Wednesday 7 January 2015

Partial Priority Again - Enlarged Board Referral

I have written a few times in the past about the issue of partial priority under Article 88 EPC, and in particular how it relates to the issue of so-called 'poisonous' divisionals and priority (see herehereherehere and here; there is also an interesting discussion of the issues on the IPKat here and here).

As I mentioned in my July 2013 post on the Nestec case at the EPO, in my view the only way to resolve the poisonous priority issue would be for a referral to be made to the EPO Enlarged Board. In the Nestec case, the issue unfortunately did not get to the stage of discussing partial priority because the patent was revoked on other grounds. It was, however, inevitable that the same issue would arise in another case, and the clear inconsistency between the Board of Appeal decisions in T 1222/11 and T 1877/08 (among others) about how to interpret G 2/98 would have to be resolved somehow. That case has now appeared.

In appeal proceedings for T 557/13 (Infineum), relating to EP0921183 (see the EP register here), the Board of Appeal have recently decided to refer one or more questions to the Enlarged Board of Appeal. As yet, the only indication of this is the minutes of the oral proceedings that were held on 17 December 2014, and no questions have yet been decided on, although there are some suggestions. The questions will, however, most certainly be about the extent to which partial priority can be claimed and what effect this has on priority applications becoming prior art under Article 54(3) EPC.

The appeal relates to a decision by the opposition division to revoke Infineum's patent (which resulted from a divisional application) based on the claims not being fully entitled to their priority and, as a result, the published parent application becoming Article 54(3) prior art. The opposition division cited G 2/98, and referred to the apparent requirement for the elements of a claim having to be "separable alternative embodiments" (see point 3.1.2 of the ground of the decision; my emphasis). Since this was not the case with claim 1 as granted, the OD decided that the claim was not entitled to priority for its entire scope and was only entitled to the application's filing date. T 1877/08 was cited in support, but the contrary decision in T 1222/11 was not published until the following year. During appeal proceedings, the proprietor made extensive arguments (see here for the grounds of appeal) about the poisonous priority issue, and why it should not happen, but did not refer to T 1222/11 until later on. They did, however, cite one of my posts (this one), which "discusses the issues at the heart of the case and highlighted some of the differing views prevalent among practitioners" (page 6 of the ground of appeal), and which apparently provided them with suggestions for a few helpful arguments.

T 1222/11 was first raised in the proprietor's further submissions in May 2014, together with some further articles on the issue. The Board of Appeal then issued a summons and, in their preliminary opinion, stated that "the issue of novelty apparently may hinge on the question of whether or not [...] the patent in suit can validly claim a right of priority as regards the use of the product of Example 1 as cold flow improver, considering in particular that said use is merely embraced by the more generic definition of the invention according to Claim 1 at issue, rather than being clearly defined as a distinct alternative of the invention defined in Claim 1" (point 4.2.2 of the provisional opinion dated 26 November 2014). The issue in question is therefore clearly one that would be decided differently, depending on whether the approach according to T 1222/11 or T 1877/08 is taken.

The proprietor has already suggested some questions that could form the basis for a referral:
Where a claim of a European patent or patent application encompasses more specific subject matter disclosed in an earlier application from which priority is claimed, for example because the claim in question contains generic term(s) or generic formula(e) encompassing both the subject matter disclosed in the priority application and alternative(s) thereto:
1. Is the expression "provided that it gives rise to the claiming of a limited number of clearly defined alternative subject matters" from Point 6.7 of Enlarged Board opinion G2/98 to be applied as the legal test for entitlement to partial priority under Article 88(2) and (3) EPC? 
2. If the answer to 1 is 'yes', how is the test from Point 6.7 to be applied in this situation, and in particular, how is a claim containing generic term(s) or generic formula(e) to be interpreted as relating to a limited number of clearly-defined alternative subject matters?
3. If the answer to 1 is 'no', what is the legal test for partial priority of a claim containing generic term(s) or generic formula(e), and how is it to be applied to such a claim?
In the situation defined above, where:
(i) the European patent was, or the European patent application is, a divisional application ("Divisional") and the claim in question fails the legal test for partial priority as clarified above;
(ii) the parent European patent or patent application ("Parent") claims the same priority and, as published, discloses specific subject matter also disclosed in the priority application and encompassed by the claim of the Divisional; and
(iii) the Parent and Divisional designate the same EPC contracting states:
4. Does Article 76(1) EPC nevertheless prevent different effective dates being attributed to the claim of the Divisional and the priority-entitled specific subject matter in the Parent?
5. If the answer to 4 is 'no', is the term 'European patent applications' in Article 54(3) EPC to be interpreted as extending to the Parent application, thereby rendering the Divisional lacking in novelty over its Parent ?
Answering these questions would resolve the issue, but we may have a little while to wait before we get the answer.

UPDATE 13 August 2015: The interlocutory decision in T 557/13 is now out. See my post on the subject here.

43 comments:

  1. Thanks David
    Well Spotted
    Not sure I like the questions, but I suspect that is because I want the answer to be "of course partial priorities are allowed". On a more serious note, I would want the questions to refer to the FICPI memorandum, which surely has to be addressed by the EBA.

    ReplyDelete
  2. Judging from the board's communication of 26 november 2014, T 557/13 is about (a patent granted on) a divisional application. The divisional and its parent disclose an example comprising the features of claim 1. Priority is valid for the example, but not for claim 1. The example as disclosed in the parent is therefore novelty destroying 54(3)-prior art for claim 1.

    Could someone explain how partial priorities could ever help in this case?

    As I understand, claim 1 would have to be split in (A) the example and (B) the complement of the example. The effective date for (A) would be the priority date, so the parent could not be used against it. The effective date for (B) would be the filing date, but example 1 in the parent does not destroy its novelty.

    The priority document discloses (A), so the effective date of (A) is indeed the priority date. This is G 2/98.

    What is the effective date for (B)?
    The parent application does not disclose (B).
    How could (B) have the filing date of the parent application as the effective date?

    Is the test for benefiting from the filing date of the parent application less strict than the test for benefiting from a priority date? That would be surprising.

    ReplyDelete
    Replies
    1. Rather than answering your question directly I suggest you read some of the previous posts on this subject (linked above), since it looks like you have not understood the issue in question.

      Delete
    2. The third paragraph should make clear I understand the basic idea.

      There seems to be an issue with that idea.

      Delete
    3. There certainly is, otherwise a referral to the EB wouldn't be needed. I still think you need to read a bit more. I'm not going to go over the issues again here, as I have explained them at length in previous posts.

      Delete
    4. So I have read all your posts now and a question remains. If you're tired with me, I hope someone else is interested.

      Your proposed solution is explained here. The core idea is this:
      "The concept that a broad claim can theoretically be divided into two parts, one of which is entitled to priority and the other which is inevitably novel over the priority disclosure, is itself sufficient to overcome the problem even if the exact form of wording is not easily derived from the application. This is because the collective scope of the resulting two imaginary claims would be exactly the same as the original claim, and Article 88(2) EPC indicates that there is no need to actually rewrite the claim to be two separate claims. There is therefore no need to find the right form of wording to satisfy added matter requirements because no amendment is needed."

      So (in the context of T 577/13) you split the claim of the divisional mentally in a first part (A) that has the priority date as effective date and a second part (B) that is the complement of (A).

      Why does (A) have the priority data as effective date? Answer: because it is directly and unambiguously derivable from the priority document (G 2/98).

      My point is that for this to work, also (B) should have some effective date (or it does not make much sense to talk about "inevitably novel"). This you have not considered in any of your posts, as far as I can tell.

      Could (B) have the effective date of the parent application?

      I say it cannot, because Art. 76(1) works quite similarly to Art. 87(1): something can only get the date of the parent application if it is directly and unambiguously derivable from it. That is not the case for (B).

      One could say: OK, let's take for (B) the whole claim. But then (B) is not new anymore.

      (Do I think that the solution would work the other way around, i.e. if the divisional is used as 54(3)-prior art against the parent? No, I do not think it does. I think any imaginary part can only play a role if it can be assigned some effective date. For that it has to be disclosed in a priority document (G 2/98), or in the application itself.)

      Delete
    5. Let's consider a slightly different example to see what partial/multiple priorities mean.

      1st priority document P1 discloses copper.
      2nd priority document P2 discloses metal.
      The application claims metal.

      Now intuitively one might say that the claim has priority P1 for the imaginary part copper and has priority P2 for the imaginary part "metal but not copper".

      G 2/98 allows "copper" to have priority from P1, so this is OK.
      However, G 2/98 does not permit "metal but not copper" to have priority from P2. P2 discloses "metal" but not "metal but not copper".

      So intuition could be misleading.

      From T 1222/11 it would actually follow that, since P1 is the first application for copper, priority cannot validly be claimed from P2 "in so far as it covers" copper (point 10). So P1 shoots a hole in P2: one can claim priority from P2 for metal, but not in so far as it covers copper. This is another way of saying that one can claim priority from P2 for "metal but not copper". The approach of T 1222/11 hence fixes the above intuition, but it remains the case that this is contrary to G 2/98 ("priority of a previous application in respect of a claim in a European patent application in accordance with Article 88 EPC is to be acknowledged only if the skilled person can derive the subject-matter of the claim directly and unambiguously, using common general knowledge, from the previous application as a whole").

      Delete
    6. The whole point of the issue is that G 2/98 does allow priority from the application as filed for B, while A gets priority from the earlier application. The examples in the FICPI Memorandum, which was expressly approved in G 2/98, has precisely the kind of example you illustrate. The question that needs answering (although I think it is obvious once you see the Memorandum) is whether you need basis for B in the application as filed. My argument is that you do not, and I think this is supported by G 2/98.

      Delete
    7. The FICPI Memorandum can be found here, by the way.

      Delete
    8. Ok, so the point is that a (hypothetical) claim to "metal but not copper" could derive priority from a document only disclosing "metal".

      It is unfortunate that this contradicts G 2/98, headnote.

      Delete
    9. Since when did anyone get a proper understanding of a decision by reading the headnote?

      Delete
    10. Anonymous, your conclusion ignores the fact that a disclaimer (here, of copper from the generic metal) doesn't affect the priority situation.

      Delete
    11. It also ignores the fact that some of the examples in the Memorandum would not pass the test according to that interpretation of G 2/98.

      Delete
  3. Rather than two hypothetical claims (one to 'copper' and one to 'metal but not copper'), I would view it as a single claim which has been notionally split into two portions.

    And I would not ask whether G2/98 allows you to derive priority from the date of filing of the present European application as filed. G2/98 is about claiming priority from an earlier application under Articles 87-89. The date of filing of the present European application derives instead from Article 80 (meeting certain formal requirements).

    My question is then whether the notional portion 'metal but not copper' actually needs to benefit from any priority at all. Does it matter what date it has? Even if it only has the filing date of the later European application, 'metal but not copper' is still not anticipated under Article 54(3) by the disclosure of copper in the priority doc P1.

    You might say "what if P2 discloses iron as a further example?" Would P2 anticipate 'metal but not copper'? However, in that case you notionally split the claim three ways. 'Copper' has priority from P1. 'Iron' has priority from P2. 'Metal but not copper or iron' arguably has no priority. None of these notional portions of the claim is anticipated under Article 54(3).

    The reason I prefer to think of notionally split portions of a single claim, rather than two or more separate hypothetical claims, is because it makes it clearer that Article 123(2) is irrelevant. We have not actually made any physical amendment to the real claim, 'metal'. So there can be no question of an amendment which extends its subject matter.

    Similarly, if the claim to 'metal' is in a divisional application, then under Article 76(1) the only question is whether the real claim wording, 'metal', extends beyond the content of the parent European application. Assuming the parent application had a corresponding disclosure of 'metal' which supports the claim language, the divisional is deemed to have been filed on the date of the parent. The Article 54(3) and priority analyses are then as above.

    Since 'metal but not copper' is a purely notional construct, Article 76(1) does not require us to find explicit support for it in the parent application. The 76(1) prohibition is on extension of the subject matter of the actual real claim, 'metal', irrespective of some purely notional split that might be proposed subsequently.

    In other words, the issues surrounding partial priorities under G2/98 only affect priorities claimed from earlier applications. I don't think they affect Articles 123(2) and 76(1).

    The issue now is for the Enlarged Board to explain what it really meant in G2/98. I think the language "a limited number of clearly defined alternatives" does have some basis in one part of the FICPI Memorandum, but is misleading when taken on its own without reference to the whole Memorandum. And a series of Board of Appeal decisions have interpreted it without considering the Memorandum.

    ReplyDelete
    Replies
    1. Rather than two hypothetical claims (one to 'copper' and one to 'metal but not copper'), I would view it as a single claim which has been notionally split into two portions.

      That is what I meant by hypothetical claims, so this is only a question of terminology.

      Before I answer your question I'd like to get your view on this example:

      P1 discloses copper.
      P2 includes P1 and additionally discloses metal.
      Application is identical to P2 and claims metal.

      If we "notionally divide" the claim into copper and "metal but not copper", is the latter part entitled to priority from P2?

      If no, would the claim then surprisingly not fully have priority from P2 and not be protected against a document D1 disclosing iron published between the filing of P2 and the application? (Or is one allowed to flexibly assign priority dates to hypothetical/notional parts of the claim as the need arises? If the "notional division" approach holds water, then how does claim 1 remain novel over D1 upon notionally dividing it in "copper" OR "metal but not copper", the latter part apparently having the filing date of the application as effective date?)

      If yes, then how about the order of G 2/98?

      Delete
    2. My question is then whether the notional portion 'metal but not copper' actually needs to benefit from any priority at all. Does it matter what date it has?

      If "metal but not copper" is to be seen as one of a number of notional "OR"-parts of a claim and some published document exists that could be relevant for it regarding novelty or inventive step, then it better have an effective date so that we can decide whether the document is prior art for that notional part under either Art. 54(2) or 54(3).

      I'm sure we agree that an OR-claim is only new and inventive if all its "OR"-parts are new and inventive.

      If the notional split is made for the first time in a communication of the examiner or in a letter of the attorney, maybe even after grant, why should all its parts be at least entitled to the filing date of the application?

      Note that I am not arguing here that it must be possible to make the notional split within the limits of Art. 123(2). I tend to think that this is required for the split to be valid, but my argument does not depend on it. So this is not about what "provided that it gives rise to the claiming of a limited number of clearly defined alternative subject-matters" might mean.

      My point is that one should be able to assign a date to each of these alternative subject-matters and that here the criterion "directly and unambiguously derivable" certainly does apply. Why should an application only disclosing "metal" entitle you to the filing date of that application as effective date for a notional subject-matter "metal but not copper"? That makes little sense and would be contrary to core principles of the EPC as expressed in G 2/98 and G 2/10, r. 4.6.

      It might be convenient to apply the notional division approach only when it suits one and without caring too much about doctrinal correctness and internal consistency of the EPC, for example only to circumvent the poisonous divisional issue, but that is is not a proper solution.

      Delete
    3. If we "notionally divide" the claim into copper and "metal but not copper", is the latter part entitled to priority from P2?

      I've already said that it doesn't matter. But no, probably not, because P2 doesn't disclose "metal but not copper".

      If no, would the claim then surprisingly not fully have priority from P2 and not be protected against a document D1 disclosing iron published between the filing of P2 and the application?

      The application is identical to P2. The claim says "metal", and so does P2. Of course it has priority from P2, which protects it from D1. It would be surprising if it didn't.

      But this priority is for the whole claim, "metal", which is what P2 actually discloses. It's not for "iron", nor for "metal but not copper".

      This is just the normal situation which exists in any application which claims priority, and it exists from the moment the application is filed with a claim to "metal". I don't see why this situation for "metal" should change subsequently. Even if a situation arises where you additionally need to consider the effective dates of notional split portions for "copper" and "metal but not copper".

      (Or is one allowed to flexibly assign priority dates to hypothetical/notional parts of the claim as the need arises?

      Why not? Isn't that what partial priorities are for?

      Suppose P2 was never filed, and D1 disclosed copper rather than iron. You would have no difficulty assigning the claim a partial priority back to P1 for copper. This is what Article 88(3) says you should do.

      So you have flexibly assigned a priority date to "copper", which is a notional part of the claim to "metal". But the rest of the claim -- the notional "metal but not copper" -- then has no priority, but just the date of filing of the application.

      [However, since D1 is a publication which can be used for obviousness against "metal but not copper", the claim may still be invalid in this case. But for obviousness, not lack of novelty. Are other metals obvious from the disclosure of copper?]

      If the "notional division" approach holds water, then how does claim 1 remain novel over D1 upon notionally dividing it in "copper" OR "metal but not copper", the latter part apparently having the filing date of the application as effective date?)

      Because in your example you don't need to consider the possible partial priorities of "copper" or "metal but not copper". Rather, you have to consider the priority of the whole claim, "metal". This comes from P2, and protects you from D1 in the same way as if you never thought about the notional split.

      This is why I emphasised that the split is purely notional. The actual claim for "metal" has not been amended. It's still there. You are assigning effective dates to three different things: "copper" (P1); "metal" (P2); and "metal but not copper" (no priority, so the date of filing of the application).

      If yes, then how about the order of G 2/98?

      I can't see a problem. The claim is to "metal". P2 discloses "metal". They are directly and unambiguously the same invention.

      Delete
    4. The application is identical to P2. The claim says "metal", and so does P2. Of course it has priority from P2, which protects it from D1. It would be surprising if it didn't.

      According to T 1222/11, P2 would not be the first application for copper, so that the application cannot claim priority from P2 "as far as" metal covers copper. In T 1222/11, the application or patent did not survive as priority was not claimed from what is here P1, so that P1 destroyed novelty of the claimed metal.

      Why not? Isn't that what partial priorities are for?

      No, that is certainly not what partial priorities are for. Partial priorities only exist to ensure that an application is not refused on the ground that not all subject-matter in it is entitled to priority. In the UK, applications used to get refused for that reason (the idea being that the applicant was trying to deceive the patent office by claiming priority when it was not fully entitled to it).

      T 1222/11, in my reading, also does not propose notionally dividing in arbitrary ways as one sees fit, just to get around a poisonous priority or divisional or parent. It does propose notionally dividing, but it makes clear that that approach may cut two ways.

      It can't be that a claim is protected from one (poisonous) document by assigning priorities in one way and from another (non-poisonous) document by assigning priorities in another way.

      Delete
    5. Anonymous: "My point is that one should be able to assign a date to each of these alternative subject-matters and that here the criterion "directly and unambiguously derivable" certainly does apply. Why should an application only disclosing "metal" entitle you to the filing date of that application as effective date for a notional subject-matter "metal but not copper"?"

      "Directly and unambiguously derivable" certainly applies if you are determining whether something has priority under Article 87. But if that fails, then the actual date of filing of the application is the default fall-back position. For that, all that matters is whether you have met certain formal requirements, see Article 80.

      In other words, "directly and unambiguously derivable" is the criterion if you want to assign an effective priority date other than the date of filing.

      In your example, you're asking for the effective date of "metal but not copper". Is it directly and unambiguously derivable from P1? No - so it's not entitled to the date of P1. Is it directly and unambiguously derivable from P2? No - so it's not entitled to the date of P2. So what's left? Just the date of filing of the application, which doesn't have any requirement for direct and unambiguous derivation.

      I still think that your problem is that although you are using the words "notional" and "hypothetical", you are not really thinking of the split as notional. Instead you are treating it as if it were an actual physical amendment, made some time after the filing of the application, e.g. when answering an official action, or countering an objection in an opposition.

      If it was actually a real physical amendment made at some subsequent date, then of course it would have to be directly and unambiguously derivable. But it's not. It's purely notional.

      Thus, on one hand you say that your argument does not depend on whether it is possible to make the notional split within the limits of Article 123(2).

      But on the other hand, you nevertheless apply the "directly and unambiguously derivable" criterion. That's the key criterion of Article 123(2).

      Delete
    6. Anonymous: "According to T 1222/11, P2 would not be the first application for copper, so that the application cannot claim priority from P2 "as far as" metal covers copper. In T 1222/11, the application or patent did not survive as priority was not claimed from what is here P1, so that P1 destroyed novelty of the claimed metal."

      The flip side is that a notional split of "metal" into "copper" and "metal but not copper" would have worked just fine, if the applicant had claimed priority from P1.

      As you say, it can cut both ways.

      "Partial priorities only exist to ensure that an application is not refused on the ground that not all subject-matter in it is entitled to priority. In the UK, applications used to get refused for that reason (the idea being that the applicant was trying to deceive the patent office by claiming priority when it was not fully entitled to it)."

      Partial priorities weren't introduced just to prevent refusal of applications in their entirety, in just one out of all the member states of the Paris Union.

      It was much less parochial -- to prevent the denial of priority if elements were added which went beyond the priority application. So if the priority document says "copper" but the claim says "metal", you can assign priority to "copper" but the rest of the claim would just have the date of filing. This is also the position of the FICPI memo.

      See Bodenhausen's Guide to the Application of the Paris Convention, paragraph (d) on page 54.

      Delete
    7. The flip side is that a notional split of "metal" into "copper" and "metal but not copper" would have worked just fine, if the applicant had claimed priority from P1.

      Not really. P2 would remain unusable for giving priority to "metal" (due to P1, see T 1222/11) and as you seem to agree the application cannot get priority from P2 for "metal but not copper" (here T 1222/11 seems to disagree, but the conclusion of G 2/98 is clear about it). So a document disclosing iron published between the filing date of P2 and the filing date of the application is novelty destroying.

      So see again my comment of 12 January 2015 at 19:51, your answer of 13 January 2015 at 01:45 and my reply of 13 January 2015 at 08:41.

      For convenience:

      P1 discloses copper.
      P2 includes P1 and additionally discloses metal.
      Application is identical to P2 and claims metal.

      D1 published between the filing date of P2 and the application discloses iron.

      T 1222/11 says: P1 is first application for copper, so the claim cannot claim priority from P2 as far as it covers copper.

      G 2/98, headnote/conclusion says: the claim's notional part "metal but not copper" cannot claim priority from P2, as P2 does not disclose "metal but not copper".

      According to T 1222/11, the claim to metal can claim priority from P1 for its notional part copper. But the notional part "metal but not copper" cannot claim priority from P2, so is not new. And T 1222/11 makes clear that the full claim "metal" cannot claim priority from P2 either, as this priority is not valid due to P1 being the first application for "copper".

      So, surprisingly, D1 is novelty destroying for metal.

      So if the priority document says "copper" but the claim says "metal", you can assign priority to "copper" but the rest of the claim would just have the date of filing. This is also the position of the FICPI memo.

      Not exactly. The FICPI memo seems to assume that our claim to metal can get priority from P2 for "metal but not copper". But as we have seen this contradicts G 2/98, headnote/conclusion. It is T 1222/11 that then makes the step to the case with a single priority document P1 disclosing copper and "metal but not copper" getting the filing date. Were the FICPI memo approach correct, then I agree it should also apply to the case with a single priority document, but it is not.

      Delete
  4. Addendum: On 10 January at 17:03, Anonymous suggests that "Art. 76(1) works quite similarly to Art. 87(1): something can only get the date of the parent application if it is directly and unambiguously derivable from it. That is not the case for (B)." (I.e. not the case for the notional split portion "metal but not copper".)

    I disagree that they work similarly. Article 88 EPC allows for multiple and partial priority dates, which is what we are considering when we notionally split the claim. But Article 76(1) is all or nothing. Either the divisional application as a whole extends the content of the parent, or it doesn't.

    So when determining the effective filing date of a divisional application, the 76(1) issue is simply whether the subject matter as a whole is directly and unambiguously derivable from the parent. Unlike a priority determination, there is no need to consider whether some parts have one date and some another.

    Article 76(1) works similarly to Article 123(2), rather than Articles 87 and 88.

    If you split a claim notionally, but do not actually make a physical amendment, you have not extended its subject matter. Under Article 123(2), there would only be an objection if you made an actual amendment which disclosed or claimed 'metal but not copper' explicitly.

    Likewise, if you split the claim notionally in a divisional application, you have not thereby extended its subject matter with respect to the parent application under Article 76(1). The divisional would only lose the effective filing date of the parent application if it actually disclosed or claimed 'metal but not copper' explicitly.

    ReplyDelete
    Replies
    1. You are right that Art. 76(1) is more similar to Art. 123(2) than to Art. 87(1). However, there is still the idea of getting an earlier effective date for subject-matter physically filed later, and this is contingent on that subject-matter being directly and unambiguously derivable from a document filed on that earlier date.

      Suppose G 1/05 had a somewhat different outcome: subject-matter in a divisional that is present in the parent gets the effective date of the parent, subject-matter in a divisional that is not present in the parent gets as effective date the physical filing date of the divisional. Now Art. 76(1) works in the same way as Art. 87(1), and I don't think this should affect whether the problem of poisonous divisionals exists or not. You're right that Art. 76(1) as interpreted by G 1/05 does not work as Art. 87(1) does, but this is only because Art. 76(1) is more restrictive than Art. 87(1) (as if an application claiming priority must be refused if priority turns out to be invalid, which was actually still the case in the UK in 1967 and the reason that the Paris Convention was amended to allow partial priority; see the annex to the board's communication).

      OK, one could argue that the order of G 2/98 is still limited to Art. 87(1) and does not apply to Art. 76(1). One could argue that, for the purpose of Art. 76(1), when "notionally dividing" a claim of the divisonal, one may also "notionally divide" a claim of the parent. The notional part "metal but not copper" in the divisional then corresponds to the notional part "metal but not copper" in the parent, so all seems fine.

      However, one is then also able to include in the divisional (when filing it) a claim to "metal but not copper", since we have seen that Art. 76(1) allows notionally dividing claims of the parent...

      Delete
  5. "One could argue that, for the purpose of Art. 76(1), when "notionally dividing" a claim of the divisonal, one may also "notionally divide" a claim of the parent. The notional part "metal but not copper" in the divisional then corresponds to the notional part "metal but not copper" in the parent, so all seems fine."

    No, that's not my argument.

    As you acknowledge, Article 76(1) is different from a priority claim under Article 87. There's no question of splitting the claim into notional portions. Article 76 does not have anything corresponding to multiple and partial priorities under Article 88.

    So for the purpose of Article 76(1), you can ignore any notional split that you might or might not have to consider for priority purposes. You just look at the actual, real subject matter of the divisional, and ask if it is directly and unambiguously derivable from the actual, real subject matter of the parent.

    If your example claim to "metal" was filed in a divisional, you would just ask whether "metal" was directly and unambiguously derivable from the parent. If yes, then the divisional is deemed to have been filed on the date of filing of the parent. End of.

    You say that by "hypothetical claims" you mean the same as my single claim which has been notionally split into two portions for priority purposes. But I don't think you do mean the same as me. Your argument supposes that when you split "metal" into "copper" and "metal but not copper", then some new (but notional) subject matter is magically created.

    This then means that under Article 76(1) you need to look for the notional "metal but not copper" subject matter in the parent application. (Even though the divisional status would be fine if no priority question arose.)

    But that's not a notional split of a single claim, which is considered purely for the purposes of determining a partial priority of that single claim under Article 88. You're notionally creating an additional (but hypothetical) claim for "metal but not copper". And then subjecting it to Article 76(1) as well.

    ReplyDelete
  6. It seems fairly simple to me: the only part of the claim that can be anticipated by the priority document is the part of the claim that is entitled to priority. The rest of the claim must be novel. I don't see how clearly defined alternatives come into it. This is a simple logical proposition and is much preferable to some abstruse effort to make a single line in a G decision into a binding rule that appears to contradict the FICPI Memorandum (see the example about pipes and hollow bodies). It does not add matter to an application as filed to be able to see that this proposition is true. Just because we have to spell it out clearly for people who don't get the point using a "notional split" of the claim does not mean we have somehow added matter.

    ReplyDelete
    Replies
    1. the only part of the claim that can be anticipated by the priority document is the part of the claim that is entitled to priority. The rest of the claim must be novel.

      To be honest I'm not entirely sure whether this is a proposal or whether you're arguing that this is the law.

      It is certainly not how priority is commonly understood to work under the EPC (and see e.g. G 2/98), or the problem of poisonous divisionals would not have arisen in the first place.

      Basically G 2/98 is clear about a claim either having priority from a document in which the subject-matter of the claim is disclosed ("same invention"), or not having priority at all. The only exception is the case where the claim defines a limited number of clearly defined alternative subject-matters (the "OR"-claim).

      Delete
    2. I have been trying to stay out of this discussion, because everything has been said before in previous posts. The point you raise, however, keeps coming up again and again, which I find puzzling. My question to you is: regardless of the "limited number of clearly defined alternatives" issue, which particular examples in the FICPI Memorandum do you think are not approved by G 2/98? I would like a clear answer on this, including why you think so, otherwise there is really no point continuing the discussion.

      Delete
    3. Anonymous, I think your argument about how priority is understood is bootstrapped onto what G 2/98 says about it.

      But the rest of us are saying that G 2/98 is what needs to be changed (or at least clarified) in the light of the FICPI memo. In particular the bit about "OR"-claims and a limited number of clearly defined alternative subject-matters.

      Delete
    4. which particular examples in the FICPI Memorandum do you think are not approved by G 2/98?

      It depends on how your read the examples and how you apply the "directly and unambiguously derivable" test. The main point is illustrated by the example that has been mentioned (and is not literally one of the memorandum examples):

      P1 discloses copper.
      P2 discloses metal (and maybe copper as well, but nothing else).
      The application claims metal.

      In my view, the answer given in G 2/98 to the referred question does not allow "copper but not metal" to have priority from P2. (Note that until G 2/10 this wasn't so clear. G 2/10 conceivably could have allowed one to claim "metal but not copper" on the basis of disclosures of the generic "metal" and the specific "copper". But it did not.)

      So let us now have a look at the examples from the memorandum.

      Example (a) is hardly problematic, especially if one assumes that the application also discloses each of chlorine, bromine, iodine and fluorine individually (as P2 clearly does). Clearly one may draft a claim to chlorine and another claim to bromine or iodine or fluorine and there seems little reason to deny a claim to "halogen" covering precisely those examples (in particular in view of point 6.7 which says using a generic term should be OK). What it gains you is that a single claim suffices instead of having to draft 4 claims. The latter would be equally possible, but the memo thinks it runs the chance of being objected to for lack of conciseness (that would be unfair though).

      Regaring example (b), I'd like to know what P2 discloses exactly. Chances are it discloses not only 10-25, but also 15-20. Does that allow one to claim the two ranges (10,15) and (20,25) (within the limits of 123(2))? To be honest, I don't really know the case law on ranges well enough to answer with certainty. The drafter of the memorandum, and even the EBA in G 2/98 if the EBA actually cared about the precise details of these examples which I don't think it did, may well have thought this possible. But assuming that P2 only discloses 10-25 and 15-20 and current case law on ranges does not allow to derive the ranges 10-15 and 20-25 from this, example (b) is contrary to G 2/98.

      In example (c), P1 discloses "the inner wall of a pipe" and P2 discloses "the inner wall of bottles or any other hollow bodies". It is not clear whether P2 also disclose "the inner wall of a pipe".

      For example (c) to work, the OR-part "the inner wall of hollow bodies not being pipes" must have priority from P2. According to the EBA's answer to the referred question in G 2/98, this is only possible if "the inner wall of hollow bodies not being pipes" is directly and unambiguously derivable from P2. I am pretty confident that according to current case law on 123(2) this is not the case (not even if P2 does disclose "the inner wall of a pipe", but for this one needs G 2/10).

      So example (c) is not valid under G 2/98 and current case law on Art. 123(2) (or rather, current case law on the "directly and unambiguously derivable" test).

      Note that in so far as examples (a) and (b) are in line with G 2/98, they do not lend support to the notion of "conceptually identifying" mentioned in point 11.5.8 of T 1222/11. In so far as these examples comply (and assuming that the application includes all the material from P1 and P2), the application really discloses the alternative subject-matters and there is no need to "conceptually" identify them. One could actually draft such claims.

      Delete
    5. Anonymous, I think your argument about how priority is understood is bootstrapped onto what G 2/98 says about it.

      It certainly is. Before G 2/98 there were divergent views on when a claim is entitled to priority. G 2/98 brought a lot of clarification.

      But the rest of us are saying that G 2/98 is what needs to be changed (or at least clarified) in the light of the FICPI memo. In particular the bit about "OR"-claims and a limited number of clearly defined alternative subject-matters.

      I understood at least T 1222/11 and Tufty to argue that G 2/98, and in particular that bit about "OR"-claims", supported and even suggested the "notionally dividing" theory.

      Delete
    6. @Anon:
      I think your answer tells us everything we need to know about your argument. Even though G 2/98 indicates that the Memorandum expresses the legislative intent behind Article 88(2) second sentence (see para 6.4), in your opinion the examples provided in the Memorandum would not all comply within Article 88(2). In my opinion, this view makes no logical sense. We will therefore just have to agree to disagree and wait until the Enlarged Board decides the issue.

      Delete
    7. Oops - typo. Replace 'within' with 'with'.

      Delete
    8. Anonymous: "I understood at least T 1222/11 and Tufty to argue that G 2/98, and in particular that bit about "OR"-claims", supported and even suggested the "notionally dividing" theory."

      G 2/98 doesn't do that by itself. You have to read it in the light of the FICPI memo. G 2/98 said that the memo represents the legislative intent, but then it added the bit about "a limited number of clearly defined alternative subject matters". And that has led a series of Boards who didn't read the memo down a path that is incompatible with it.

      Delete
    9. Anon - why bring A123(2) into it? That is a total red herring. The sentence at issue in G 2/98:

      "The use of a generic term or formula in a claim for which multiple priorities are claimed in accordance with Article 88(2) EPC, second sentence, is perfectly acceptable under Articles 87(1) and 88(3) EPC, provided that it gives rise to the claiming of a limited number of clearly defined alternative subject-matters."

      In example (c), the Memorandum helpfully sets out the two alternative subject matters that allowing partial priority for the generic claim to C (method of coating the inner wall of hollow bodies) gives rise to. These are A "method of coating inner wall of pipes"; and B "method of coating hollow bodies not being pipes". There are two alternatives, which fits "limited number". They are also clearly defined, which meets the second part of the test. Thus I would argue that example (c) of the Memorandum is within G 2/98.

      I do not think A123(2) comes into it, because there is no amendment. There is no suggestion in G 2/98 that the clearly defined alternatives that are the result of the partial priority must have basis as an amendment under A123(2).

      Delete
    10. Even though G 2/98 indicates that the Memorandum expresses the legislative intent behind Article 88(2) second sentence (see para 6.4), in your opinion the examples provided in the Memorandum would not all comply within Article 88(2). In my opinion, this view makes no logical sense.

      You asked me what examples in my view did not comply with G 2/98. I have answered and explained why.

      If what you mean is that example (c), even in view of G 2/10 (in particular point 4.5.3) complies with G 2/98's "priority of a previous application in respect of a claim in a European patent application ... is to be acknowledged only if the skilled person can derive the subject-matter of the claim directly and unambiguously, using common general knowledge, from the previous application as a whole" (replace "claim" by "notional part"), then we indeed have to agree to disagree.

      In G 2/98, the EBA felt it had to explain what was meant by "where appropriate, multiple priorities may be claimed for any one claim". This led it to page 104 of this document and to the memorandum. The memorandum indeed expresses the legislative intent underlying Art. 88(2), second sentence, EPC: OR-claims can have multiple priorities. That the examples in this memorandum are now somewhat outdated in view of how the case law regarding the disclosure test and the concept of "same invention" has developed does not contradict the fact that the memorandum still expresses the legislative intent.

      Delete
    11. In example (c), the Memorandum helpfully sets out the two alternative subject matters that allowing partial priority for the generic claim to C (method of coating the inner wall of hollow bodies) gives rise to. These are A "method of coating inner wall of pipes"; and B "method of coating hollow bodies not being pipes". There are two alternatives, which fits "limited number". They are also clearly defined, which meets the second part of the test. Thus I would argue that example (c) of the Memorandum is within G 2/98.

      Although I do not agree that the claim of example (c) defines two "clearly defined alternative subject-matters" in the form of A and B, this is not my argument.

      My argument is that B cannot have priority from P2.

      This is because for B to have priority from P2, P2 must disclose B, i.e. B must be directly and unambiguously derivable from P2. This is the outcome of G 2/98 (see the text below "Conclusion").

      B is directly and unambiguously derivable from P2 precisely when one could claim B on the basis of P2, i.e. within the limits of 123(2). (I am simply writing "123(2)" as shorthand for the well-known disclosure test.) The answer to this is "no".

      Since P2 does not disclose B, B cannot claim priority from P2. But example (c) of the memorandum assumes it can.

      Delete
    12. Anonymous said "That the examples in this memorandum are now somewhat outdated in view of how the case law regarding the disclosure test and the concept of "same invention" has developed does not contradict the fact that the memorandum still expresses the legislative intent."

      I would write that as "the way the case law has developed does not agree with the legislative intent as expressed by the examples of the memorandum."

      And the way the case law has developed produces absurd results. Section 88 was intended to conform with Article 4F of the Paris Convention (multiple and partial priorities). Did the drafters of Article 4F in 1958 really intend that a priority document could be used to invalidate a patent, rather than protecting it from intervening disclosures?

      That's why the Enlarged Board needs to revisit the issue, and set the case law back on track.

      Delete
    13. Did the drafters of Article 4F in 1958 really intend that a priority document could be used to invalidate a patent, rather than protecting it from intervening disclosures?

      Certainly not, but I think this is more a problem of Art. 54(3) than of Art. 4F.

      I suspect that the drafter of the memorandum silently assumed that disclaimers were generally allowable. I think in those days this was the case at least in certain jurisdictions. Assuming disclaimers are allowable, then if "metal" has priority from P2, also "metal but not copper" has priority from P2.

      According to G 1/03, disclaimers are in general not allowable (or at least they have to satisfy Art. 123(2) just like any other amendment). The limited exceptions include the introduction of a disclaimer to establish novelty over an Art. 54(3) document. This is exactly what is needed to neutralise poisonous divisionals. Example:

      P1 discloses copper.
      Application discloses copper and metal, claims metal.
      Divisional discloses copper and metal.

      Now "copper" in the divisional is in principle novelty destroying for the claimed "metal". But this claim can be split into two claims: one to "copper" and one to "metal but not copper" (using a G 1/03-disclaimer). The claim to "copper" is undisputably new. The claim to "metal but not copper" is also new, as G 1/03, point 4, confirms that it has priority from P2. Under these circumstances it makes no sense to insist on the claim being rewritten as two claims, so the claim to "metal" survives.

      Unfortunately, in G 2/10 the EBA found it necessary to limit G 1/03 to disclaimers of undisclosed subject-matter. See points 3 and 4.7. (Of course point 4.7 puts in doubt whether anything of G 1/03 is still valid.)

      In G 1/03, the EBA argued that Art. 54(3) was not intended to prevent an applicant from getting a patent on subject-matter "not covered by the disclosure" of the earlier application and that the legislator must have assumed the then-existing practice of delimitation against earlier applications to remain valid. The EBA therefore allowed a disclaimer in order to split the invention into the part present in the earlier application and the part disclosed for the first time in the later application.

      In case of a parent and a divisional disclosing a species with priority and a genus without, it does not seem to have been the intention of Art. 54(3) to prevent the applicant from getting a patent on the genus. Allowing (hypothetical) delimitation against the "earlier" 54(3)-species would neutralise the latter's poisonous effect and, as far as I can tell, not create the inconsistencies and incoherence of the "notionally dividing" approach.

      Btw, it seems to me that with such a G 1/03-based approach all examples from the memorandum again work just fine. So it is not the case that accepting the validity of those examples (which in my view is not done in G 2/98, but opinions clearly differ) would inevitably lead to the approach of T 1222/11.

      Should the EBA choose the G 1/03-based solution, it should emphasise that points 2.2 (disclaim not more than necessary) and 2.4 (clarity and conciseness) of the order are not relevant since the claim does not actually need to be rewritten, and that point 2.3 (may not become relevant for inventive step and sufficiency of disclosure) can be ignored as well on the understanding that for the purpose of Art. 54(2), 56 and 83 the claim is to be treated as unsplittable (unless it really is an OR-claim, obviously). This would also be a good opportunity for the EBA to kill point 4.7 of G 2/10.

      Delete
    14. I noticed two mistakes in my previous comment of 16:33.

      Now "copper" in the divisional is in principle novelty destroying for the claimed "metal". But this claim can be split into two claims: one to "copper" and one to "metal but not copper" (using a G 1/03-disclaimer). The claim to "copper" is undisputably new. The claim to "metal but not copper" is also new, as G 1/03, point 4, confirms that it has priority from P2. Under these circumstances it makes no sense to insist on the claim being rewritten as two claims, so the claim to "metal" survives.

      The phrase "as G 1/03, point 4, confirms that it has priority from P2" should be ignored, as there is no P2 in the example I am discussing.

      Btw, it seems to me that with such a G 1/03-based approach all examples from the memorandum again work just fine.

      This is wrong. For our purposes, the G 1/03-based approach only allows to disclaim what would otherwise be 54(3)-prior art. So in example (c), a document disclosing the method for pipes and published between the filing of P1 and P2 would be novelty destroying 54(2)-prior art, whereas the memorandum assumes it is only 54(2)-prior art for the method applied to "hollow bodies not being pipes" which would mean the claim is new but very likely not inventive.

      Delete
    15. I wonder why these posts always end in a complicated discussion of G 2/98 and effective dates of priority. Even if the Enlarged Board of Appeal would manage to find a solution to the priority issue along the lines of T 1222/11, then we would just face new questions of uncertainly in the future with yet further questions of establishing the fine line of priority basis. The narrow approach in this respect practiced by the EPO in terms of what is directly and unambiguously disclosed is obviously not only relevant to priority, but also to novelty in general as well as Art 123(2). If the Enlarged Board of Appeal will go along the lines of T 1222/11, then we will for sure face even more critical issues with respect to legal certainly in the future since the same principles would come into play for novelty purposes facing third party prior rights.

      Generally, I believe the problem is Art 54(3) and not one of priority. Also, I believe that by addressing the problem of Art 54(3) instead of priority, we would get easier around the real problem with own prior rights and we would avoid interference with third party prior rights (which I guess we all believe works fine in Art 54(3)). We could completely avoid the sensitive priority discussion if Art 54(3) would not have effect on own prior rights. Then we would solve the toxic divisional problem and problems with published priority applications being novelty destroying to later filed applications, and still maintain a system where third party prior rights may destroy novelty based on the existing narrow priority approach (which seems reasonable).

      Delete
  7. Dave, I agree. The problem is that the Enlarged Board said what it said in point 6.7 of G 2/98, about a "limited number of clearly defined alternative subject-matters". And a series of subsequent decisions has followed it literally. Only the Enlarged Board can undo the resulting mess.

    ReplyDelete
  8. Replies
    1. It is indeed. I've just spent the last couple of hours reading it and writing about it.

      Delete