I'd have to say that I don't think either the pregnant cow case or the discussion of unilateral vs. bilateral contracts is germane to the SCO-Linux situation. First, contract law varies by state, so the cow case may only apply (if it's even relevant and still valid) in the state from which it came. Other states may very well not have such law. Second, the GPL is not a contract, either unilaterally or bilaterally.

Unsurprisingly, courts are left puzzled about when to void for mistake. One casebook says, "The case law in this area is confused beyond reconciliation. Courts cannot agree on what is 'mutual' and what is 'unilateral' and in many jurisdictions cases can be found in which relief is granted in both situations, however they are defined."

But what if Caldera/SCO had been in a contract, under which they released the code and later discovered code they didn't know was in there?

I found some material that appears to indicate that even if the GPL were a contract, and even if Caldera had released materials under the GPL without awareness that certain code was in the release, the pregnant cow case wouldn't necessarily save them.

First, here is the case, Sherwood v. Walker, 66 Mich. 568, 33 N.W. 919. As he correctly points out, it's a Michigan case, not a US Supreme Court decision, so it applies specifically to Michigan. It is a famous case, however, often taught in contract law, because it addresses the issue of when a mutual mistake allows a party to get out of the contract. It was a replevin action, [PJ: note that the original link to http://www.personal.psu.edu/ faculty/w/d/wdo2/ba243/contracts/indexp24.htm no longer resolves, so I have introduced another definition page], which is an "action to recover an item of personal property."

The facts of the case, explained here, were these:

Seller Walker owned breeding cows, worth between $750.00 and $1,000.00 and barren cows, worth about $80.00. Buyer Sherwood inspected an apparently barren cow, Rose 2nd of Aberlone, and decided to buy her. A price was agreed on 5.5 cents per pound but before the exchange of money and cow, Walker found Rose was pregnant and refused to part with her. The court said that if both parties thought the cow was barren (a question for the jury), the contract was voidable on grounds of mutual mistake.

Further, if both parties in Sherwood had been taking a risk on whether the cow might be fertile, the contract would not have been voidable [.doc file]. When will relief be granted for mutual mistake? Here is one explanation:

A mistake of fact, shared by both parties, which goes to the basis of the bargain, entitles either party to rescind the contract. (See, Sherwood v. Walker, 33 N.W. 919 (Mich. 1887).) If both parties are mistaken and neither is at fault or both are equally to blame, mistake may prevent the formation of a contract. Unilateral mistake does not render a contract voidable except where the mistake is due to the fault of the other party or the other party knows or has reason to know there is a mistake. (Rest.2d § 153).

When BOTH parties make a material mistake, it is voidable if:
Enforcement makes K [contract] burdensome on one party
During K, the fact in question was not in doubt
NOTE Re:Chen.the court will not say it, but if a contract is executed, they will not change the positions of the parties, rather, they will find another reason not to overturn contract. . . .A basic mistake on subject matter voids a contract.

So if there is a misunderstanding as to a material term and neither party knows or has reason to know of the misunderstanding, there is no contract. Rest. 2d Rest. 2d, §20(1). . . . Where one party either knows of the misunderstanding or objectively should know of the misunderstanding, the contract is interpreted against them. . . .

Business transactions. It will be presumed in business settings that the parties intend to be bound -- for example, even if one party made an offer in jest and the other party seriously and reasonably accepted the offer, the contract will be binding.

And what about the issue of Caldera/SCO's duty to investigate what it was releasing? Was the company able to investigate if the code it was releasing contained the allegedly infringing code? To look carefully at the terms of the GPL so as to understand them? Well, folks, that's what open source means. Anybody can look at the code. No one else was in a position to know if their proprietary code was there, because it was closed source code that only they could identify as theirs. If they didn't bother to investigate, despite having full opportunity to do so, would a a court reward their lack of due diligence? Obviously we can't say for sure, since courts themselves are so divided. But what do you think?

If you are interested, there is a tutorial on mutual mistake here. And you can read more here.

The attorneys originally quoted seemed to be thinking about SCO's relationship with its own customers under contract with them, judging from what they said. And while, as one of them honestly pointed out, it's impossible for anyone to say for sure how a judge or jury might rule in a mutual mistake case, even if this were such a case, the simple fact is that Caldera released code to the public at large , not just to its own customers, and they did it under the terms of the GPL, which isn't a contract. Anyone could download it off its site, and it came with the GPL attached. So where is the contract they can now ask a court to void? If there is one certain thing, it's that you can't distribute GPL code under any other terms than the GPL and you can't hit reverse and take it back later.

Update: For the sake of historians, since the article by 2009 has been moved, here is what the attorneys said about the pregnant cow case, from Stephen Shankland's article of June 12, 2003, Will a pregnant cow decide SCO's future?:

The issue isn't as clear-cut as either SCO or its opponents would have it, said John Ferrell, an intellectual-property attorney with Carr and Ferrell. "If anybody tells you they have the definitive answer, they're crazy," he said.

But he'd give the edge to SCO in the situation, not because of its interpretation of the GPL, but because of a legal principle stemming from the 1887 sale of a pregnant cow in Michigan. That case established the so-called doctrine of mutual mistake, under which a contract can be nullified if two parties - in this case SCO and a company using Linux - misapprehended the true nature of what was in the contract.

Mark Radcliffe, an attorney with Gray Cary, also believes SCO has a case. On the argument that the company opened the Unix code, he said, "I think that's a tough argument to make. The fact that distributing software with proprietary code in it into open source - I think a court would have difficulty swallowing that one," he said. "This may be a test of the open-source licence." ...

Moglen likened the situation to selling a book. SCO's argument would be like a publisher saying, "Sorry, had we read the novel we would have known it was a good novel and would have charged more for it," Moglen said. But Ferrell said that with the doctrine of mutual mistake, SCO could employ to its advantage an argument that it didn't know what exactly was in the Linux product. "If neither party knew what it was getting under the contract, it may be possible for SCO to avoid the terms of the contract," the contract in this case being the GPL....

How the SCO case is resolved will affect how the computing industry treats open-source software and its underpinnings, Ferrell said.

"This particular issue is really important to the GPL, and (its resolution) is just inevitable," Ferrell said. "At some point we're going to need to have a judge sit down and sort it all out."

• The GPL is a "unilateral contract to establish a limited public license trust".
• summary of his case
• On Lessig's Blog in 2004 on GPL being a contract, not a license and this: "The GPL is DOA in a court of law. The open source licenses at OSI as well as the Creative Commons License are also preempted by Section 301 of the Copyright Act." [PJ:The latter is precisely the failed SCO argument in SCO v. IBM.]

Tell Me This is "Just a Coincidence"

In older contracts, Microsoft agreed to pay all legal fees for volume license customers who got sued because of Microsoft, but only up to the value of the software they bought.

Under the new provision, which took effect March 1, Microsoft removed the liability cap in intellectual property suits and altered other parts of the agreements that potentially expand its liability.

The company also expanded its product warranties for licensing customers from 90 days to a year and expanded the minimum notice given to customers regarding software audits from 15 days to 30 days.

"The former clause allocated too much risk to third parties," said Laura Didio, an analyst at the Yankee Group. "Intellectual property issues relating to Microsoft software are entirely in the control of Microsoft."

"In real world terms, the changes won't likely dent Microsoft's legal budget. The indemnity provision rarely comes into play," said Morris Kremen, associate general counsel for licensing at Microsoft. Customers, though, hated the old one.

Excuuuse me, but what jumps off the page is that "the indemnity provision rarely comes into play", meaning this is a big FUD issue, for starters, and Microsoft wasn't indemnifying anybody in any real way either, until the SCO attack was planned:

A phantom menace

Microsoft, however, was defending against a phantom menace. Kremen could recall only one case where a plaintiff brought a copyright infringement action against Microsoft's customers rather than the multi-billion dollar company. (In that case, IBM and Microsoft actually picked up the defense anyway and obtained a verdict in favor of their customers).

Neither Kremen nor Mark Bolender, senior attorney at Microsoft, could recall any cases where customer sued a Microsoft customer for software security breaches or personal injury relating to a Microsoft product.

Even if a suit were brought, the old indemnity provisions, which are similar to contractual provisions issued by other software companies, would likely have been adequate to protect licensees.

"It is hard to imagine a royalty claim that is higher than the cost of the software," Kremen said.

Sorry, Microsoft. Friendships are built on trust. Even business relationships are built on a measure of it . This announcement inspires less trust, not more, because of the coincidental timing, and, frankly, Ms. Didio being trotted out again.

The fundamental problem Microsoft faces isn't Linux. Their problem is their business depends on customers being tech-stupid and FUD-inexperienced. That only can work for so long. At one time, nobody knew anything about computers, so they relied on Microsoft products to ease them into it and hold their hands. That just isn't the case any more, except for my mom. Her generation is the last one that believes such FUD and will put up with poor quality and poor service and proprietary ways, because they imagine the BSOD they are staring at happened because they did something wrong as they typed and they imagine proprietary software is normal, like air.

When she reads about the new license, she'll think it's great, until we talk it over. When I read about the new license, I just snort. And that is Microsoft's real problem. They lost our trust. We grew up. And we know how to use computers. We know now that it's valuable to be able to look at code yourself. Some of us went into business, and we are the new office managers and other types that decide what software to buy. Microsoft treated us like we were criminally-inclined. And that told us who they are. And their licenses for XP onward were so mean-spirited, so weighted against the customer, that we got disgusted. GNU/Linux gave us a choice, and we took it. How do you reverse that? I can't think of any way, can you? No one goes back to a bad relationship of their own free will. Especially because in the interim, we fell in love with the new software.