Partial priority is one of those areas of European patent law that seems straightforward once the Enlarged Board has explained it, right up until somebody tries to apply it to an inventive-step attack.
The basic principle from G 1/15 is now familiar. A claim containing generic “OR” subject-matter can enjoy partial priority for an alternative disclosed in the priority application, without that alternative having to be expressly carved out in the wording of the claim. The claim can instead be conceptually divided into the part entitled to priority and the remaining part that is not.
This is useful for dealing with the formerly troublesome problem of “poisonous priority”. What is rather less clear is what happens when a priority-year publication discloses the priority-entitled embodiment, but an opponent wants to use that embodiment as the starting point for an inventive-step attack against a different part of the claim.
Two recent decisions, T 2615/22 and T 195/25, approach this question from slightly different directions. The second decision also has some fairly pointed things to say about an earlier decision, T 307/22, which it describes as “isolated”. This is not normally a sign that everything is entirely settled.
A brief reminder about partial priority
According to G 1/15, entitlement to partial priority cannot be refused merely because the relevant alternatives are expressed generically in the claim. The exercise involves identifying the subject-matter disclosed directly, or at least implicitly, unambiguously and sufficiently in the priority document, and then treating that subject-matter as a notional alternative within the claim.
The EPO Guidelines on multiple and partial priorities describe this by reference to different alternatives within the same claim having different effective dates. They also make clear that separate priority disclosures cannot simply be mosaicked together to create a combination that was not disclosed in either.
So far, so reasonably straightforward. The difficulty comes when the same physical product or method appears in a publication during the priority interval.
Is that publication prior art against the whole claim because some of the claim has the later filing date? Or is its disclosure protected because the thing actually shown in the publication is within the priority-entitled part?
The answer, perhaps unsurprisingly, appears to be that it depends on exactly what is being attacked and precisely what the intervening document discloses.
Water balloons in T 2615/22
T 2615/22 concerned the familiar type of device in which several water balloons can be filled at once. The claimed apparatus had a housing connected to a fluid supply, multiple hollow tubes and inflatable containers attached to the tubes. Elastic rings both held the containers in place and sealed them when they were released.
The case involved three priority applications. For present purposes, the important priority disclosure was of a particular construction having:
- a threaded opening at a first end of the housing;
- a second end of the housing; and
- hollow tubes attached to or inserted into holes at that second end.
The claims before the Board were broader. Among other things, they covered non-threaded openings and arrangements in which the tubes did not have to be attached to holes at a second end.
The Board concluded that the priority applications directly and unambiguously disclosed the claimed functional arrangement of the elastic rings. However, priority was available only for the more particular housing construction actually disclosed in the priority documents.
The claim was therefore notionally divided into at least two parts. One part covered the threaded, second-end, attached-tube arrangement and enjoyed priority. The rest of the claim, including non-threaded and differently constructed alternatives, had only the European filing date.
The intervening Bunch-o-Balloons disclosures
Documents D7 to D14 related to the commercial Bunch-o-Balloons product and had been made available after commendably precise. D9 and D12 showed the same particular construction as the priority applications: a threaded opening and tubes attached to holes at a second end. That disclosed subject-matter fell within the priority-entitled part of the claims. For that part, D7 to D14 did not form part of the prior art.
For the remaining parts of the claims, the publications fell within the relevant prior-art period. They did not, however, disclose those broader alternatives. In particular, they did not disclose a housing with a non-threaded opening or one lacking the particular second-end arrangement.
The intervening publications therefore did not anticipate either part:
- for the part they disclosed, the claim had priority; and
- for the parts lacking priority, the publications did not contain the necessary disclosure.
This is partial priority doing precisely the job it is supposed to do. A later publication of the same embodiment does not become novelty-destroying merely because that embodiment sits inside a broader generic claim.
Inventive step was another matter
The opponent also wanted to pursue inventive-step attacks based on the Bunch-o-Balloons disclosures.
The Board did not decide those attacks. It noted that the question whether the non-priority parts of the claims involved an inventive step over D7 to D14, together with common general knowledge or D17, had not been addressed in the opposition proceedings.
The case was remitted because novelty and inventive step required different analyses. In particular, inventive step required application of the problem-solution approach and consideration of the distinguishing features.
The decision consequently left the door at least partly open. It treated D7 to D14 as belonging to the prior art for the non-priority parts of the claims, although it did not decide exactly how disclosures of the priority-entitled construction could properly be used in an inventive-step argument directed to those other parts.
That unanswered question was rather more squarely confronted in T 195/25.
Steel slag in T 195/25
T 195/25 concerned a method of making carbonate-bonded construction blocks from steel-making residues.
Claim 1 required granular material and a binder including steel slag to be mixed with water, compacted at a pressure below 20 MPa, subjected to a reduction in water content and then cured using carbon dioxide.
Specific examples in the patent used:
- a compaction pressure of 12 MPa;
- a first water-to-slag ratio of 0.20; and
- a second water-to-slag ratio of 0.10.
The same examples appeared in the priority application D11 and in a later PhD thesis, D12. It was common ground that these examples were entitled to priority from D11. The broader subject-matter covered by claim 1 did not necessarily enjoy the same priority.
The opponent’s argument was ingenious, although ultimately unsuccessful. It proposed taking the 12 MPa example from D12 and modifying it in view of D5 or D16. By changing the compaction pressure, the skilled person would supposedly arrive at subject-matter within claim 1 but outside the specific embodiment entitled to priority.
In effect, the opponent sought to start inside the priority island and then sail out into the later-dated waters surrounding it.
The Board was having none of this.
If it isn’t prior art, you cannot start from it
The Board held that where a claim enjoys partial priority for a specific embodiment, the same embodiment disclosed in a non-patent document published during the priority year cannot be used as the starting point for an inventive-step attack against the part of the claim that does not enjoy priority.
D12 had been published after D11. The relevant examples in D12 were identical to those in D11 and in the patent. The subject-matter of those examples was therefore protected by the priority claim and did not form part of the state of the art under Article 54(2) EPC.
Article 56 EPC provides that inventive step is to be assessed having regard to the state of the art. The Board concluded that only subject-matter forming part of the Article 54(2) state of the art could be used in that assessment.
The D12 examples could consequently not be relied on “as a starting point or otherwise” when assessing inventive step.
This is the crucial part of the decision. The fact that the proposed endpoint of the obviousness analysis lay outside the priority-entitled embodiment did not somehow turn the starting embodiment into prior art.
Put another way, partial priority applies to subject-matter, not merely to the final wording against which novelty is assessed. If a particular technical teaching has the priority date, that teaching cannot be stripped of its effective date and used as an Article 54(2) springboard merely because the opponent proposes changing one of its features.
What about T 307/22?
The opponent relied on T 307/22, in which an embodiment appearing both in a priority document and in a priority-year publication had been permitted as the starting point for assessing inventive step.
The Board in T 195/25 declined to follow that approach. It instead regarded its conclusion as consistent with T 2073/22, T 1482/21 and T 211/06.
The Board went further when dealing with the requested referral to the Enlarged Board. It described T 307/22 as appearing to be an “isolated decision” supporting the opponent’s view. The alleged support from T 665/00 was dismissed as speculative because the disputed prior use in that case had not ultimately been employed in an inventive-step attack.
Despite this apparent divergence, no question was referred. The Board maintained that Article 56 provided the answer: if the relevant subject-matter was not part of the Article 54(2) state of the art, it could not be used for inventive step.
This is certainly a clean answer. Whether all Boards will consider it quite so straightforward remains to be seen.
Are the two decisions inconsistent?
At first sight, there is a possible tension.
In T 2615/22, the intervening Bunch-o-Balloons disclosures were treated as prior art for the parts of the claim not entitled to priority. An inventive-step attack based on those disclosures was remitted for consideration.
In T 195/25, the intervening thesis could not be used as a starting point against the part of the claim lacking priority because the embodiment it actually disclosed was itself entitled to priority.
The decisions can, however, be reconciled if attention is paid to two separate questions.
The first question is:
Which part of the claim is being attacked, and what is its effective date?
That was the main focus in T 2615/22. The Board divided the claim into priority and non-priority parts and tested novelty separately against each.
The second question is:
What precise technical teaching in the intervening document is being relied on, and is that teaching itself part of the Article 54(2) state of the art?
That was the decisive point in T 195/25.
An intervening document might contain additional disclosure falling outside the priority-entitled subject-matter. That additional teaching may be prior art against the non-priority part of the claim. Nothing in T 195/25 appears to prevent such an attack.
What an opponent cannot do, according to T 195/25, is rely exclusively on the embodiment protected by priority and then use the inconvenience of inventive-step reasoning to transport that embodiment into the unprotected remainder.
This seems correct. Otherwise, the result would be rather odd. A document would be incapable of anticipating the priority-entitled embodiment but could nevertheless be used as the starting point for modifying that very embodiment. It would be prior art for inventive step while not being part of the Article 54(2) state of the art from which Article 56 says inventive step is to be assessed.
A possible difficulty with the remittal in T 2615/22
Although the two decisions are distinguishable, T 195/25 may affect what can now be done with the remitted inventive-step issue in T 2615/22.
The Bunch-o-Balloons materials apparently disclosed the particular arrangement entitled to priority. They did not disclose the broader alternatives for which priority was unavailable.
If an inventive-step attack begins with precisely that protected arrangement and then proposes modifying it into, for example, a non-threaded version, the reasoning in T 195/25 suggests that the attack should not get started. The disclosed starting arrangement is not Article 54(2) prior art.
There may still be a valid attack if one of the intervening documents contains some separate technical teaching outside the priority disclosure, or if an earlier document provides a proper Article 54(2) starting point. Merely observing that the target part of the claim has a later effective date should not be sufficient.
This may be the most useful general lesson from T 195/25. Partial priority requires a two-sided comparison. It is not enough to identify the effective date of the claimed endpoint. The effective status of the proposed starting disclosure must also be established.
Some practical consequences
For patentees, the decisions confirm the value of putting detailed embodiments and worked examples into the priority application. Even if the eventual claim is broader, the particular subject-matter originally disclosed may retain its priority date as a notional part of the claim.
That can have consequences beyond novelty. Following T 195/25, a later paper, thesis, poster or product disclosure reproducing the priority example cannot be used as a convenient starting point for an inventive-step argument simply because the final claim covers more than the example.
For opponents, an attack needs to be mapped rather more carefully. It will be necessary to identify:
- the part of the claim that lacks priority;
- the subject-matter actually disclosed by the proposed starting document;
- whether that disclosed subject-matter is itself outside the priority-entitled part; and
- why it provides a proper Article 54(2) starting point for the problem-solution approach.
It is not enough to begin with a protected embodiment and announce that the skilled person could have changed it.
For those drafting priority applications, the message is much the same as it usually is: disclose the important combinations properly the first time. Broad claims can come later. A clearly disclosed embodiment may enjoy a surprisingly useful afterlife when priority is examined subject-matter by subject-matter.
Conclusion
T 2615/22 shows the now orthodox application of partial priority to novelty. A generic claim is conceptually divided, and an intervening disclosure is tested separately against the priority-entitled and non-priority alternatives.
T 195/25 takes the analysis one step further. It holds that the priority-entitled embodiment is not merely protected against anticipation. It is also unavailable as the starting point for an inventive-step attack directed to the later-dated remainder of the claim.
The emerging principle appears to be:
A claim may have different effective dates across its scope, but a specific embodiment entitled to the earlier date does not become prior art merely because an opponent proposes modifying it into a part of the claim having the later date.
That seems both logically and legally sound. Whether the approach in T 307/22 remains merely an isolated curiosity, or eventually prompts a referral, will be worth watching. The Boards have occasionally been known to disagree about supposedly straightforward points of priority law. That is, after all, how we ended up needing G 1/15 in the first place.
