Corps de l’article

I. Introduction

In recent years, the parameters of ordinary damages for breach of contract have been somewhat blurred by decisions purporting to award “contract damages” without adhering to the “contract measure.” This has not yet had a noticeable deleterious effect on wider contract and damages jurisprudence. However, given that the decisions in question are from the Supreme Court of Canada, it appears necessary to challenge and correct the misunderstandings present. In this article, I will address a key misunderstanding evident in relation to the nature of contract damages in the majority decision in CM Callow Inc v Zollinger.[1]

Many will already be familiar with Callow’s facts and its outcome, but to contextualize the later discussion and identify the source of my concern, I will briefly recap the facts and holding. The case, in short, involved a winter maintenance contract (including work such as snow removal) between Callow and a group of condominium corporations in Ontario represented by a joint use committee (the defendants) that covered the period from November 1, 2012, to April 30, 2014.[2] Callow was apparently led to believe that its performance under the contract was satisfactory despite some early performance issues.[3] The Supreme Court accepted that this led Callow to believe that the winter maintenance contract was not in jeopardy, which subsequently turned out to be untrue.[4] In reality, the defendants had decided in March or April 2013 to terminate the contract, but only did so in September 2013 so as not to jeopardize Callow’s performance under a separate summer maintenance contract.[5] Termination occurred on ten days’ notice as permitted, but came as a surprise to Callow given what had been communicated about the quality of the work done.[6] A majority of the Court held in two separate judgments that this deception was a breach of the defendants’ duty of honest performance and confirmed an award of damages equivalent to the value of the balance of the contract.[7] The Supreme Court’s view of the remedy was that it was necessary because Callow had foregone the opportunity to bid on other contracts for the second winter, and that it was justifiable (in the majority’s opinion) under the expectation measure despite the Court’s emphasis on reliance and the fact of the termination clause.[8] It is these latter two points relating to the assessment of the award that I will focus on in my critique of the case.

Before getting to the crux of my critique of Callow, however, I will first explain the nature of contract damages for context. I will then turn to explain how the majority position in Callow is incorrect, as well as the position taken in relation to the same matter in the Court’s earlier decision in Bhasin v Hrynew.[9]

II. Expectation vs. Reliance

Expectation damages are inherent in contract.[10] So inherent, in fact, that they are frequently referred to simply as “contract damages,” and the expectation measure as the “contract measure.”[11] As such, one could be forgiven for thinking that the expectation measure and the idea behind it are universally understood. In light of the recent decisions referred to above, however, these fundamentals bear some review.

The expectation measure is a paradigm for the assessment of damages that is geared towards vindicating the injured party’s expectation interest.[12] An expectation interest is quite literally the relevant party’s interest in obtaining benefits they could have expected to get had the relevant “injury” not occurred.[13] It is this interest that contract is most concerned with, and it is clear that contract is the legal context where regard for such an interest and the application of such a measure for damages makes the most sense, since contracts are prospective by nature and thus invariably involve expectations — typically of benefit. Of course, it is possible to find occasions outside of contract in which expectations are also engaged.[14] But, it is worth noting that the expectation interest and associated measure of damages first arose in contract and have prevailed there since at least Robinson v Harman.[15] One should also note that expectation’s emphasis on benefits or betterment is particularly appropriate for contract, given that the relevant “injury” in a contract case is frequently not an injury in the common sense of having made the relevant party worse off in some way, but is often instead no more than a failure to have made the relevant party better off.[16] I note that this approach makes contract damages seem rather generous, and in some respects that is a fair assessment.[17] There are, however, intrinsic limits as to how far the expectation measure can go in relieving a disappointed plaintiff, although they may not be readily apparent. Fortunately, these limits become clear if one contrasts expectation with its long-time rival, reliance, which I will do after first explaining the reliance concept in contract law next.

The reliance interest and reliance theories of contract first entered modern contract discourse in 1936 with the publication of Lon Fuller and William Perdue’s seminal article, “The Reliance Interest in Contract Damages.”[18] This so-called “reliance interest” was explained by Fuller and Perdue as the plaintiff’s interest in being able to rely on their contract in making other and further decisions.[19] The reliance measure, which is the paradigm of damage assessment geared towards vindicating a plaintiff’s reliance interest, is assessed so as to ensure that the plaintiff is no worse off as a result of a breach than they would have been had the contract not been breached.[20] This, the learned authors contended, was ethically superior to expectation as manifestations of corrective rather than distributive justice in the Aristotelian sense, and arguably already reflected in the law itself.[21] Although, none of the modern authorities appeared to actually admit to protecting a distinct reliance interest or employing a reliance measure, or at least not at the particular time that the authors were writing.[22] Academic and professional audiences were undeniably impressed by the authors’ analysis, however, and it is fair to say that reliance theories of promissory liability only entered the fray of modern contract theory as a result of Fuller and Perdue’s work.

Although reliance theories only came to prominence in contract discourse after Fuller and Perdue’s work, one should note that support for their position can be found both before and after their seminal article.[23] The former can arguably be found as far back as the reign of Elizabeth I, at which time assumpsit was arguably understood to be an action to recover the value of reliance a party had placed on a promise, rather than the value of the promise itself.[24] The latter can be found in the English Court of Appeal’s seminal decision in Anglia Television Ltd v Reed nearly 30 years later.[25] Readers will recall that Anglia Television is noteworthy for being the first decision in the modern era to explicitly hold that wasted expenditure could be recovered in contract notwithstanding uncertainty about the relevant contract’s profitability for the plaintiff, which gave credence to the argument that reliance was an interest protected by contract quite apart from expectation. In the particular case, this led to Anglia being able to recover production costs for a show that were wasted because of the defendant actor’s decision to withdraw from the project, even though the plaintiff production company could not show that the project would have been profitable and that the production costs would not have been wasted in any event.[26] But, as we know, if there was support for the view that contracts protect reliance rather than expectation, or that reliance constitutes a separate interest protected under contract, that support was short-lived in the wider jurisprudence.

Reliance’s fall from favour in the Commonwealth came later than it appears to have in the United States,[27] but only a few short years after Anglia Television suggested that reliance may have a life of its own in contract. The end began with the seminal British Columbia Supreme Court decision of Justice Berger in Bowlay Logging Ltd v Domtar Ltd in 1978.[28] Thereafter, appellate courts in other leading Commonwealth jurisdictions disavowed reliance as a distinct interest protected in contract, including England in 1983[29] and Australia in 1991.[30] The reasons for this uniform rejection will be explained next using Bowlay as an example. Before doing so, I should reiterate that my reasons for exploring Bowlay’s rejection of the reliance interest are to explain the intrinsic limits on contract damages referred to above and to establish — specifically for the later discussion of the Supreme Court’s recent work — what contract damages are, and just as importantly, what they are not.

The litigants in Bowlay were parties to a contract under which Bowlay was to cut timber on Domtar’s behalf.[31] Domtar, for its part, was to supply trucks to transport the timber to its yard in the town of Golden, British Columbia.[32] Unfortunately for Bowlay, however, logistical challenges caused Domtar to provide fewer trucks than anticipated. Bowlay alleged that this was a breach of the parties’ contract and that it prevented Bowlay from being able to carry out the contract efficiently, which led to Bowlay being forced out of business.[33] Bowlay sued Domtar for breach of contract on this basis and claimed damages for the out-of-pocket expenses that it incurred while attempting to perform its part of the bargain.[34] What Bowlay notably did not sue for, though, was lost profits.[35]

Domtar defended Bowlay’s claim against it on the basis that Bowlay would have lost money on the contract even if there had been no breach, and that it would have only lost more had the contract continued.[36] Justice Berger, for his part, held that Domtar, as the defendant, bore the onus of proving that the plaintiff’s allegedly wasted expenses would have been wasted in any event, even if the contract had been performed without issue.[37] Fortunately for Domtar, Justice Berger found that the onus was met in this case and that Bowlay could not recover any of the expenses it alleged were wasted on account of Domtar’s breach.[38] In reaching his conclusion, Justice Berger made the following observation that is apposite for present purposes:

The law of contract compensates a plaintiff for damages resulting from the defendant’s breach; it does not compensate a plaintiff for damages resulting from his making a bad bargain. Where it can be seen that the plaintiff would have incurred a loss on the contract as a whole, the expenses he has incurred are losses flowing from entering into the contract, not losses flowing from the defendant’s breach.[39]

The upshot of the above is that reliance is not an independent interest protected in contract, or at the very least not an interest of coordinate stature with expectation, because it is what the plaintiff could have expected their position to be had the contract been performed that marks the upper bound for recovery.[40] And regardless of whether a plaintiff has relied honestly or not in incurring expenses or foregoing other or further opportunities, contract will not knowingly put the plaintiff in a better position than they would or could have been had the contract been performed without issue.[41] As such, while reliance may be an appropriate alternative method of assessment when the value of the plaintiff’s expectation cannot be ascertained, it is not an alternative the plaintiff can elect simply because it would yield a pecuniary advantage. Instead, the fact or probability of such an advantage means that such a measure should not be available at all.

Having now said all of the above, I believe the stage is now set for the critique I intend to level at the Supreme Court’s decisions mentioned earlier. However, before departing, I am compelled to observe that I believe the foregoing to be correct not only in substance as a reflection of the present state of the law, but also as a matter of principle with regard to the underlying unarticulated norms that animate our understanding of what contract is at common law. To understand what I mean, it helps to remember that to some extent every contract entails taking a position about the future. There are an infinite number of positions that one can take, but at its simplest, the position is almost always that the agreement will prove to be advantageous. And in that way, every contract is, in some sense, a wager about how the future will unfold. Literal wagers are, of course, presumptively invalid, but given that one of contract’s understood functions is to facilitate the allocation of risk, it is safe to say that the common law forgives no small amount of betting short of actual gambling.[42] Short of actual gambling, though, it appears that there ought to be another limit, which is that contract will not suffer arrangements that amount to “heads I win, tails you lose.” That is why the foregoing discussion of expectation and reliance appears to my mind as though it must be correct, because there must be some limit to the allocation of risk under contract unless said risk is otherwise explicitly assumed. If the situation were otherwise, and reliance were truly an independent interest unfettered by expectation, every contracting party would become their opposite’s underwriter. But expectation will not permit of this, and it is expectation that prevails and limits recovery to ensure a plaintiff is no better off than they would have been had the contract not been breached, as well as no worse off.

III. The Supreme Court’s Misunderstanding of Damages in Contract

I have previously written at some length about the inaccuracies or inconsistencies in the Supreme Court’s earlier decision in Bhasin.[43] My hope was that these pieces might help shift the development of the duty of honest performance onto a more doctrinally sound footing. This work has apparently fallen on deaf ears for the most part, but one piece was cited by Justice Brown in the concurring minority in Callow for the proposition at the heart of my critique of the majority in this case.[44] That proposition is that the award granted in Bhasin, and consequently the duty of honest performance itself, vindicated the plaintiff’s reliance interest in the circumstances and had nothing to do with expectation, despite the Bhasin Court’s insistence to the contrary.[45] If the majority in Callow had considered my critique instead of drawing upon and employing the inaccurate language used by the Court in Bhasin, it is possible I would have no criticism to offer with regard to the remedy in this case at all. However, given that this has not happened, it appears that I will have to revisit the inaccuracy in the earlier decision, as it is discussed by Justice Kasirer for the majority in Callow, before explaining how the majority has made much the same mistake.

Bhasin involved a similar dispute to Callow. The plaintiff, Bhasin, had worked for some time as an “enrollment director” in Alberta for the defendant, Heritage Education Funds Inc. (known as “Can-Am”), pursuant to an enrollment director’s agreement between the two.[46] Under this agreement, Bhasin acted as a retail dealer marketing education savings plans provided by Can-Am.[47] Bhasin’s position was effectively that of a small business owner and akin to a franchisee, although not technically a franchisee under Alberta law.[48] The parties’ agreement was set to automatically renew unless either party gave at least six months’ notice of termination before the renewal date of October 4, 2001.[49] Bhasin suspected that his contract might not be renewed and asked Can-Am about that possibility in August 2000. Can-Am dissembled in response, but effectively misled Bhasin about its intention to terminate his agreement, which it then did by giving notice of non-renewal on May 4, 2001, six months before the renewal date (i.e. the last possible moment).[50] Can-Am’s deception and unexpected termination of its contractual relationship caused Bhasin to lose the whole value of his business as his workforce left for a rival, which prompted Bhasin to bring suit.[51] The Supreme Court’s conclusion after the appeal had reached them was that Can-Am had breached its duty of honest performance (introduced in this case) by misleading Bhasin about its intention to terminate the contract when Bhasin approached Can-Am in August of 2000, despite the fact that Bhasin was not entitled to notice at that point and that Can-Am had not actually breached any of the terms of the contract in delivering notice when it did.[52] The remedy provided in response to this wrong was damages of $87,000, said to be equivalent to the value in his business that Bhasin might have been able to salvage with the benefit of the additional notice he would have had if Can-Am had not misled Bhasin in August of 2000.[53]

The Supreme Court’s decision in Bhasin is notable for a number of reasons, but three aspects of the facts and the ruling bear emphasis before examining its inaccuracies. The first (as noted above) is that the defendant (Can-Am) clearly had not committed any breach of the parties’ contract. Can-Am had rather cynically misled Bhasin about its intention not to renew the contract with him, but it had abided by the contract’s notice provision. The second is that this “non-breach misfeasance” by Can-Am prompted the Court to recognize the existence of a general organizing principle of good faith in Canadian contract law, and a specific actionable duty stemming therefrom that requires parties to be honest in the performance of their contracts. The third, and most important for present purposes, is that Justice Cromwell, writing for the Court, took the position that the new extra-contractual duty of honest performance was a contract doctrine, and that its breach sounded in contract damages.[54] As I have written elsewhere, this is difficult to accept for a number of reasons, but it is particularly hard to understand in light of the way the Court assessed the award.[55] In short, the Court awarded Bhasin the loss he might have avoided if Can-Am had been honest when asked about the potential non-renewal of its contract with him, instead of misleading him to think his contract was safe before then giving notice of termination at the last possible second.[56] As I will explain below, the only way to explain this award (or that in Callow) on the basis of expectation involves turning the very idea of expectation on its head. One further observation about the Court’s award will be made first, however, in order for the larger point to be made clear.

What I would draw the attention of the reader to before I proceed to engage with the expectation point explicitly, is the fact that what the Court’s award did in Bhasin was undo the consequence of Can-Am’s representation being false.[57] It did so by attempting to put Bhasin in the position he would have been had Can-Am told the truth (i.e., that his contract would be cancelled) by undoing the effect of Can-Am’s deception on Bhasin’s behaviour and thus his fortunes.[58] What the award categorically did not do was put Bhasin in the position he would have been in had Can-Am’s representation been true, which distinguishes it utterly from any kind of expectation award.[59] An expectation award would have required the Court to award Bhasin the value of something akin to a renewed contract because this is what Can-Am’s deception encouraged Bhasin to actually expect.[60] In fairness to the Court, I should point out that there was good reason not to award Bhasin the value of a renewed contract in the circumstances of the particular case, and that I regard the actual award made as defensible.[61] But, the fact remains that while the Court’s award to Bhasin may have been right in substance, there was a surplus of things wrong with its rationale (particularly with respect to remedy).[62] And now, the majority in Callow appears to have deepened these errors further by matching mistakes in their reasons with mistakes in the outcome too.

We are now finally at the point where I can unpack and explain the mistakes made by the majority in Callow with respect to the remedy granted in that case. I note that I am not concerned with the methodological issues raised by the majority’s introduction of civil law concepts to answer common law questions, and that I intend to only address the errors in the portion of the majority’s judgment that purports to apply common law principles to the problem in this case. These mistakes are, first, the majority’s incorrect application/explanation of the expectation measure to justify a remedy that clearly vindicates reliance, and second, the fact that the award granted exceeds the settled scope of reliance remedies in contract and ought not have been granted at all.

A. Incorrect Explanations and the Nature of Expectation

Above, I noted that the Court’s approach to expectation in Bhasin and the majority’s approach in Callow turn the idea on its head. It does so in my view because each group has made or accepted a false equivalence between things that we might generally expect should not happen, and things that we can justifiably and specifically expect should happen. On its face, I admit that the distinction is perhaps not obvious, but if one unpacks the Court’s (perhaps) unwitting expansion of expectation to include the former as well as the latter, one can see that it is definitely a distinction with a difference. To understand what I mean, consider the supposed “expectation” that each group of jurists purports to be vindicating. In each case, it is the general “expectation” that one’s opposite contracting party will not lie.[63] Now, consider battery, negligence, or almost any other tort. In all cases there is some harm suffered by the plaintiff that they would generally not expect because the perpetrator had no right to do it,[64] which means that the plaintiff in all such cases could be said to have generally expected that they would not be harmed or that these things will not have been done to them. If we compare the two, it becomes clear that there is no meaningful difference. For if a general expectation that a contracting party will not lie falls within the ambit of expectation for the purposes of contract, then a fortiori so too does all of tort because there is no difference between a general negative expectation in relation to honesty[65] in contract, and the general expectation that would-be-tortfeasors will leave us alone. Both are negative in the sense that they pertain to how the world should not be, and neither bears any relation to even our most basic understanding of what it means to expect, because both pertain to things we literally do not expect (i.e. bad things that people are not supposed to do).

The effect of the conflation described above on the measure of damages associated with expectation is similarly hard to accept once it is unpacked. In short, if the wrong is doing the unexpected, and the point of the remedy is to undo this wrong, then damages must correspondingly be assessed so as to place the injured party in the position they would have been in had the wrong that the plaintiff did not expect, not happened. For those familiar with damages in tort, it may be difficult to see how exactly this differs from the tort measure intended to effect restitutio integrum.[66] But, this is how the remedies in Bhasin and Callow have undeniably been assessed. In Bhasin, the amount awarded was assessed so as to undo the decision or decisions made by Bhasin following Can-Am’s deception (decisions that he would not have made had Can-Am not deceived him when he asked about renewal of his contract).[67] As I have already said, it is impossible to explain how exactly such an award can vindicate an expectation interest given that the Court pointedly did not grant an award equivalent to what Bhasin expected, which was a renewed contract. As such, it must either reflect a misunderstanding of expectation, or have created one in the name of expediency to justify a particular result. Either way, it appears to have led to the remedy awarded in Callow, which is likewise inconsistent with the orthodox (and correct) understanding of the expectation measure, but less obviously so even though it is even less justified in substance. It is less obviously incompatible with the orthodox understanding of expectation because it does, in fact, grant Callow what he expected, which was that his contract would continue through its full term.[68] The Court’s explanation of this award is that it was necessary on the basis that Callow had abstained from bidding on other work that would have made up for the loss of the maintenance contract with the defendants, because of the impression/belief falsely encouraged by the defendants to the effect that there was nothing wrong with his contract performance to date, and that there was no danger of early termination.[69] On its face, this award might look somewhat more acceptable because it appears to vindicate a positive specific expectation indirectly, rather than a general negative one as described above. However, as I will explain below, this award is in fact even less justifiable in substance than the award made in Bhasin, and does significant violence to the idea of expectation that has prevailed in contract since the 19th century.

B. Exceeding the Restrictions Placed on Reliance

Earlier, I explained the nature of expectation and reliance as distinct bases for contract enforcement, and further clarified that while reliance may act as an alternative method of assessment, it is not an independent interest that gives rise to an alternative remedy.[70] This was made clear in Bowlay, and subsequently reiterated by courts of high authority across the Commonwealth.[71] As such, it would appear settled that no remedy can be provided with respect to reliance when to do so would put the plaintiff in a better position than they could have realistically expected to be in had the contract been performed without any breach.[72] It had also appeared settled that the defendant’s standard of performance, against which the plaintiff’s future position is to be assessed, is the least onerous method or minimum standard that the defendant was obliged to meet under the contract.[73] Unfortunately, however, the majority and concurring minority’s decisions in Callow bring both of these propositions into serious question and leaves us to grapple with two unsavoury possibilities: either the Court is wrong in principle and has sown confusion with its dicta, or alternatively, the law has been inadvertently and drastically changed, and likely not for the better. I prefer the former view of the two, and I will now explain how the majority and minority decisions in Callow run afoul of the two propositions set out above.

I will start with the second of the two propositions above, and the one that may have attracted less attention of the two. In short, it requires that if damages are to be assessed for the defendant’s failure to provide what they promised under the contract, then the value of what was promised is to be assessed to be as little as the defendant could have gotten away with providing or doing, without being in breach.[74] I note that this may seem generous to the defendant, but it appears to be an apt corollary of our requirement that the defendant do no less than what they promised to, to also hold that they need not do more. Further, there would be no obvious limit to the defendant’s liability if it were otherwise. And for perhaps these reasons, even the Court in Bhasin respected this restriction when they declined to award Bhasin the value of a renewed contract, since not renewing the contract was the least onerous thing Can-Am could have done.[75] By contrast, even the minority in Callow appears to have misunderstood this limit on the assessment of expectation damages,[76] which is to say ordinary damages in contract, because they concurred with the majority in the result.[77] That result was an award of damages equivalent to the value of the balance of the contract after early termination by the defendants, despite the fact that the defendants had a contractual right to terminate on ten days’ notice without cause.[78] If the Court in Callow had applied the “least onerous means of performance” limitation, it is clear that damages equivalent to the value of the balance of the contract would not have been available despite the fact that Callow arguably relied on what he was told about the defendants’ satisfaction with his services.[79] This is because it must be assumed that a party in the defendants’ shoes would have simply terminated the contract when it deemed it expedient to do so because the contract contemplated that very choice.[80] This leads to the next proposition that the Court in Callow contravened, as I will explain next, because this disregard for the limits on expected performance has led the Court to contravene the limits applicable to the use of the reliance measure in contract as well.

The mistake made in Callow with respect to the application of a reliance measure is not entirely obvious because of the majority’s insistence that they are not applying a reliance measure at all, and are instead applying the expectation measure to assess Callow’s award. For the reasons outlined above and provided by the concurring minority, that clearly cannot be the case given they have assessed the remedy specifically so as to undo the effect of Callow’s decision not to find or bid on alternative contract work following the defendants’ deception.[81] As such, I take it as given that the award granted in Callow is an award assessed on the reliance measure. If that is the case, it appears plain that the Court in Callow has misapplied the measure by putting Callow in a better position than he could have expected to be in had the contract been performed, because the Court has awarded Callow the value of the balance of the contract left to be performed at the time of termination, despite the fact that the least onerous method of performance for the defendants would have been to terminate. In other words, the Court has supplanted the expectations that a reasonable contracting party could or should have had with respect to a contract that was terminable on notice without cause, and effectively replaced them with the party’s subjective expectations by awarding them damages for a “bad bargain” (i.e. damages for a loss that could, and in this case would, have arisen under the contract even if it had been performed without any breach).

To expand on the potential inevitability of Callow’s loss and why the Court’s decision is so difficult to justify according to established principles, I would draw the reader’s attention to the compelling dissenting judgment of Justice Côté, and the nature of the deception and the parameters of the right exercised by the defendants. The defendants deceived Callow with respect to their satisfaction with his performance, which the Court held would have misled Callow about the risk of early termination.[82] This is not entirely untrue, but as Justice Côté very clearly explains, it is inaccurate in the sense that the defendants’ deception really only pertained to a risk of termination.[83] That is to say, the deception about the defendants’ satisfaction with his services only pertained to one reason among a number of potential reasons for termination. Other reasons include a better offer from a rival operator, an unusually mild winter with too little snow to justify the contract, or a sudden cash flow problem for the defendants that forced them to cut back on services like Callow’s. Of course, in addition to this, there was always the possibility that the defendants could have cancelled for no reason at all, or no good reason, and as such it appears hard to sustain the view that Callow did not just get what he bargained for. And if that is the case, then all that can really be said is that this was a bad bargain, but as explained above, the law is clear that a bad bargain begets no remedy simply for being bad.[84]

IV. Why does this matter?

Having worked through my objections to the Court’s approach to reliance and expectation in the cases discussed above, I can now conclude with why exactly anyone might care. And the reason, in short, is the erosion of certainty in contracts if the expectations embodied in contracts are to be so easily set aside, and if the allocation of risk agreed to in a contract is to be so easily upended. The point about expectations is likely to be clear enough from the discussion above, but the point in relation to risk is this: if recovery for not only bad bargains, but risks that the plaintiff has specifically agreed to bear, are now to be recoverable in contract, then we have crept towards the kind of agreements that the law ought not suffer to enforce. As explained above, recovery for such risks renders contracts a kind of one-sided wager, a “heads I win, tails you lose” type scenario in which one party has effectively become their opposite’s underwriter. Some may retort, of course, that the allocation of risk in some contracts is already one-sided, and I would have to concede that this is true. However, there is some element of choice involved, and when parties like Callow enter the market looking to “buy” remuneration in exchange for their services and do in fact accept certain terms to do so, there is something to be said for the view, according to the old adage, that agreements should stick when “you pays your money, and you takes your pick.”