The hearing lasted about two days and focused mainly on procedural matters. The Palo Alto weather was miserable. A cold, torrential rain wrapped us in a gloom worthy of Bleak House, the Dickens novel about the grotesque inefficiency of the English chancery courts. My most vivid memory was when the arbitration panel allowed Tom Barr to question Ninomiya. We all perked up because we thought something interesting was about to happen. Ninomiya was Fujitsu’s top executive for hardware and software development and had made many decisions central to the case. Barr wanted Ninomiya to describe exactly how Fujitsu was currently using IBM’s programming materials in its software development.
It should have been a moment of high drama. Barr was a great courtroom lawyer and no doubt had looked forward to this confrontation. Instead he found himself moving in slow motion with a stone-faced enemy who wouldn’t engage. He would ask a question of Ninomiya. It would be translated from English to Japanese. Ninomiya would then mumble some short answer in Japanese. One of the translators would translate his response into English. Another translator would sometimes object to the first translator’s translation. All three translators would then get into a dispute while the rest of us sat there impatiently. Barr would then characterize the answer as unresponsive and ask the same question again, phrased somewhat differently.
This went on for hours. It was excruciatingly boring. The high point came as Barr was trying to get Ninomiya to explain how Fujitsu had decided which programs belonged on the DP list. Barr wanted to force Ninomiya to acknowledge in front of the panel that the DPs contained copied material. But Ninomiya parried with a double negative that gave the translators quite a workout. As I recall, the English version was something like, “Fujitsu put a program on the DP list when it could not be concluded that the program had not been influenced by an IBM program.” Even in translation, it was obvious that Ninomiya had been well coached by counsel. (Note the words “influenced by” rather than “copied.”) Recently I learned that one of the young Morrison & Foerster lawyers had nicknamed him “Garbo” because he was so elusive.
During the next ten months, like a traveling circus, we met in various cities around the United States, Japan, and even Canada. The parties fought about everything. For example, the panel needed some education on operating system software and the technology underlying the conflict, but these seminars were difficult to arrange because the parties so deeply distrusted each other and worried we would somehow be brainwashed in this process. Although our instructors were always neutral experts who weren’t affiliated with either party, the lawyers fought about who would educate us and they fretted about what these experts might say. Each side insisted on having a representative pres-ent at every session so they could carefully monitor everything we were told.
The parties also submitted an astounding volume of letters, memoranda, documents, and exhibits to educate us about their conflict. The parties’ submissions also gave us a much clearer sense of the legal questions we would need to address, although the answers each proposed were diametrically opposed. Fujitsu argued that it had broad rights to IBM information under the Externals Agreement, and IBM argued that its obligations were extremely narrow. IBM argued that copyright law provided very broad protection for its software, while Fujitsu claimed that copyright law applied narrowly, if at all, to IBM’s operating system software.31
Unfortunately, copyright law itself didn’t provide much guidance. For one thing, it wasn’t clear whether U.S. or Japanese law applied. But either way, the case posed very difficult questions about the scope of protection.
If copyright law applied to operating system software at all—and we assumed it did—it obviously prohibited line-by-line copying. IBM’s Final Report gave some compelling examples of this. (One “smoking gun” involved a subroutine, or loop, which an IBM programmer had named after his wife, Lucy. The corresponding Fujitsu program included the very same loop—also named “Lucy.”) But we didn’t know how important those examples were in the context of the full program. In other instances, IBM claimed that Fujitsu had simply paraphrased the IBM material. This, too, would be a violation.
But we had a growing sense that the heart of IBM’s case involved a much more subtle claim: that Fujitsu had copied the “structure, sequence and organization” of an IBM program without actually quoting or paraphrasing any lines of IBM code. When, if ever, would this be enough to prove the “substantial similarity” required for infringement? Deciding this issue would require us to apply the most elusive doctrine found in copyright law, the distinction between “ideas,” which are not protected, and “expression,” which is.
Even with literary works, it can be hard to draw that line. The following hypothetical is often used in law school copyright courses: Imagine that the descendants of William Shakespeare still owned a copyright on his play, Romeo and Juliet. Would the Broadway musical West Side Story violate the copyright? The actual words in the two productions are entirely different. But the story line in West Side Story was plainly copied from Shakespeare’s play. The plots are similar. (Two families are feuding. Boy from one feuding family falls in love with girl from the other. They decide to secretly marry and run off together. After a series of mishaps, all ends tragically with death of both.) How similar do the two plots have to be before the later work violates the copyright of the original? Many years ago Learned Hand, a distinguished federal circuit court judge, said that drawing this line requires an “ad hoc judgment” based on a careful comparison of the two works. There’s no rule to tell you how to do it. It all depends on the particular facts and a detailed comparison of the two works.
And there was the rub. IBM had brought scores of claims against Fujitsu, each concerning a different software program. To draw the line between idea and expression, we’d have to conduct a detailed analysis of each disputed program and its IBM counterpart. How similar would the sequence, structure, and organization have to be to prove a violation? Each program contained hundreds of thousands (in some cases millions) of lines of code. Would we have to examine each program, function by function, data area by data area, line by line? That would take years.
IBM hoped to limit this process by delivering a knockout blow in the form of a motion for summary judgment, but to my mind this hope was nearly delusional. One of the requirements for summary judgment is that no material facts are in dispute. The idea/expression distinction and most of Fujitsu’s other defenses clearly involved “mixed questions of law and fact.” So we denied IBM’s motion32—and its request for triple license fees, which we held to be punitive—and ruled that we would have to analyze the disputed Fujitsu programs one at a time, by comparing each with its IBM corresponding program. We told the parties that a detailed comparison of the first pair of programs would be due in forty-five days.
The parties immediately began to fight about which pair of programs to compare first. There were three pairs under consideration. On behalf of IBM, Barr argued that the panel should start with a pair that he thought would reveal blatant line-by-line copying. Fujitsu, naturally, wanted to start with a pair that was deep in the gray area, involving similarities in structure, sequence, and organization. The third example had features of both.
While the panel was struggling with this issue, Barr suddenly changed course. “Why choose now, before you’ve seen the evidence?” he asked the panel. “IBM can submit detailed memoranda on each of the three Fujitsu programs comparing them to the corresponding IBM program. You can choose later, after you’ve read our analysis and anything Fujitsu wants to submit,” he purred. Perhaps, fearing that the panel was about to choose the wrong set of programs, he wanted to make sure we saw the worst of Fujitsu’s copying before we made a decision. Whatever his motives, I saw jaws dropping in horror among his troops. Barr was improvising. He obviously hadn’t discussed this offer with the enlisted men and women who would actually be looking at millions of lines of code. From their ashen faces, I figured some of them wouldn’t be seeing their families for a whi
le.
Meanwhile, I had my own, more basic concern. Was conventional arbitration a sensible way to resolve this conflict? While we were looking backward at old disputes, new ones were cropping up. IBM was still investigating Fujitsu’s alleged misdeeds and adding claims, and Fujitsu was still developing new programs. I feared the new conflicts would multiply faster than we could dispose of the old ones. I thought, This is crazy. Like Sisyphus with his boulder, the panel might never reach the end of its labors. Moreover, a case-by-case process would miss the forest for the trees.33 It wouldn’t directly answer the core question: To what extent can Fujitsu use IBM programming material in the future?
I began to believe that the most effective way to help the parties was to adopt a radically different approach, one that looked forward as well as backward. But designing a new process was not part of an arbitrator’s job description.
I have never been good at hiding my impatience. Even when I’m not speaking, my facial expressions often give me away. Shortly after the panel denied summary judgment, we held a session at a beautiful Japanese resort in Hakone National Park, in the shadow of Mount Fuji. During a break in the hearing, I decided to take a walk in the woods nearby—and ran into Tom Barr. I may have made some offhand comment to the effect that we all had “our work cut out for us.” If I didn’t say it, that certainly was what I was feeling. Barr may have sensed my frustration with what lay ahead, and perhaps even shared it.
In any case, he said something that would change the course of the entire proceeding: “Why don’t you and Jack put your heads together and figure out some more efficient way to cut through all this crap and get this matter resolved?”
I was stunned by his directness but leaped at the opportunity. “Do you mean IBM would be open to some sort of mediation process, with Jack and me helping the parties see if some kind of deal might be negotiated?”
Barr said, “Yes! I can assure you IBM would participate.”
I found this an appealing prospect. Barr didn’t say why he had not included Macdonald in his suggestion, but that made sense to me. It would protect Macdonald, who, after all, had been chosen by both parties. It is highly unconventional—and some would say improper—for a sitting arbitrator to change roles and become a mediator. The two roles are completely different. The former is a decision-maker who imposes a solution on the parties; the latter is a facilitator who helps the parties negotiate a solution with each other. If this mediation didn’t work, either party could probably disqualify Jack or me from returning to an arbitral role. Leaving Macdonald out of the mediation process was the solution: he could carry on as arbitrator.
I immediately went to find Jack. He liked the idea. The two of us went to Macdonald and told him about Barr’s suggestion. Macdonald thought the whole thing was a bit unorthodox and probably doomed to failure, but worth a try—as long as Fujitsu was willing.
I naïvely thought, Of course Fujitsu will be willing. It’s obviously a great idea to look for a global solution that doesn’t require comparing scores of programs! But to my surprise, it wasn’t that easy.
When Jack and I first proposed mediation to the Morrison & Foerster team, we told them that the idea had initially come from the other side. Bob Raven and Dave Nelson listened with poker faces and solemnly said they would consult with their client and get back to us. Eventually Fujitsu agreed to participate, on one condition: that they not be required to meet or negotiate directly with IBM. They would participate only if Jack and I met with both sides separately, gener-ated proposals for the parties, and refined those ideas through shuttle diplomacy.
As a mediator, this is not my preferred method of mediating. I prefer working primarily with the parties and their lawyers together, in the same room. That way, each side hears directly from the other side their story of what happened in the past and what their needs and future interests are. In most cases, I believe, this is the best way to solve hard problems and promote mutual understanding.34
But here, I was forced to realize that was impossible. We had one party that was very cautious—indeed, fearful that somehow IBM was going to trick them again. From Fujitsu’s perspective, they had already been burned once in a negotiation with IBM: the 1983 Agreements had resolved nothing and IBM had used them to assault Fujitsu. Now here was IBM suggesting mediation. How could that be good for Fujitsu?
Looking back, I realize that the entire arbitration process must have seemed terribly alien and risky to Fujitsu. Fujitsu probably resented being forced to participate in a process so dominated by Americans, and in which it was so dependent on American lawyers. I recently learned from a Fujitsu executive that, in the eyes of some of his colleagues, the arbitration was “four against one”: (1) IBM was an American company; (2) Cravath was an American law firm; (3) the panel consisted entirely of Americans;35 and (4) Morrison & Foerster was an American law firm.
Nelson and Raven showed great sensitivity to this problem. They were always at pains to keep their clients fully informed and to strengthen their sense of control.
One ingenious idea helped improve the quality of communication between Morrison & Foerster and its clients. Bob Raven suggested to Fujitsu, “Lend us one of your young in-house lawyers. Have him move to San Francisco and work out of our San Francisco offices. You have my word he will be allowed to sit in on all of our internal meetings when we discuss the strategy we are developing for the arbitration. Your Mr. Katoh would be perfect.” Masanobu Katoh was fluent in English and served from then on as a key link between the American lawyers and the Fujitsu top brass, most of whom were engineers. With a smile, Katoh recently told me that when he was assigned to the San Francisco post, one of his Fujitsu colleagues told him he was being sent as a “hostage” to the Americans.36
Cultural differences permeated the mediation process as well. For Jack and me, there was a big difference between the kinds of conversations we could have with each side. When we met with the IBM team, lawyers Tom Barr and Dan Evangelista, IBM’s general counsel, had broad authority to act on their client’s behalf. They and the IBM executives were comfortable exploring new ideas with us, telling us what they thought, and even making decisions on the spot. That was never the case with Fujitsu. Bob Raven and Dave Nelson attended all the mediation sessions but never claimed authority to make decisions for their client. Naruto would typically speak on behalf of Fujitsu but he would usually just report Fujitsu’s reaction to what had been discussed at the last meeting. He might indicate which aspects were acceptable and which ones were not. Once I understood Fujitsu’s objection, I would invariably suggest some modification and ask whether this would solve Fujitsu’s problem. The Japanese would typically say nothing one way or the other. They would be completely nonresponsive. I was initially very frustrated by this. I wondered, Why can’t they simply say what they think? Why won’t they work with us?
After one of these sessions, in which the Japanese had once again declined to respond to one of my “brilliant” ideas, Nelson pulled me aside and gave me a coaching tip. When you are negotiating with a Japanese company, he said, don’t expect movement at the table when something new is proposed. The Japanese negotiating team needs time to discuss the idea within the team and build a new consensus before they can proceed. In a Japanese organization, no one at the negotiating table typically has the authority to deviate from the position announced at the outset of a meeting. “Bob, be patient,” Nelson said. “In due course, after they have discussed it internally, I suspect they will come around.”
Barr was probably aware of these cultural differences and Raven’s more limited authority. At one point in the mediation, Jack and I suggested that the two top lawyers get together and try to work out a couple of details. They happened to meet on Valentine’s Day, in an IBM conference room that some staffer had decorated with hearts and valentines. When Barr entered the room and saw the decorations, he didn’t say “Be My Valentine.” He quipped, “Remember the St. Valentine’s Day Massacre!” (That was when Al Capone
’s men lined up six members of an opposing gang and shot them.) Raven laughed. Barr went on to make a concrete proposal to deal with the questions the panel had asked counsel to work on. Perhaps wanting to needle Raven, Barr said to him, “Bob, in all seriousness, I want to assure you that the IBM Corporation has given me full authority to make a deal on the two questions we’ve been asked to address. What authority has Fujitsu given you?” Raven, who had a good sense of humor, grinned and said, “Tom, I want to assure you that I have full authority to report back to Fujitsu everything you say today.”
What did we actually do during the mediation phase? Ultimately we designed an entirely new process that was custom-made for this dispute. Here’s how we got there.
As a first step, Jack and I tackled the mess that was the DP list. This turned out to be fairly easy because the parties were now both motivated. We had already dispelled IBM’s fantasy with regard to getting triple damages for the “new DPs” that Fujitsu had wanted to add. Fujitsu, meanwhile, still wanted immunity for those programs and was prepared to pay.
After less than four weeks of shuttle diplomacy, we had a deal: Fujitsu would add 295 programs to the DP list and would pay IBM a lump sum of $30 million for past use. Going forward, as with the original DPs, it would pay a semiannual license fee for each customer installation. Thus, in one fell swoop, the panel and the parties were freed from the task of slogging through those particular programs case by case.
Bargaining with the Devil Page 19