Article 123 (2) EPC and corresponding national provisions prohibit an applicant or patentee from amending a patent application or patent such that its subject-matter extends beyond the content of the application as originally filed. The statute is the same throughout Europe, but the practice is not always so. An issue where the Case Law of the EPO and the national courts seems to increasingly diverge is the admissibility of intermediate generalisations.

What are intermediate generalisations? Imagine the following exemplary situation: An EP application has a broad claim 1 consisting of features A, B and C. The application further contains several working examples which are directed to various combinations of specific embodiments, e.g. A’, B’, C’ and D’ (example 1) and A’’, B’’, C’’ and D’’ (example 2). Prior art turns up anticipating a combination of A, B and C. Is it now possible to amend claim 1 to a combination of features A, B and C’’? Or is your only possible fall-back position the specific combination of A’’, B’’, C’’ and D’’?

1 The EPO Practice

Over time, the Boards of Appeal of the EPO have developed a pretty strict application of Article 123(2), on the basis that, under the European first-to-file system, an applicant should only be able to rely on the technical information which he has “clearly and unambiguously” disclosed on filing in defining his invention. In particular, the Boards have consistently asserted that an applicant must not be allowed to improve his position with regard to the information originally disclosed in the as-filed application when making prosecution amendments.

In decisions such as T 962/98, T 1496/11 and many others the Boards of Appeal have consistently held intermediate generalisations to be impermissible because they arguably result in the disclosure of a combination of features which corresponds neither to an originally-disclosed combination of features, such as in a specific disclosed embodiment, nor to an originally-disclosed generalisation from that embodiment, such as an original claim, but rather to a combination of features in between the disclosed generalisation and the disclosed embodiment. Such a combination would have some, but not all, of the features of the disclosed embodiment. According to decision T 201/83, intermediate generalisations are exceptionally permissible only in the absence of any clearly recognisable functional or structural relationship between the retained and omitted features of the specific embodiment, such that the retained features are not inextricably linked with some or all of the omitted features.

It is therefore very common in European proceedings, such as examination and opposition, for claims to be held invalid because they had been amended to define such an impermissible intermediate generalisation.

2 The UK Practice

In general, the British courts, both officially and unofficially, have approved of the broad aim of bringing national standards of matters of patent validity into harmony across Europe, and with the standards of the EPO. In cases such as Vector v Glatt and European Central Bank v DSS, the Court of Appeal has referred with approval to the principles governing added subject matter laid down by the Boards of Appeal, and with particular reference to the concept of intermediate generalisation developed at the EPO.

However, a recent decision of the English Court of Appeal has indicated that the British courts may well be more flexible on this point of law than the EPO.

The case, AP Racing Ltd v Alcon Components Ltd, was heard at first instance in the Patents County Court (recently renamed to the Intellectual Property Enterprise Court). The case concerned brake calipers used in motor sport, and dealt with, in addition to added subject matter discussed below, matters of infringement, insufficiency, and obviousness. The patent with which the case was principally concerned was GB 2 451 690, granted by the UK IPO.

At first instance, AP Racing sought to enforce a claim to a body for a fixed type disc brake caliper, the body of which inter alia had peripheral stiffening bands with a profile that was asymmetric about a lateral axis of the body when viewed in plan. According to the granted specification, the purpose of the peripheral stiffening bands, which extended around the periphery of the caliper body, was to resist the bending motion moment generated during braking. This allowed the caliper body to use less material overall, and thus to become lighter. Due to the use of the peripheral stiffening bands in the caliper body and the removal of material elsewhere, the body was given a distinctly asymmetrical appearance when viewed in plan. However, the application as originally filed did not specifically explain that the peripheral stiffening bands had a profile asymmetric about a lateral axis of the body when viewed in plan, as claimed. Rather, specific embodiments were described in which the body comprised a J-shaped or hockey-stick-shaped peripheral stiffening band which extended around the leading (or forward) end of the caliper body. Additionally or alternatively, it was explained that the body could comprise a similarly-shaped peripheral stiffening band which extended around the trailing (or rear) end of the caliper body.

The judge at first instance explained the principles of added matter and intermediate generalisation as applied by the UK Courts in previous decisions, with reference also to the case law of the EPO. The judge then held the claim to be invalid on the basis that the application as originally filed taught only that the body of the caliper had a distinctly asymmetrical appearance. While peripheral stiffening bands having a specific configuration, namely the J-shaped configuration, which was necessarily asymmetrical, were disclosed, the particular generalisation of stiffening bands having an asymmetrical appearance which was not necessarily the same as the specific asymmetrical appearance of the J-shaped embodiments was not disclosed.

The decision was appealed, on the question of added subject matter only, to the Court of Appeal. In the judgement on appeal, the Appeal Court reversed the decision of the first instance court, holding that the amendment did not constitute an impermissible intermediate generalisation. In the view of the Court, because the application as originally filed disclosed a particular class of embodiments having peripheral stiffening bands which were, in fact, asymmetric about a lateral axis, the recitation in the claim that the peripheral stiffening bands had a profile asymmetric about a lateral axis of the body when viewed in plan did not provide the skilled person with any further information which he would not have obtained from the application as originally filed. Particularly, the Court explained that the claims of a patent application were understood by the reader to perform a different function from the disclosure in the body of the specification, namely to delimit the area of the proprietor’s monopoly. The Court explained that the law did not prohibit the addition of claim features that state, in more general terms, that which is described in the specification, as long as such features do not disclose new information about the invention.

In the view of the Court of Appeal, the skilled person would understand from the application as filed that the design of the peripheral stiffening bands, in particular the fact that they wrapped around the corners of the caliper in a J-shape, was technically significant in imparting stiffness to the caliper, in terms of resisting the bending moment generated during braking. In the view of the Court, since the disclosed J-shaped peripheral stiffening bands were necessarily asymmetrical, the statement in the claim that such bands were asymmetrical would not provide new information to the skilled person. That the stiffening bands of the disclosed embodiments were asymmetric would be clear to the skilled person reading the application as filed, directly and unambiguously. All that the skilled person would take additionally from the specification as granted, as compared with the specification as filed, was that the claim had been drafted so that the scope of protection covered asymmetric peripheral stiffening bands generally, and not solely peripheral stiffening bands of the J-shaped form shown in the embodiments. Therefore, the intermediate generalisation had not taught the skilled person anything new, and therefore had added no new subject matter.

3 The German Practice

The German practice also applies a more flexible and applicant-friendly approach than the EPO. Intermediate generalisations of an originally disclosed working example are generally considered admissible, if one or several features of the working example that together and per se are responsible for the technical success of the invention are taken into the claim (X ZB 9/89 Spleißkammer; X ZR 117/11 Polymerschaum).

The Polymerschaum decision is particularly instructive for understanding the approach taken by the German Federal Court of Justice. The patent claim in this case related to a process for preparing a polymer foam, wherein expandable polymer microspheres and a molten polymer mixture are to be melt mixed under certain temperature conditions. The FCJ held that it is sufficient as a basis for this feature that it can be taken from the working examples that microspheres are to be added to a polymer melt. The court stated (margin no. 52, translation by the authors):

“According to the established case law of the Senate it is necessary for affirming the original disclosure of the subject matter of a patent claim that the skilled person was able to “directly and unambiguously” derive the technical teaching of the claim from the as filed documents as a possible embodiment of the invention (cit. omitted). In this case law the Senate has also approved generalizations of originally disclosed working examples to avoid undue restrictions of the applicant when using the disclosed content. He has in any event deemed a “broad” claim unobjectionable from the point of view of added matter, if a working example presented itself as an embodiment of the more general teaching as circumscribed in the claim and if this teaching in its claimed generality could already be taken from the application as filed as belonging to the invention as filed (cit. omitted), be it in the form of a claim formulated in the application as filed, or be it from the whole context of the application as filed. Such generalizations have been mainly admitted when only one or only a few of the features of a working example that taken together, but also in themselves promote the success of the invention, have been included in the claim (cit. omitted). The generalization of a chemical compound has, however, also been approved (FCJ, judgment of 18 December 1975, X ZR 51/72 , BGHZ 66, 17, 30 – Alkylendiamine I).”

4 Conclusion

The above discussed two judgements suggest that at least the English and German patent courts will be more prepared to uphold claims that the EPO might have considered to be invalid for contravening Article 123(2) EPC. From a practitioner perspective, this would suggest that, if protection in the United Kingdom and/or Germany were important, the filing and prosecution of a GB and/or DE national application to grant as an alternative to or in addition to an application before the EPO may be a useful strategy, since it may well be possible to obtain amended claims of broader scope than those achievable before the EPO.

Mark Jones (Hoffmann Eitle London) & Thorsten Bausch (Hoffmann Eitle Munich)


To make sure you do not miss out on regular updates from the Kluwer Patent Blog, please subscribe here.

Kluwer IP Law


  1. As the two patents discussed in UK and in Germany, were never subject of procedures before the EPO, it seems not entirely fair to say that national courts are more lenient on the matter than the EPO.
    One example in which both ways are possible could be a PCT application in which FR has been designated (becomes automatically a Euro-PCT), and Germany via the national route.
    In general, the question at stake here is the following: are A”,B” and C” inextricably linked or not? If the answer is positive, then the claim with A,B,C” offends Art 123(2), as A” and B” have been deleted.
    When it comes to intermediate generalisation, one could also think of applying the “essentiality” test, cf. H-V-3.1, provided it is correctly applied, in order to help in the matter. See also T 2311/10, when it comes to correctly applying the essentiality test.

    1. Nice to see D X Thomas commenting. I always enjoy his contributions. Perhaps he could comment on the “not even the slightest doubt” standard to be met, when it comes to the allowability under Art 123(2) EPC, of an amendment made after filing.
      I ask because this egregious test has no solid foundation in the DG3 caselaw. When one looks at the early cases, in which the first mention of “beyond reasonable doubt” occurred, they concern particular things like Art 123(3) and whether the content of an affidavit is to be given weight, and not the general issue whether to allow the amendment. Why should the burden be on the applicant, to prove beyond the slightest doubt that the amendment adds no matter to the application? It isn’t fair, logical or reasonable, is it, Daniel?
      The DG3 excuse for the aberrant standard is that the EPO has to be particularly vigilant to protect the public from the applicant who is bent on “improving his position” after filing. Doesn’t the public also need just as much protection from the applicant who improves his position during the Paris year. And isn’t the applicant entitled to the same degree of protection from an over-zealous DG1, asserting that there is no novelty over the disclosure of D1. Every prosecution amendment can be seen as an applicant trying to improve his position, legitimately, after receipt of an EPO search report that alerts him to art he hitherto did not know. Why impose “beyond reasonable doubt” for 123(2) but not for Art 54 or 87? I say, it is neither fair nor reasonable, and will cause world-wide annoyance and offence, squandering the good name of the EPO round the world. May i ask, what do you think about that, M. Thomas

      1. Sorry for the late reply, but I did not realised that a new comment had been filed.

        In my opinion, the discussion has no reason to be.

        I merely interpret the wording “beyond reasonable doubt” as a no more than a different way of expressing the concept of “directly and unambiguously derivable” which is nowadays commonplace in decisions of the Boards. The wording “beyond reasonable doubt” stems from a very old decision, T 388/88, which I concede has been quoted in some later decisions.

        When looking at German decisions I could not find any corresponding decision. I have not looked at French decisions, but I doubt my findings will be substantially different.

        On the other hand, it is quite clear that the Boards of Appeal have tightened their interpretation of what is directly and unambiguously derivable, but I fail to see a different standard for added subject-matter and novelty and priority.

        What is nevertheless clear is that the Boards of Appeal have not and will not adopt the views of the German Federal Court (BGH), the contrary is true, as I personally think that the BGH has rather adopted the view of the Boards. Equivalents belonged to novelty under old German practice, never under EPO practice.

        On the other hand, I also think that it does not needs much to have a broader range of possibilities when drafting the original specification. For instance, when in the description at the beginning of a § one finds the word “alternatively”, the applicant should not be surprised if he is barred from claiming later a combination of both alternatives, even such a possibility would be obvious for the person skilled in the art. Case law of the Boards is quite clear in this respect. On the other hand, a formalistic approach when assessing added subject-matter is not correct either, see T 2619/11.

  2. Afterthought:
    One supposes that the public needs just as much protection from the Applicant
    who improves his position during the Paris Convention year, by adding matter,
    as from the Applicant who adds matter after the end of that year, during prosecution
    at the EPO.

    Yet, while the
    Applicant who adds during the Paris year benefits from Art 89 EPC, the one who
    adds after the Paris year ends (and ONLY the one who amends after the Paris
    year ends) has to struggle to get the amendment in, and succeeds only when the
    Absence of new matter is beyond the slightest doubt.

    Why doesn’t DG3
    protect the public from early amenders just as diligently as it does against
    those who amend later? I thought Art 54, 87 and 123(2) were all judged
    with identical severity. But they’re not, are they?

Comments are closed.