I was asked to speak about how one should prepare opinions in the aftermath of the Supreme Court’s decisions of the past couple of years.¹ That’s a tall order for a 30-minute talk. As I started to prepare, I quickly found this evolving into yet another talk on the doctrine of equivalents.

There were a couple of reasons for that. We have yet to see the fallout from the Pfaff case on the requirements for an on-sale bar. Although the Supreme Court and more recently the Federal Circuit have pronounced the "totality of the circumstances" test to be dead², I don’t think we’ve seen the last of it yet. The Supreme Court still requires a "commercial" sale or offer for sale to have been made.³ I think it is only a matter of time before some creative litigators use this "commercial" requirement as a vehicle for bringing in the same kinds of evidence that we have seen used before to establish the "totality of the circumstances". We’ll just have to wait and see.

Markman has been beaten to death by the commentators, and I’m afraid I have little to add. The mechanics of resolving Markman issues in a litigation context has little bearing on an opinion-writer. More importantly, I think the main attitudes that the Federal Circuit brings to bear on claim interpretation issues are reflected in its decisions in doctrine of equivalents cases. I find a single major theme running through both the DOE cases and the claim interpretation cases. By discussing this theme in the context of the DOE decisions, I think you can get a flavor of how it operates in the claim interpretation area.

On the other hand, we have a fair number of post-Warner-Jenkinson decisions out of the Federal Circuit now, so we have some raw material to work with in the doctrine of equivalents arena. We have enough that we can extract the rules that are most important in the aftermath of Warner-Jenkinson. So, I decided to focus today on analyzing DOE issues in the context of providing advice to a client.

When I first learned about the doctrine of equivalents, I was told three things. The doctrine existed to prevent "fraud on a patent", whatever that was. "File wrapper estoppel" was a kind of legal antidote to the doctrine of equivalents. And a "pioneering" patent usually got the benefit of a liberal application of the doctrine of equivalents.

I didn’t exactly feel as though I was ready to conquer the world.

As years went by, and I read more and more of the cases, but I can’t say that my understanding of the DOE got much clearer. There are, of course, more legal rules than the three I was taught way back when. But a great many Federal Circuit decisions jumble the different legal concepts together.⁴ So, when I said a moment ago that we can now "extract" some rules from the Federal Circuit cases, "extract" was a key word. The Federal Circuit mixes and matches rules, and says a lot of things that it really doesn’t mean, or that it subsequently retreats from. The Maxwell⁵ case is an example. In Maxwell, the court pronounced a flat rule that unclaimed subject matter in a specification is dedicated to the public, and outside the reach of the DOE. That rule lasted less than two years before the court repudiated it in YBX Magnex.⁶

So, today I’ll try to identify some of the major rules and underlying trends that have survived Warner-Jenkinson. Unlike most talks I’ve heard on the DOE, I’m not going to focus very much on individual cases. I’ll try to paint with a broader brush, and identify the decisions that I think support the broader rules. I’ll try to relate some of these rules to the main theme I see running through the cases. And, I won’t dwell very much on cases that are resolved on what I call "technological equivalence". "Technological equivalence" is the term I use to refer to equivalence as a technological, rather than a legal, matter. It goes to whether a particular technical difference is "insubstantial". You don’t have to consider technological equivalence unless you’ve failed to resolve the matter through application of the legal rules I’ll try to describe. Whether a thing is technologically equivalent to another thing is so fact-dependent that I find many of the cases to be of little practical value in resolving the specific problems. When I analyze DOE issues, I almost always leave the matter of technological equivalence to last.

I’ll start with a suggestion that I hope is practical. At the very beginning of your analysis, you should conduct a kind of a threshold test, a reality check, which is this: you should identify a claim limitation that is not literally met, and correlate that to some specific structure on the allegedly infringing product. Remember, you should have already interpreted the claim as part of the literal infringement analysis, and, also as part of that analysis, you should have already identified the claim elements that were not literally met. So, the first step in the DOE analysis is simply to match up some structure in the product with each of those claim elements. At this stage of the analysis, I don’t apply any rigorous test for technological equivalence.

This suggestion flows directly out of Warner-Jenkinson’s requirement that the DOE is to be applied on an element-by-element basis. This isn’t a new test–it is essentially an affirmation of the Federal Circuit’s pre-existing Pennwalt⁷ rule. The interesting thing at the beginning of your analysis is simply, can you correlate the claim element to some specific structure? All the subsequent analysis flows from this step. If you find yourself thinking, ‘this device works the same way as the claimed device’, rather than ‘this part of the product works the same way as this claim element works’, you are not applying the DOE on an element-by-element basis. And if you can’t do better than that, you may have your answer already.

The next rule I want to mention is another pre-Warner-Jenkinson rule that has survived. The Federal Circuit has expressed this rule better than I can, so I’ll just quote their words:

Wilson Sporting Goods v. David Geoffrey and Associates⁹, In other words, you can’t use the doctrine of equivalents to cover unpatentable subject matter.

Although Wilson Sporting Goods is a pre-Warner-Jenkinson case, the Federal Circuit confirmed its continuing vitality in Litton. In Litton, the Federal Circuit remanded the case with explicit instructions to consider the effect of the Wilson Sporting Goods rule.¹⁰

This rule is completely independent from the issues of PHE and technological equivalence. I’ve found it to be a very powerful and very underused tool. Even many litigators that I’ve worked with seem reluctant to rely on it. I think the reason for this is they worry that the interweaving of validity and infringement issues may confuse the court.

But there are several advantages to using this rule. First, the necessary facts should be available to the lawyer. You need the patent, file history, prior art and the product you are evaluating. All of those things are easily obtainable.

Second, the analysis should be rather straightforward. What you are primarily looking for is the existence of the allegedly equivalent structure in the prior art along with a suggestion to combine it with the remaining elements of the patent claim. I would go so far as suggest, then, if you find the analysis becoming complicated you will probably not be able to safely resolve the DOE question on this basis. At the least, you will have to appreciate the difficulty in explaining it to a judge or jury and you will be able to factor that into the advice you give your client.

The third advantage involves the burden of proof you must consider. This analysis allows you to bring invalidity-type issues, which normally must be proven by the defendant to a clear and convincing standard, into an infringement context where the burden of persuasion is on the patentee and a preponderance of evidence standard applies. You don’t have to conclude that the claims are invalid in order to take advantage of this rule. All you have do to is conclude that an expansive application of the claims through the DOE would encompass unpatentable subject matter. This allows you (and a court) to concede the validity of the claims, if narrowly interpreted and narrowly applied through the DOE.

So, we have one way of analyzing DOE issues without dealing with PHE or technological equivalence.

Another major rule is this: the DOE cannot be applied in a way that would have the effect of writing an element out of the claim. Once again, this is an old rule, but it has taken on unprecedented significant in the post-Warner-Jenkinson era.

Like the rules I discussed earlier, this one is different than PHE, because it does not depend at all on what happened during the prosecution of the patent. It also is unrelated to the analysis of technological equivalence. Instead, it springs directly from the notice function of the claims as articulated in the Warner-Jenkinson decision¹¹.

For one in the position of providing advice to a client, this rule has a clear upside–the facts needed to apply this rule should be known in nearly any situation I can imagine. You need to know the structure of the accused product, plus the contents of the patent and its file history. So, a conclusion resting on this rule should not be subject to attack based on later-discovered facts, if you’ve done your homework. The downside is that the rule is subjective to a degree–applying the rule takes a certain amount of judgement. But my experience is that there are many cases in which this rule is clearly applicable.

The Federal Circuit tends to apply this rule most often (so far) where the claim element at issue is rather concrete, such as in the case of spatial relationships and numerical ranges. An example is the Sage Products case¹². The claim element in question there was to be "at the top" of a container according to the claim, but was inside the container in the accused device. Ethicon Endo-Surgery¹³ is another case where the Federal Circuit applied this rule. The claim in Ethicon required that a thing be "in a staple cartridge", whereas in the accused device it was located somewhere else. Going back a little earlier in time, the Conopco¹⁴ case involved a numerical limitation, a ratio of "about 40:1". An attempt to stretch this value to around 160:1 was found to violate this rule¹⁵.

In Sage Products, the court essentially found that "top" meant "top". "Top" is a simple, Anglo-Saxon word that represents a concept that your average 5-year-old can understand. The patentee should have known that "top" means "top", and could easily have selected a different word or phrase if he had intended something different. In a case like that, the Federal Circuit isn’t going to rescue the patentee. The court would call rescuing the patentee in that way "writing the limitation out of the claim".

I mentioned before that I find a major theme running through the recent DOE cases that informs many of those decisions. It involves the notice function of the claims, and is this:

This is equally true of claim interpretation cases.

The rule I’ve just been discussing, that you can’t use the DOE to erase a limitation, is this theme in action. The court is starting to use this rule more and more often to defeat DOE claims. I think you’ll see it even more in the future.

Look how the Federal Circuit expressed it in Sage Products:

What I want to focus on is the distinction between "structure" and "function" in the second sentence. This distinction is rapidly gaining in significance. The traditional function-way-result test for equivalence is essentially a functional one. The emphasis is on what the allegedly equivalent structure does, not what it is. And the emphasis on function alone has troubled several of the judges at the Federal Circuit for some time. You might recall Judge Lourie’s dissent in the Hilton-Davis case¹⁷. He pointed out that applying the function-way-result test alone would lead to the conclusion that ibuprofen would infringe a patent on aspirin under the DOE, even though the two compounds are not chemically similar. Having a rule that says you can’t stress functionality to the point that the claim limitations become superfluous is a way of dealing with Judge Lourie’s concerns. So the court is evolving a more stringent structural test. In essence, this rule is becoming a kind of "super-equivalents" rule. That is, it is a threshold test for equivalents that is over and above the traditional function-way-result test for equivalents. What this rule does is to require a certain amount of structural similarity to exist before one even gets to the functional equivalents test. In other words, lack of structural similarity can trump the functional similarity that is the focus of most equivalents analyses under the traditional function-way-result test. And, in fact, the Northern District of Georgia has interpreted Sage Products and Ethicon in exactly this way, in Baracuda International Corp. v. Hoffinger Industries Inc.¹⁸ In that case, the court had denied summary judgement motions for noninfringement, and the jury actually found the accused device to infringe under the DOE. The judge granted the defendant’s motion for JMOL, effectively reversing himself and wiping out the jury verdict. The court noted that it "erroneously focused exclusively upon functional equivalence as opposed to structural equivalence", in earlier denying the summary judgement motion.

You might keep in mind that until Warner-Jenkinson, the Federal Circuit had another tool to help decide close DOE cases. That tool was the inquiry into the subjective intent of the accused infringer¹⁹. The court could resolve close questions by deciding whether the defendant was an enterprising designer or an evil copier. The judges most hostile to the DOE, you might recall, were the ones who were trying to set up an inquiry into the equities as a threshold test for applying the DOE. Now that the Supreme Court has taken this tool away, the emphasis on the notice function of the claims has replaced it. The Supreme Court certainly helped this along by emphasizing the importance of the claims’ notice function in the Warner-Jenkinson opinion.

So, this is a powerful rule that I think will be applied more often as the primary check on the application of the DOE. We see it being applied especially in cases where a claim limitation is concrete and easily understood by even lay persons–numbers and spatial relationships that are capable of being expressed unambiguously.

You’ll notice that I haven’t talked about technological equivalence or prosecution history estoppel yet. That’s intentional. A major point I hope you take with you today is that there are legal rules that you can apply to resolve DOE issues without getting deep in the morass of equivalence or PHE issues.

But I can’t talk about the DOE without discussing prosecution history estoppel just a little.

An very interesting case on PHE is Vehicular Technologies Corp. v. Titan Wheel International Inc²⁰. In Vehicular Technologies, the Federal Circuit split the world of PHE in half by distinguishing two types of estoppel. It identified the first type of estoppel as "prosecution history estoppel", but didn’t define what was encompassed by that. I’ll try in just a minute. The court did define the second type of estoppel, which I refer to as a Type II estoppel. A Type II estoppel prevents a patentee from asserting that a certain structure is equivalent as a technological matter to a particular claim element. The court took pains to state that this estoppel was different than and separate from PHE. What is PHE, then? The implication is that PHE prevents a patentee from asserting that certain types of structures infringe under the DOE, even if those structures are technologically equivalent to the claim element. PHE therefore applies irrespective of the technological equivalence of the accused structure. Technological equivalence therefore does not need to be considered if there is a PHE. On the other hand, the Type II estoppel is in effect a predetermination that certain types of structures are not technologically equivalent. PHE issues do not need to be considered.

You may think that the distinction between these estoppels is difficult to grasp. I think it will be especially difficult for the Federal Circuit to maintain. Many of the decisions the Vehicular Technologies court relied on as support for its Type II estoppel referred to it as "PHE"²¹. As I looked at some of those cases when I prepared for this talk, I wasn’t sure whether they should be classified as PHE or Type II estoppel cases.

But, it probably doesn’t matter whether the court maintains this distinction or not. The key point to remember, and to use when you are analyzing an issue, is that estoppels come in different flavors. Once you recognize that, you can apply a number of simpler rules to resolve a great many estoppel issues. So, I think the distinction between the two types of estoppels is a terrific analytical tool. In the end, it makes no difference whether you call the estoppel "PHE" or something else.

The Type II estoppel arises from statements made by the patentee about:

1. What the structure of his claim element is (more commonly, what his structure is not);
2. What the function(s) of his claim element are;
3. The way his claim element performs its functions; and
4. The results that are obtained with his claim element.

You’ll very easily recognize these as being the three elements of the Graver Tank equivalence test, plus the structural "super-equivalents" test I spoke about earlier. To look for Type II estoppels, you examine the patent and the file history for statements as I just identified. Then you apply those statements to the structure in the product that you’ve already identified as the potential equivalent to the claim element. You are looking to see whether that patentee has already said that a structure that looks like yours, or functions like yours, or functions the way yours does, or gets the results your does, is not equivalent to his claim element. If he has, there is a Type II estoppel.

Statements creating these estoppels are easy to find, and they appear everywhere. The background statement in the application itself, where prior art is discussed, is a very fertile ground. In the same vein, information disclosure statements are very important. We have at least two cases now where gratuitous statements in prior art statements purporting to distinguish the prior art created Type II estoppels. They are Jack Ekchian v. Home Depot, Inc. and Litton. Comparative examples are often a gold mine. Then, of course, are the arguments made in response to art rejections, where the patentee often makes very clear statements about particular features being different than his claims, or working in a different way, or leading to different results. Although I haven’t tried to do a rigorous analysis of the cases, I suspect that most if not all cases in which the estoppel is based on argument, rather than amendments, will turn out to be Type II estoppel cases.

A couple of good cases to look at for examples of how a Type II estoppel works are Vehicular Technologies and a pre-Warner-Jenkinson case, Hoganas AB.²² Vehicular Technologies involved a patentee’s statements about the functions of a claim element. I’ll say more about that in a moment. In Hoganas AB, the patentee said that a certain claim element was of such a size that it didn’t function the same way as a particular in the prior art. The court refused to apply the DOE to a different structure that worked the same way as the prior art.

So, the beauty of the concept of a Type II estoppel is that it allows us to separate this big hairy estoppel beast into more manageable pieces. You can look for specific statements as to structure, function, way and result that might prohibit the patentee from asserting that a particular thing is equivalent to an element of his claim.

Last month, the Federal Circuit issued an opinion in Sextant Avionique, S.A v. Analog Devices, Inc.²³ I mention this because it is potentially an important case on PHE. It creates a rule that is likely to have a fate similar to that of the Maxwell case.

Sextant Avionique involved a claim amendment. It was not clear why the amendment was made, so the court applied the Warner-Jenkinson case and presumed that it made been made for purposes relating to patentability. There is automatically an estoppel in such a situation, so the question became, what was the scope of the estoppel? Judge Lourie found that the estoppel was complete–an amendment that is presumed to be related to patentability is entitled to no range of equivalents whatsoever²⁴.

I have my doubts if other panels will follow such a draconian rule. Note also that it only applies when it is presumed that an amendment is made for reasons related to patentability. It doesn’t apply when the prosecution history makes it clear that an amendment was made for patentability reasons.

What is interesting is that Judge Lourie based this decision squarely on the notice function of the claims²⁵.

Finally, I’ll talk for a few moments about some ways to look at technological equivalence, or "insubstantiality" of the differences. I want to tie this discussion in with the broad theme I mentioned before that I see running through the DOE cases. That theme, you’ll remember, is that the public notice function of the claims is the paramount policy consideration. I find at least four corollaries to this theme:

1. The old function-way-result test is not enough. This is reflected in the Sage Products/Ethicon line of cases I talked about before.
2. The risk of uncertainty lies with the patentee, not the public²⁶.
3. Ties go to the accused infringer.
4. All information needed to interpret and apply the DOE should be in the public record²⁷.

The fourth of these corollaries comes into play in those decisions in which the court has to decide whether equivalence exists on a technological level. The basic question here is "How do we measure equivalence?" More and more, the answer is becoming, "by finding out what the patentee told the PTO about it". So, statements in the application or prosecution saying "this element does this and does that" will become the test for whether some different structure performs the same function as that element. Statements that a claim element works in a certain way become the test for whether the "way" portion of the equivalents test is met. And the same goes for statements about results²⁸.

This line of thinking has reached a head in Vehicular Technologies. In Vehicular Technologies, the patentees said in their specification that a particular a claim element did several things. So, the patentee in effect had identified several functions for the element. The majority found that in order for the function arm of the equivalents test to be satisfied, an alleged equivalent structure had to perform all of those functions. That generated an outraged dissent from Judge Newman, who called the new rule the "all advantages" rule²⁹.

I don’t know if the court will retreat from that position. If Judge Newman gets an opportunity to author an opinion, I’m sure she will try to repudiate the rule or hold the Vehicular Technologies case to its facts. But I don’t think it’s such a terrible rule. It doesn’t read limitations into the claim for purposes of evaluating literal infringement. And to find that a structure is equivalent to a claim element because it performs only some of its functions strikes me as arbitrary. The public would have to guess what functions would be the "key" ones for weighing equivalence.

A couple of other cases are interesting because they illustrate another way in which the court is emphasizing the notice function of claims. These are Tanabe Seiyaku Co. v. ITC³⁰ and Insituform Technologies Inc. v. Cat Contracting Inc.³¹

In each of these cases, the patentee elected to present claims that were very narrow in certain features. In Tanabe, for instance, a mixture was defined in terms of precisely five pairs of materials. In three of the pairs, both members were precise compounds–no generic description appeared at all. In the other two pairs, one of the members was defined in terms of a narrow genus. The court refused to apply the DOE even to what is known in chemistry as a "homologue". Now, when you start out doing patent work in the chemical field, the classical example of equivalence that everybody uses is that of homologous compounds. "Ethyl" is equivalent to "methyl" because ethyl merely contains an additional carbon atom that usually doesn’t affect properties very much. So, for a court to deny coverage for even a homologue reflects an extremely narrow view of the DOE. In Tanabe, it is quite clear that the court depended heavily on the mere narrowness of the claim language itself to defeat the equivalence claim. That is, by choosing to claim narrowly, perhaps even artificially so, you run the risk of a court refusing to grant any range of equivalents because of the mere fact your claims are narrow. To quote the court³²:

There are some other factors, peculiar to the Tanabe case, which the court pointed to in its decision. But I think that they probably didn’t make a difference in the end. We are likely to see an evolving rule to the effect that claim elements that are narrowly drafted will be accorded a narrow range of equivalents, irrespective of the lack of prior art or other constraints³³. The Tanabe decision is really nothing more than another aspect of the court’s increased reliance on structural similarity as a super-equivalents test.

I’ll close with some statistics. The professional statisticians will sneer at these. I don’t claim to be professional in either my sampling methods or treatment of data³⁴. But I did identify 29 post-Warner-Jenkinson decisions in which the Federal Circuit ruled on DOE issues. Of these, the Federal Circuit found a product to infringe under the DOE in only 5 cases. In 21, the Federal Circuit found no infringement under the DOE. In four, there was a remand. The discrepancy in the numbers is because there were multiple products in at least one case.

I also looked how the individual judges handled these cases³⁵. Patentees had a winning record only with Judges Archer and Smith. And those two judges handled only a small number of decisions. The numbers for Judges Newman and Rich are especially interesting. When Judge Newman was new to the bench, she was the Will Rogers of the judiciary–she never met a patent she didn’t like. But since Warner-Jenkinson, she is 9:5 against the doctrine of equivalence. Judge Rich is 5:2 against. As you might expect, Judges Clevenger, Plager, Lourie, Rader and Mayer are heavily anti- the doctrine.

I set out to do two things today. First, I wanted to try to separate out some of the individual legal rules that play into the doctrine of equivalents analysis. We should realize that the analysis isn’t just "insubstantiality" of the differences and PHE. There are other rules that we can apply to focus our evaluation and hopefully simply it. By applying these rules, you often can resolve DOE questions more easily and better articulate your conclusions.

The other thing I tried to do was to identify the paramount importance of the notice function of the claims in the recent case law, and show how that plays into some of the rules that the court is applying in DOE cases.