal, but are rather presenting it as a heuristic tool to help a court determine what genuinely constitutes shared meaning.24 But we are frankly skeptical that any state or federal court could readily engage in this exercise in imagination, or would want to.
Courts of first instance are, of necessity, tied to the record before them. In matters relating to contract interpretation and enforcement, it is a contested question how much of the context surrounding the making of the contract is properly admissible.25 We agree with Kar and Radin that, as regards the kind of contracts under discussion, having the court pay attention to the context under which boilerplate text is presented and “accepted” makes sense; our prior discussion of how we would apply the objective theory of interpretation assumes as much. This means having recourse to documents, electronic images, and testimony that show how the actual text at issue was formed. But if, instead, a court, following the Kar and Radin suggestion, were asked to use its imagination to conjecture whether a given term could have been the product of a fictional oral conversation where “both parties were observing the cooperative norms that govern language [according to Grice],”26 the court would hardly know where to begin, let alone what rule of law empowers the court to render a decision based on such imagined facts not part of the record. So, while we agree with Kar and Radin that ignoring a great deal of boilerplate language is closer to reality than crediting it as if it were a negotiated text, we think the discovery of the actual shared meaning in any given case must be based on facts in the record, and reasonable inferences therefrom, rather than on a somewhat squishy exercise in imagination.
With these comments regarding their general theory in mind, we turn now to Kar and Radin’s application of their approach to the very salient topic of boilerplate that requires that disputes be arbitrated, which usually also implies that class actions cannot be brought.27 This topic is of great interest to us,28 and, judging from the several times they treat it, to Kar and Radin as well.
The Federal Arbitration Act29 provides that “[a] written provision in any . . . contract . . . to settle by arbitration a controversy thereafter arising out of such contract . . . , shall be valid, irrevocable, and enforceable, save upon such grounds as exist at law or in equity for the revocation of any contract.”30 The Supreme Court in recent years has accepted case after case under the Act, and has proven to be enthusiastic for giving it the broadest application.31 The result of these decisions, generally stated, is that the Act applies fully to state, as well as federal, courts and that the only legal doctrines that meet the “save upon” clause are doctrines that qualify as part of a state’s general law of contracts (and maybe not even then).32 Many of the cases, as Kar and Radin point out, have allowed firms to enforce mandatory arbitration provisions adopted through boilerplate contracts.33 The only way to meet such decisions, short of congressional action to revise the Act, is to provide a general theory of contracts that would make these provisions unenforceable, or sometimes unenforceable, on grounds applicable to other contractual provisions as well. This Kar and Radin claim to have done. (Indeed, one might suspect that one of the main motivations for writing the article was to provide a general theory of contracts, satisfying the requirements of the Supreme Court, that argues against the routine enforceability of these particular boilerplate clauses.)
Kar and Radin provide two lines of argument. One rests on the claim, already discussed, that it is part of the cooperative use of language to form a contract that each side give and receive legally enforceable commitments. Arbitration provisions obliterating class actions for claims that would otherwise not be worth litigating, Kar and Radin say, are in “hidden conflict” with this principle and therefore should be treated as unenforceable “pseudo-contract.”34 For reasons already stated, we are not inclined to follow this argument. It seems to us to rely on unstated normative premises as to the adequacy of various remedies that are not fairly derivable from Kar and Radin’s general linguistic framework.
We prefer the more capacious argument that Kar and Radin make about long boilerplate contracts in general: “terms and conditions” to which the other party merely clicks “I agree” do not, and are not meant to, generate shared meaning.35 Kar and Radin express this conclusion in terms of their hypothetical oral negotiation, supposing, not unreasonably, that any conversation would have been terminated long before the last provision had been read.36 While as already said we see little use in that hypothetical test, it seems to us that the same result follows from the well-known realities, often part of the record, that the form terms were not in fact read and were not reasonably expected to be read. The boilerplate terms, to use Kar and Radin’s language, were not introduced “into a cooperative linguistic exchange that aimed at producing a common meaning of the parties.”37 Once the term providing for arbitration, like any other term, is not enforceable for this reason, it simply drops away in favor of the legally provided default rule, which, in this case, is enforcement through a civil action, including a class action when otherwise allowed.
The conclusion would seem to be, as Kar and Radin put it: “Parties are, of course, still free to agree in advance to arbitrate a broad range of disputes. But they cannot agree to arbitration without actually doing so as part of the common meaning of the parties.”38 Which, of course, raises this question: If presenting long boilerplate text into a routine transaction, and then relying on an unrealistic “duty to read” and a clicked “I agree” violates norms of cooperation and gives the objective observer no reason to think that the text has been agreed to, what suffices? Kar and Radin sensibly say that the answer depends on context, and advise that “[v]endors that hope to use boilerplate text to produce a common meaning of the parties would therefore need to choose their battles wisely and find innovative ways of cooperatively communicating those terms that are the most essential to their proposed contracts.”39 The example Kar and Radin offer to flesh out this generality — discussing a different form provision — gives a purchaser of a license to hear a song two choices (single-platform listening or multiple-platform listening) at explicitly different prices;40 much the same commonly happens when one rents a car and has to choose among various insurance options. But if the underlying issue is whether the parties share meaning as to what has been agreed, the offering of a choice would not seem to be required; and a state statute or court decision saying that parties need to be offered a choice specifically to accept or reject arbitration would seem to run afoul of the Federal Arbitration Act, at least as the Supreme Court understands it.41
As regards arbitration, then, the argument seems to come to this: once we understand the enforcement of a contract term as based on the existence of an agreed shared meaning, or what can reasonably be thought to be one, a boilerplate agreement to resort exclusively to arbitration has to have been made both clear and obvious to the other party to be enforceable; but once that has been done, the party wanting the arbitration provision is free to insist on it as the price of agreeing to the contract on its part. That, to us, seems as far as an argument based on the interpretation of the parties’ use of language in context can go.
No doubt this proposal is preferable to those cases that seem to accept the merest chance that a text might be read by someone as adequate to make it binding. But what is its practical effect? Assuming Kar and Radin’s framework, as we have reconstructed it, were in force, we suspect that many firms would decide that a provision for mandatory arbitration, sans class action, would be, to repeat Kar and Radin’s words, among “those terms that are the most essential to their proposed contracts.”42 However forthright these firms had to be, they would try to insist on its inclusion in their contracts with consumers and with employees. Would they succeed? This is no longer a question of interpretation, but rather a matter of the social and economic dynamics of particular situations. Individuals faced with terms that firms insist on are rarely in a position to bargain for different language. Of course, they can, legally speaking, walk away; but they will in fact walk away only if there is a realistic alternative that is available. The real question as to practical effect, then, is whether making a given firm’s arbitration term perspicuous to the parties asked to adhere to it will generate enough resistance among enough people to provide the basis for other firms in the market to compete by offering contracts without an arbitration provision. If so, as regards arbitration, using the Kar and Radin analysis will generate a real increase in contractual freedom. If not, it will only generate an awareness of a choice that exists in theory but has not been made real.
We do not know what will in fact happen — and given the importance of the issue, we are hesitant to remit it to uncertain market dynamics. Kar and Radin’s approach is, in the end, tethered to the fact that it is a method of interpretation employed in the quest for freedom of contract. As the great sociologist Max Weber long ago pointed out, freedom of contract often increases the options open to those with economic or social power, but often decreases the freedom of those who lack it.43 Even taking freedom of contract as the background norm, society often removes particular items from its sphere. How far to go in making any particular social arrangement subject to contractual freedom depends on judgments as to both the importance of the arrangement and the circumstances in which the freedom will be used. Making arbitration a matter to be arranged at will between the parties, when the parties are consumer and firm (rather than two commercial entities) or single employee and large employer (rather than union and employer) seems a mistake to those who, like us, believe that access to the courts is a foundational right that protects against all sorts of social abuses. Kar and Radin’s approach is an improvement over cases that allow mandatory arbitration to be hidden in fine print; it at least raises the possibility of creating countervailing social forces. But as regards arbitration specifically, our first choice would be to amend the Federal Arbitration Act to undo much of what the Supreme Court has found in it.
*Byrne Professor of Administrative Law, Harvard Law School.
**Senior United States District Judge, United States District Court for the Southern District of New York.