According to Article 54 EPC, an invention is new if it does not form part of the state of the art, which is everything that has been made available to the public before the filing date of the application. Under Article 54(3), the state of the art also includes the content of another European application having a filing date before that of the application but which was published on or after the filing date of the application. This second definition is however only used when considering novelty, and not for inventive step under Article 56.
If a European application validly claims priority to an earlier filed application, this has the effect that the filing date of the earlier application counts as the filing date of the application. This applies to both Article 54(2) and (3).
All this means that an earlier (first) European application could be used as prior art for a later filed (second) European application that claims priority to the first one, if that priority claim is invalid (and provided the first application is also published). This could happen, for example, if the first application disclosed and claimed only specific examples, while the second application claimed an invention more generally. Claims of broader scope might therefore not have a valid claim to priority. Consequently the earlier application, disclosing examples falling within the scope of the later broader claim, could provide a novelty-defeating prior disclosure.
Normally this would be correctable, as it is usual to make sure that all of the material from the first application is included in the second. The applicant could, for example, narrow the scope of the claims in the second application so that they do have a valid claim to priority. In some cases, however, this might not be possible, or desirable. To come up with a far-fetched hypothetical situation to illustrate this, let's say that a first EP application discloses an invention incorporating a range from 0.325 to 0.415 (the units don't matter), while a second EP application, claiming priority to the first application, discloses and claims only a range from 0.330 to 0.415. The claim to priority is invalid because the range in the second application is not present in the first one, meaning that the claim to priority is not for the same invention (Article 87(1)(b) EPC). The first application, having an earlier filing date and having been published after the filing date of the second, is therefore prior art under Article 54(3). Since the first application discloses a range that entirely encompasses the second, and the difference between the two ranges is very small, the claims in the second application are not novel (although this is possibly an arguable point).
Although far-fetched, this is the actual situation that gave rise to decision T 680/08, which issued in April 2010. The decision has already been commented on by others here and here. The issue has also arisen before in T 1443/05. In both cases, the solution to the problem was (possibly dubiously) to include a disclaimer.
Cases where the priority filing can cause problems for a later filed application may be fairly rare. However, the principle in T 680/08 might, according to some interpretations, apply also for divisional applications. According to Malcolm Lawrence and Marc Wilkinson of HLBBshaw, a divisional application could effectively 'poison' its own parent if it turns out that the parent is not entitled to priority. The reasoning (which, in the HLBBshaw article, is very long-winded) is based on the divisional application being citable under Article 54(3) against its parent as a result of it having an earlier filing date (due to the priority claim) than the parent application (which lacks the priority date). The ramifications for this would, naturally, be quite serious. But is this reasoning even plausible?
Cases where the priority filing can cause problems for a later filed application may be fairly rare. However, the principle in T 680/08 might, according to some interpretations, apply also for divisional applications. According to Malcolm Lawrence and Marc Wilkinson of HLBBshaw, a divisional application could effectively 'poison' its own parent if it turns out that the parent is not entitled to priority. The reasoning (which, in the HLBBshaw article, is very long-winded) is based on the divisional application being citable under Article 54(3) against its parent as a result of it having an earlier filing date (due to the priority claim) than the parent application (which lacks the priority date). The ramifications for this would, naturally, be quite serious. But is this reasoning even plausible?
The issue also came up during opposition proceedings relating to EP0846450, although was not key to the patent being eventually revoked. The question of whether a priority document could be used as Article 54(3) prior art was considered by an expert called by the patentee, Professor Dr.-Ing. Ulrich Vollrath, a professor of patent law at the Rheinian-Westfalian University of Technology (his contribution, which is well worth reading, can be found here). When asked whether a part of a European patent application enjoying priority of an earlier US application under Article 87 or 89 EPC could be novelty-destroying for claims of the same application not enjoying priority, his firm answer was no for the reason that there had to be two applications for there to be a conflict. However, in answer to a further question of whether the outcome would be any different if the claims not enjoying priority had been divided out of the European application enjoying priority, his answer was also no because, in the case of divisional applications, for the purposes of examination of priorities and the resulting effects the parent and divisional applications had to be considered as if all claims remained in the same application. There would therefore be no earlier application to be considered, as both would be effectively the same application. Both had to be considered to be interchangeable, having equal footing as a result of a procedural split between them. What was admissible in the former application could not become inadmissible just by filing a divisional.
I am much more persuaded by the arguments of Professor Vollrath than those of Messrs Lawrence and Wilkinson. However, I may be wrong. What are the chances that an EPO board of appeal, tasked with a suitable fact situation, would consider a divisional application to be Article 54(3) art for its own parent? I would think very slim, but not impossible. What do you think?
UPDATE 18 August 2015: The above post has been cited in a referral to the Enlarged Board of Appeal regarding the issue of partial priority and poisonous divisionals. For more information see my recent post here.
