National Tax Institute ("the Institute")—an organization that puts on conferences—sued Topnotch, a resort hotel at Stowe, Vermont, with a capacity of 90 rooms. The suit concerned the meaning of a contract provision between the parties as to the number of rooms to be provided to the Institute at a reduced rate. The district court granted summary judgment in favor of Topnotch, and this appeal by the Institute followed.

Since 1995, the Institute has run an annual week-long seminar for accountants and attorneys at Topnotch. The dates have been in October, typically at or near the peak foliage season when travel in and to New England tends to be popular. Under its contract with Topnotch, first negotiated in 1994 and since amended several times, the Institute secures a block of rooms from Topnotch at a discounted "group rate" and then provides the rooms to those who attend the seminar, charging them a marked-up room rate and a seminar fee.

The original contract negotiated in September 1994 used a standard Topnotch agreement for group rates; but the terms critical to the present dispute appear in a detailed addendum drafted by the Institute's representative. The contract specified, for October 1995, a block of 40 rooms for one week at $90 each per night. The addendum obligated Topnotch to re-offer the contract for the next ten years "with identical terms and conditions (excluding group rates) should the [Institute] decide to book another meeting in any year from 1996 through 2005."

The addendum then listed specific October dates and a group rate for each of the next ten years. The rates set in 1994 (later revised) ranged from $99 to $132 per room, depending upon the year, with a $92 rate in any year that the Institute elected a later week in October. The renewal option given to the Institute applied "separately for each year." Also in the addendum appeared a further provision—around which this case centers—stating: "Additional rooms may be blocked at the group rate subject to availability."

In November 1995, the parties renegotiated the dates and the size of the block reserved for the 1996 conference; they increased the Institute's block to 45 rooms in exchange for moving the dates to ones preferred by Topnotch. In August 1996, the dates for the 1997 conference were again changed by agreement, raising the block to 50 rooms. The 1996 revisions made clear that the 50-room block would be available through 2005, stating: "The group shall have a first right of refusal [sic] on the following dates at the following group rates for which the hotel shall set aside fifty (50) rooms per night for each peak night...."

In September 2000, the Institute, seeking to enlarge its attendance, wrote to Topnotch requesting 70 rooms for all future years through 2005, relying upon the "additional rooms" clause quoted above. Topnotch replied that the Institute had no option for more than 50 rooms; it also remarked that Topnotch did not intend to extend the arrangement beyond the specified years because the Institute's contracted-for rates fell short of the group rates Topnotch was willing to offer after 2005. Compared with Topnotch's retail rates, which ranged from $230 to $740 depending on the year and type of room, the (revised) flat rates offered to the Institute were much discounted, starting at $132 in 2001 and rising to $144 by 2005.

A few months later, the Institute sued Topnotch in Massachusetts state court, seeking (along with damages) a declaration that the "additional rooms" clause entitled the Institute during the term of the contract to reserve, at the group rate, any rooms not yet booked by others. Topnotch removed the case to federal court on diversity grounds. Following discovery, including experts' reports, Topnotch moved for summary judgment, saying it had full discretion whether to offer at the group rate more than 50 rooms.

In a decision filed on May 27, 2003, the district court sided with Topnotch. The opinion rested primarily on "the rule of the last antecedent" and on the conclusion that Topnotch's position was the only reasonable construction of the "additional rooms" clause. The court deemed that provision unambiguous, adding that if the proffered extrinsic evidence were considered, it too would support Topnotch.

On the Institute's appeal, our review of the grant of summary judgment is de novo, drawing inferences in favor of the Institute. Filiatrault v. Comverse Tech., Inc., 275 F.3d 131, 134 (1st Cir.2001). Because the parties have plausibly treated Massachusetts law as governing on substantive issues, we may follow their lead. AccuSoft Corp. v. Palo, 237 F.3d 31, 40 n. 5 (1st Cir.2001). We agree with the district court that the contract is not ambiguous—only Topnotch has advanced a plausible reading of the disputed provision. See Lohnes v. Level 3 Communications, Inc., 272 F.3d 49, 53 (1st Cir.2001). The proffered extrinsic evidence, if considered, would not alter the result.

If one looked solely at the bare language of the amended contract, it might well be difficult to know which of the two proffered readings to adopt for the disputed sentence: "Additional rooms may be blocked at the group rate subject to availability." The Institute says that a room is "available" if not already reserved by another customer; Topnotch says that it is up to the management to decide whether to make the group rate "available" for more than 50 rooms.

The Institute drafted the addendum, and courts sometimes construe uncertain contract language against the drafter. Kerkhof v. MCI WorldCom, Inc., 282 F.3d 44, 51 (1st Cir.2002). But the canon is a qualified one, Principal Mut. Life Ins. v. Racal-Datacom, 233 F.3d 1, 4 (1st Cir.2000), a default rule that arguably has more force where the parties differ in sophistication or where standard forms are used (e.g., insurance contracts). In any event, the canon has little to do with actual intentions and should only be used, as a last resort, if other aids to construction leave the case in equipoise. Id.

The district court invoked a different canon, one presuming that the last antecedent term (here, "group rate") is what the parties intended to be modified by the qualifier that follows (here, "subject to availability"). See Deerskin Trading Post v. Spencer Press, Inc., 398 Mass. 118, 495 N.E.2d 303, 307 (1986). The canon formally applies, but it is not conclusive; and anyway the opaque sentence in question—inserted seemingly at random between unrelated provisions—was not drafted with any noticeable precision.

Often, related provisions may cast light on meaning. Here, the Institute points out that in the original contract, another term provided for "[g]roup rates [to] apply for days before and after the group dates, at the discretion of the hotel's management and subject to availability." This reference to "discretion" presents a contrast, abstractly helpful to the Institute, with the clause in question; but, again, the document is too haphazard to justify assigning much weight to the discrepancy.

What is clear is that the number of rooms to be made available at the discounted rate was a matter of importance expressed in precise figures. When in 1996 Topnotch increased the block to 50 rooms, the addendum was reworded to say: "The group shall have a first right of refusal [sic] on the following dates at the following group rates [listed by year] for which the hotel shall set aside fifty (50) rooms per night for each peak night." At the same time, the Institute also received in writing a fixed allocation of five rooms at the group rate for two nights just before and after the conference.

Even more helpful is an understanding of the context and nature of transactions like this one. Agreements, especially commercial arrangements, are designed to make sense. If one reading produces a plausible result for which parties might be expected to bargain, that reading has a strong presumption in its favor as against another reading producing an unlikely result (e.g., windfall gains, conditions that cannot be satisfied, dubious incentives). See Fishman v. LaSalle Nat'l Bank., 247 F.3d 300, 302-03 (1st Cir.2001).

Discounts are common in business; often they simply reflect the reduced per-unit costs for large orders, due to economies of scale. In other instances (airline advance purchase fares, for example), the motive instead may be to secure a base of customers at bargain rates while extracting more revenue per customer from others who will pay more (e.g., for last-minute reservations or the option to change flights without penalty). This is an especially attractive strategy where unused capacity saves little cost and cannot itself be "saved."

In practical terms, charging everyone a high price may leave many unfilled seats or rooms, and charging everyone a discounted price may attract too little revenue to cover total costs. Often the key in such situations—as true for hotels as for airlines—is for the seller to reserve enough uncommitted capacity for at least some high-paying customers while a possibly lesser but still vital contribution to revenues is obtained from the discount clients. The economists' term is "price discrimination."

From this perspective, it made perfect sense for Topnotch to reserve some but not all rooms for the Institute at low prices and to increase the number of discounted rooms in exchange for the Institute's acceptance of less desirable dates. The original number was 40; the sequence of amendments shows that it was twice raised, first to 45 and then to 50, but only in exchange for the Institute's accepting other dates than those originally agreed upon. This freed up more uncommitted rooms for periods in which high-paying guests would be more numerous.

By contrast, the idea that Topnotch would have agreed to rent all its rooms at about $90 to $130 for high foliage season dates for the following ten years makes no sense.1 In the normal course of business, no rational hotelier would likely make such a contract, encompassing even the rooms that would probably (and predictably) be rented at three or four times the group rate. The Institute's reading of the contract assumes just such an improbable bargain.

If this exhausted our clues as to the parties' intent, Topnotch would easily prevail. The Institute, however, points to additional evidence "extrinsic" to language and inferred purpose. The term "extrinsic evidence" is imprecise but includes proof of negotiations between the parties, their post-contract conduct, and general trade practice. See Den Norske Bank AS v. First Nat'l Bank of Boston, 75 F.3d 49, 52-53 (1st Cir.1996). The opening question is whether and when extrinsic evidence may be examined and allowed to alter the conclusion drawn from language and context.

The old-fashioned and still common answer is that extrinsic evidence may be considered if language is ambiguous but not otherwise (and whether language is ambiguous is a question for the judge). Lanier Prof'l Svcs. v. Ricci, 192 F.3d 1, 4 (1st Cir.1999); Den Norske, 75 F.3d at 52-53, 55. This should be taken with a grain of salt, partly because extrinsic evidence may in fact reveal an ambiguity not otherwise patent, Smart v. Gillette Co. Long-Term Disability Plan, 70 F.3d 173, 179-80 (1st Cir.1995), but also because language may point only slightly in one direction and extrinsic evidence strongly in another.

In all events, here the Institute has no evidence of actual negotiations worth mentioning; its few proffered instances prove nothing.2 Evidence on course of performance—if the amendments in 1995 and 1996 adding rooms to the block are treated as such—tends to help Topnotch. But each side did obtain experts about industry practice, and the Institute's expert proffered what at first glance looks like extensive evidence in its favor. Upon closer examination, the picture changes.

The Institute's expert, John Foster, is a partner in an Atlanta law firm and has impressive credentials as a hotel sales manager and as a lecturer on law affecting hotels and conventions; he also has personal experience in negotiating contracts both as a hotel manager and as a lawyer. He gave his opinion in the Institute's favor on the disputed "additional rooms" provision based primarily upon "custom and usage in the industry."

Foster agreed with Topnotch that hotels do engage in "yield management" (a polite term for price discrimination). But, he continued, it was and is the custom for "hotel owners and managers to use specific and unequivocal language ... if they want to restrict requests for reservations beyond physical availability"; thus (in his opinion), the disputed language must refer only to physical availability. Foster quotes or appends hotel and cruise ads offering bargain rates to consumers (e.g., three days in Florida at $100 a night).3

Such consumer ads tell us nothing about how to construe a detailed, but badly drafted, long-term contract directed to a specific type of transaction (convention bookings). Nor do the supposed drafting habits of hotel managers provide much information about an addendum which was not drafted by the hotel but by a customer who had an adverse interest. If Foster had offered evidence that hotels customarily gave bargain rates during high season for all their rooms, this might be a different case; but he didn't (and they probably don't).

Finally, Foster's expert report also presents responses from two hotel sales managers whom he polled by e-mail, both saying that Topnotch should have attached restrictive language to the sentence. This material did not purport to be responsible poll evidence and would almost certainly be inadmissible at a trial as merely an effort to smuggle in views of individuals not present for cross-examination. See Fed.R.Evid. Rule 703 & Advisory Committee's Note to 2000 Amendment. It adds nothing to Foster's own opinion.

In sum, even if the Institute's extrinsic evidence were considered, it could not justify a different result. The critical disputed sentence did no more than make clear—as Topnotch has argued—that the limited number of discount rooms to be provided (first 40, then 45, later 50) could be enlarged if the Institute wanted more rooms and if Topnotch chose to provide them at the same discount rate. This added nothing to the bargain beyond expressly keeping the door open for a mutually agreed enlargement; no wonder the sentence was left without elaboration.

Affirmed.