Lew Alpert and his production company, Alpert Productions, Inc. (together "Alpert"), won a judgment against Goldhirsh Group, Inc. ("Goldhirsh") in United States District Court for the Southern District of New York (Lawrence M. McKenna, Judge ), on a counterclaim for tortious interference with business relations. Goldhirsh moved for judgment as a matter of law, arguing that the evidence was insufficient to support the jury's verdict on such a claim. The district court denied the motion, see Goldhirsh Group, Inc. v. Alpert, No. 94 Civ. 3922, 1996 WL 122441 (S.D.N.Y. March 20, 1996) (unpublished Memorandum and Order); Goldhirsh now appeals.

Alpert is a freelance television producer. Goldhirsh is the publisher of Inc. Magazine ("Inc."), a periodical for and about growing companies. In late 1991, Alpert approached Goldhirsh with the idea of producing sixty-second television "vignettes" on subjects of interest to small businesses, attaching Inc.'s name to the TV spots, and selling them to sponsors. The first thirty seconds of the spot would be a business tip from Inc., and the second thirty seconds would be a related commercial for the sponsor.

Proceeding cautiously, Goldhirsh agreed to two pilot vignettes, to be sponsored by the Avis Corporation ("Avis"). The vignettes were well received, and Alpert and Goldhirsh decided to expand the idea to include a new series of half-hour television shows entitled "Inc. Magazine Getting Down to Business" (along with more vignettes). Under their agreement, Alpert was allowed to use Inc.'s name for the TV shows. Alpert would pay all production costs for the new shows and would also search for more sponsors. If the revenue from the sponsorship payments exceeded costs, Alpert and Goldhirsh agreed to a 50-50 split of the profits.

Avis sponsored all the new shows. They were broadcast on CNBC. While everything went well, Alpert's production costs exceeded revenues, so there was no profit. Undeterred, Alpert and Goldhirsh agreed to yet another run of shows. For this third series, however, Alpert secured American Telephone & Telegraph, Inc. (AT & T) as a sponsor. AT & T promised to pay $575,000 to sponsor eight episodes and related vignettes. The shows were produced and scheduled to run on CNBC on weekends from late September 1993 to early January 1994.

In November, while the third series of shows was underway, Alpert and Goldhirsh were already discussing the possibility of a fourth run of shows, and Alpert was talking to potential sponsors for that fourth series. Alpert, however, was now weathering severe financial troubles. In fact, Alpert had fallen behind in paying CNBC for the airtime devoted to the third run. CNBC called Alpert and threatened to pull the show scheduled for Thanksgiving weekend, 1993, unless he came up with all the money he owed CNBC. Unable to make the payments, Alpert took his troubles to Goldhirsh. Three Goldhirsh employees--James Spanfeller, publisher of Inc.; Howard Kaplan, associate publisher of Inc.; and John Reardon, vice president and controller of Inc.--were worried that, if the show was pulled, it would damage Goldhirsh's long-term relationship with AT & T, the sponsor of the third run. Goldhirsh lent Alpert $180,000, which Alpert promised to repay when it received the final sponsorship payments from AT & T. At the same time, Alpert and Goldhirsh reached a formal agreement to produce the fourth run of shows.

AT & T paid Alpert, as anticipated, in early 1994. But Alpert did not forward $180,000 to Goldhirsh, as he had promised. Instead, he poured the money back into production and development costs and paid off various debts associated with the show. When Goldhirsh inquired about the money, Alpert lied and said that AT & T had not yet paid him. But Reardon, who had frequent contact with McCann Erickson, AT & T's advertising agency, found out that AT & T had indeed paid Alpert.

Reardon confronted Alpert. Alpert confessed that he had been paid by AT & T, but had spent the money. Reardon demanded immediate repayment of the $180,000 debt to Goldhirsh, warning Alpert that, if he did not repay, Goldhirsh might bring "criminal fraud" charges against him. Reardon then called Kaplan and Spanfeller, his confreres at Goldhirsh, telling them that Goldhirsh would be terminating its agreement with Alpert and pulling the Inc. name from all of Alpert's shows. Reardon also proposed to his confreres that Goldhirsh file "criminal fraud" charges against Alpert unless he promptly repaid the loan.

By now, Alpert was well on the way to securing sponsorship contracts with AT & T and the United States Postal Service ("USPS") for the fourth series of shows. Reardon, Kaplan, and Spanfeller decided that, because both AT & T and USPS had longstanding relationships with Inc., Goldhirsh should inform both of them of its decision to pull Inc.'s name from all of Alpert's shows. They agreed that Kaplan would call McCann Erickson (AT & T's advertising agency) and Young & Rubicam (USPS's advertising agency) and tell them that (1) Goldhirsh was terminating its agreement with Alpert and pulling the Inc. name from his shows; (2) the matter was now "legal," and Goldhirsh could therefore say no more; and (3) if anything changed, Goldhirsh would call them.

Kaplan then called two people at each agency (four people in all), reaching three in person and leaving a message for one. Kaplan testified that he told each of these people only that (1) Goldhirsh was terminating the agreement and pulling Inc.'s name; (2) the matter was now "legal"; and (3) if anything changed, Goldhirsh would call them. After these calls, Alpert's prospective sponsorship deals with AT & T and USPS fell through.

Goldhirsh sued Alpert in the United States District Court for the Southern District of New York, alleging fraud, breach of contract, money had and received, conversion, and unjust enrichment--all based on the unrepaid $180,000 loan. Alpert counterclaimed against Goldhirsh for breach of contract and tortious interference with business relations--both based on Alpert's loss of the potential sponsorship deals, a loss that Alpert attributed to Goldhirsh's reproachful calls to the advertising agencies.

The jury found for plaintiff Goldhirsh on its money had and received claim, awarding it $180,700. The jury, however, also found for defendant Alpert on the tortious interference counterclaim, awarding Alpert $750,000 in tort damages.

Goldhirsh, defendant on the counterclaim, moved for judgment as a matter of law or, in the alternative, a new trial, on the ground that no rational jury could have found on the evidence submitted that Goldhirsh had tortiously interfered with Alpert's business relations. The district court denied the motion. Goldhirsh now appeals.

Goldhirsh's basic contention is that the evidence was insufficient to support a rational finding that Goldhirsh tortiously interfered with Alpert's business relations. We agree.

We review the denial of a motion for judgment as a matter of law de novo. See Valley Juice Ltd. v. Evian Waters of France, Inc., 87 F.3d 604, 613 (2d Cir.1996); Song v. Ives Labs. Inc., 957 F.2d 1041, 1046 (2d Cir.1992). While it is hornbook law that we may not substitute our view of the evidence for the jury's when that evidence allows multiple legitimate inferences, see 9A Charles A. Wright & Arthur R. Miller, Federal Practice and Procedure Civil 2d § 2524 at 255-56 (1995); H.L. Moore Drug Exch. v. Eli Lilly & Co., 662 F.2d 935, 940 (2d Cir.1981), we cannot let a verdict stand where it is grounded on "such a complete absence of evidence ... that the jury's findings could only have been the result of sheer surmise and conjecture." Mallis v. Bankers Trust Co., 717 F.2d 683, 688-89 (2d Cir.1983). As noted in Michelman v. Clark-Schwebel Fiber Glass Corp., 534 F.2d 1036 (2d Cir.1976):

[i]f ... after viewing all the evidence most favorably to [the prevailing party], we cannot say that the jury could reasonably have returned the verdict in his favor, our duty is to reverse the judgment below. The jury's role as the finder of fact does not entitle it to return a verdict based only on confusion, speculation or prejudice; its verdict must be reasonably based on evidence presented at trial.

Michelman, 534 F.2d at 1042; see also County of Suffolk v. Long Island Lighting Co., 907 F.2d 1295, 1311 (2d Cir.1990).

The line between permissible inference and impermissible speculation is not always easy to discern. When we "infer," we derive a conclusion from proven facts because such considerations as experience, or history, or science have demonstrated that there is a likely correlation between those facts and the conclusion. If that correlation is sufficiently compelling, the inference is "reasonable." But if the correlation between the facts and the conclusion is slight, or if a different conclusion is more closely correlated with the facts than the chosen conclusion, the inference is less reasonable. At some point, the link between the facts and the conclusion becomes so tenuous that we call it "speculation." When that point is reached is, frankly, a matter of judgment. See generally United States v. Shonubi, 895 F. Supp. 460, 482-88 (E.D.N.Y.1995), rev'd on other grounds, 103 F.3d 1085 (2d Cir.1997).

Referring specifically to legal reasoning, it has been noted that

[t]rust in inference is simply the belief that if there is a firm basis for the starting point the derived judgment is acceptable. The difference between speculation and inference lies in the substantiality of the evidence constituting the premise. Inductive reasoning claims the premises constitute some evidence for the conclusions and in law we speak in terms of the probability and likelihood that the premises buttress the conclusions.

Wisconsin Mem'l Park Co. v. Commissioner of Internal Revenue, 255 F.2d 751, 752 (7th Cir.1958).

Thus, inference and speculation are not different processes, but labels attached to the same process, and used to describe a level of certainty we have in a conclusion based on the quantity and quality of the facts from which that conclusion is drawn. In law (as elsewhere), this level of certainty derives, as discussed, from (1) the strength of our belief that the facts supporting the conclusion are true; and (2) the weight of the correlation between the facts and the conclusion. In this sense the difference between inference and speculation is truly practical rather than logical, and best understood as "a function of the capacities and limitations of the system's mechanisms for creating deference to jurors' decisions and thus for generating acceptable verdicts." Charles Nesson, The Evidence or the Event? On Judicial Proof and the Acceptability of Verdicts, 98 Harv.L.Rev. 1357, 1371-72 (1985).

Applying these principles here, we find that the jury's verdict on Alpert's tortious interference counterclaim is based not on legitimate inference, but on impermissible speculation.

A defendant becomes liable for tortious interference with a plaintiff's business relations when four conditions are met: (1) there is a business relationship between the plaintiff and a third party; (2) the defendant, knowing of that relationship, intentionally interferes with it; (3) the defendant acts with the sole purpose of harming the plaintiff, or, failing that level of malice, uses dishonest, unfair, or improper means; and (4) the relationship is injured. See Purgess v. Sharrock, 33 F.3d 134, 141 (2d Cir.1994).

Alpert's theory in this case rests foursquare on the Kaplan phone calls to the potential sponsors' advertising agencies. According to Alpert, Kaplan surely "must have" told the advertising agencies that Alpert was a "crook" (or words to that effect), that Goldhirsh was going to file "criminal fraud" charges against Alpert, and that AT & T and USPS should not do business with Alpert. These disparaging comments, Alpert asserts, were the dishonest, unfair, or improper means giving rise to his tort claim. But, as Goldhirsh points out, there was absolutely no evidence that Kaplan made any such comments.

Most significantly, Alpert never called a single employee of the advertising agencies to testify about the calls made to those agencies. Indeed, the only witness who testified about the calls was Kaplan, and his testimony was clear, categorical, and cogent: he told the agencies only that Goldhirsh was terminating the agreement and pulling Inc.'s name from the shows, and that he could not discuss the matter further because it was "legal." Alpert's attorney pressed Kaplan on cross-examination as to whether he had made any reference to "criminal fraud" during the calls. Kaplan responded: "No, I did not." True, Alpert's attorney argued in summation that Goldhirsh "told AT & T and the [USPS] that Mr. Alpert was a person who would commit criminal fraud.... Common sense tells us that what happened was that Mr. Kaplan told [the agencies] Mr. Alpert was a crook, don't do business with him, pull out of those deals." The summation, of course, is not evidence, and arguably was an improper comment since there was no evidence to support that comment.

Thus, no witness testified that Kaplan made any disparaging remarks in his calls to the advertising agencies. And one witness--Kaplan--testified that no disparaging remarks were made. The jury certainly had the right to disbelieve Kaplan. But, "[i]f all of the witnesses deny that an event essential to the plaintiff's case occurred, the plaintiff cannot get to the jury simply because the jury might disbelieve these denials. There must be some affirmative evidence that the event occurred." 9A Wright & Miller, Federal Practice and Procedure Civil 2d § 2527, at 288; see also Martin v. Citibank, N.A., 762 F.2d 212, 217-18 (2d Cir.1985). As Judge Learned Hand stated, while "a party having the affirmative might succeed in convincing a jury of the truth of his allegations in spite of the fact that all the witnesses denied them ... a verdict would nonetheless have to be directed against him." Dyer v. MacDougall, 201 F.2d 265, 269 (2d Cir.1952). To phrase it yet another way: "[w]hen the testimony of a witness is not believed, the trier of fact may simply disregard it. Normally the discredited testimony is not considered a sufficient basis for drawing a contrary conclusion." Bose Corp. v. Consumers Union of the United States, Inc., 466 U.S. 485, 512, 104 S. Ct. 1949, 1966, 80 L. Ed. 2d 502 (1984). Therefore, even if the jury thoroughly disbelieved Kaplan, that disbelief is insufficient to support a verdict for Alpert in the absence of affirmative evidence that Kaplan told the advertising agencies that Alpert was a crook or that Goldhirsh would be bringing "criminal fraud" charges against Alpert.

The district court held, and Alpert now argues, that there was sufficient circumstantial evidence to support a legitimate inference that Kaplan had denigrated Alpert in his phone calls to the advertising agencies. According to Alpert, such evidence includes: (1) that Reardon told Alpert, in his phone call demanding money for the unrepaid loan, that Goldhirsh might file "criminal fraud" charges against Alpert; (2) that Reardon proposed, in his call with Kaplan and Spanfeller, that Goldhirsh file criminal fraud charges against Alpert; (3) that Reardon, Kaplan, and Spanfeller talked about going forward with the shows without Alpert; and (4) that George Gendron, Inc.'s editor-in-chief, sent an "e-mail" to Reardon painting Kaplan as a "loose cannon."

None of these items of evidence--whether alone or in the aggregate--give rise to a reasonable inference that, in his calls to the advertising agencies, Kaplan called Alpert a crook or referred to "criminal fraud." What Reardon earlier said to Alpert in his call demanding payment, while perhaps not wholly irrelevant, in and of itself cannot support the conclusion that Kaplan said the same thing to the agency employees in a later, separate call. That Reardon repeated this language to his colleague Kaplan in their conference call also says nothing about whether Kaplan later told the advertising agencies that Alpert was a crook, particularly when it is recalled that Reardon explicitly told Kaplan not to say anything negative about Alpert to the agency employees. And Gendron's e-mail criticism of Kaplan--that Kaplan was "reckless[ ]" and had a " 'slam dunk' attitude"--was undeniably a comment about Kaplan's overall performance on the project; such a general criticism is insufficient to generate a "reasonable" inference that Kaplan said something in particular about Alpert in his calls to the agencies.

Alpert also contends that Kaplan "likely" made disparaging comments about Alpert in the calls to the advertising agencies because Kaplan was "motivated by fear and anger." This assertion is unsubstantiated by testimony or other evidence. And finally, Alpert relies on the fact that no advertising agency representatives ever called back in an effort to revive a project from which they would have earned large commissions. This circumstance, Alpert contends, supports an inference that Kaplan's statements were so compelling that they definitively terminated the relationship. But it does not necessarily follow that, because Kaplan's statements may have been "compelling," they were improper. And, after all, Kaplan specifically told the agencies that, if there was a change in Goldhirsh's position, Kaplan would call them. We cannot, therefore, read much into the agencies' failure to call back Goldhirsh.

In short, there simply was no affirmative evidence that Kaplan called Alpert a crook or made any reference to "criminal fraud" in his calls to AT & T's and USPS's advertising agencies. "[T]he jury's findings could only have been the result of sheer surmise and conjecture." Mattivi v. South African Marine Corp., 618 F.2d 163, 168 (2d Cir.1980). The question that haunts this record like the ghost of Banquo is: why did Alpert never call as a witness any one of the four persons to whom Kaplan allegedly made disparaging statements? Though circumstances may sometimes support a reasonable inference as to the content of a conversation, even in the absence of direct testimony by a participant, such content cannot reasonably be inferred on this record.

Because Alpert failed to prove that Goldhirsh intentionally interfered with his business relations through sheer malevolence, or dishonest, unfair, or improper means--an essential element of his claim--the district court should have granted Goldhirsh's motion for judgment as a matter of law.1

We have considered all of Alpert's arguments and find them without merit. The district court's denial of Goldhirsh's motion for judgment as a matter of law on Alpert's counterclaim for tortious interference with business relations is REVERSED, and the case is remanded with instructions to enter judgment for Goldhirsh on that claim.