Does the requirement for a "technical contribution" in a novel selection survive? (if selecting from convergent lists) (T 1621/16)

The EPO has famously strict requirements with respect to support for amendments. An important principle in the prosecution of patent applications is that any amendments the applicant makes to the application must have basis or support in the application as filed. The EPO is particularly strict with regards to amendments based on the combination of features from different parts of a patent specification. A recently published decision from the Boards of Appeal (T 1621/16) clarified the criteria for assessing whether a selection from multiple “convergent lists” (i.e. lists with features of increasing preference) is supported. 

The Board of Appeal found that the selection from convergent lists is different in principle from a selection from a normal, non-convergent lists. However, to have basis in the application as filed, the Board reasoned that the selection from multiple convergent lists must not be identified with a previously undisclosed technical feature and must be supported by a pointer elsewhere in the specification. Importantly, basis and novelty are inextricably linked. In particular, the same test for "disclosure" is used. Therefore, if applied by other Boards of Appeal, this decision also has implications for the assessment of novelty of selection inventions from multiple convergent lists.

Legal Background – Selection from lists

At the EPO, a patent application or patent may not be amended to introduce subject-matter that was not disclosed in the application as filed (Article 123(2) EPC). The purpose of this provision is to prevent an applicant from claiming more than they originally had possession of at the filing date. The EPO’s support requirements are stricter that than that of USPTO. The EPO particularly has a narrow understanding of what can be considered as “disclosed in the application as filed”, at least compared to the USPTO. One way in which the EPO’s strict understanding of disclosure manifests is in the EPO’s criteria for allowing amendments based on the selection of features from multiple lists.

An application may include multiple lists of different features of an invention. Selection of a feature from a single list is considered to have basis. However, at the EPO, it is not permissible to select and combine features from multiple lists, unless the combination is specifically disclosed elsewhere in the specification. The case law of the Boards of Appeal on whether this type of selection is allowable is consistent and well-established. Where the EPO case law is less consistent, is with regards to the selection from multiple so-called “convergent lists”. Convergent lists are those in which items of the list are singled out as being more or less preferred. Two examples of convergent list are:
White chocolate and kiwi cheesecake
  1. A cheesecake, preferably a chocolate cheesecake, most preferably white chocolate cheesecake.
  2. A cheesecake comprising fruit, preferably green fruit, most preferably kiwi.
As with the selection from a single non-convergent list, an amendment based on the selection of a single item from a convergent list, e.g. "chocolate cheesecake" would have basis. But what if you select and combine items from 2 convergent lists? You may wish, for example, to select and combine options with different preference ratings from a pair of convergent lists, e.g. to claim white chocolate cheesecake comprising green fruit. To date, the case law on whether such selections are permissible has not been consistent (see T 2273/10, T 2237/10).

Convergent versus non-convergent lists

The recently published appeal decision T 1621/16 concerned Procter & Gamble's patent EP 2264138. The patent related to washing up liquid. Claim 1 of the patent specified the composition of the washing up liquid. The main question on appeal was whether an amendment during prosecution which selected features from two convergent lists, has basis in the application as filed.

The Board of Appeal (3.3.06) first clarified that the selection from convergent lists was a different in principle to the selection from non-convergent lists. In non-convergent lists, the Board of Appeal reasoned, each element is distinct. The selection and combination of elements from two such lists therefore singles out a distinct combination from among several distinct alternatives.

By contrast, in a converging list, each progressively narrower element falls under the scope of the proceeding elements. For example, white chocolate cheesecake is a subset of chocolate cheesecake, which is a subset of cheesecake. Consequently, each element of a convergent list does not represent a distinct feature. Instead the elements represent more-or-less restricted versions of the same feature. The Board of Appeal thus reasoned that the combination of elements from multiple convergent lists did not represent an arbitrary selection of distinct elements.

The Board of Appeal in T 1621/16 therefore appeared to be heading in the direction of a more lenient approach to the selection from multiple convergent lists compared to the selection from non-convergent lists. However, the Board of Appeal emphasised that their reasoning did not mean that selections from multiple converging lists will always have basis (r. 1.7.3). The Board of Appeal instead considered the selection from non-convergent lists as analogous to another form of selection from lists, the deletion of an element or elements from a list ("shrinking lists").

What does shrinking lists mean?

A specification may include two lists of features of a chocolate and fruit cheesecake, such as:
  1. ...wherein the chocolate is dark, milk, white or orange chocolate.  
  2. ...wherein the fruit is grape, apple, lime or starfruit.
According to established case law, the deletion of, for example, milk chocolate and lime cheesecake from these lists would not represent added matter, provided that two conditions are satisfied. First, the deletion of the elements did not single out a previously undisclosed combination of elements. Second, shrinking of the lists should not lead to the definition of a previously undisclosed invention, i.e. a combination having a previously undisclosed technical effect (Case Law of the Board of Appeal, II.E.1.6.3).

The criteria for selecting from converging lists

The Board of Appeal reasoned that, if selecting from convergent lists was equivalent to shrinking lists, the same criteria should be applied. Thus, the selection from a convergent list should also not single out a previously undisclosed combination of elements, and should not be associated with a previously undisclosed technical effect. 

If, for example, the selection of a white chocolate and green fruit cheesecake was identified with a previously unidentified technical effect (e.g. resistance of the cheesecake to wasps at a picnic), the selection would not have basis. Furthermore, if there was no pointer anywhere else in the specification to the combination of white chocolate with green fruit in a cheesecake, an amendment to white chocolate and green fruit cheesecake would also lack basis. In the case in question (relating to the less-appetising field of washing up liquid), the Board of Appeal found that both of these criteria were satisfied. The amendment was therefore found to have basis in the application as filed. 

The requirement for a "technical contribution" of selection invention survives? 

The principles of basis and novelty are inextricably linked. The test for novelty and added matter apply the same Gold Standard test that a disclosure must be clearly and unambiguously derivable (Case Law of the Board of Appeal, II.E.1.2.1). The recent Board of Appeal decision on the selection from convergent lists therefore has implications for the novelty of inventions in view of prior art disclosing converging lists. Applying the Board of Appeal's reasoning in T 1621/16, the selection of multiple features from converging lists in the prior art would be novel if the selection was associated with a previously undisclosed technical feature. The question remains whether the identification of a previously undisclosed technical feature of the selection would also be necessary for novelty. 

The novelty requirement that a selection invention must relate to a "technical contribution" has otherwise fallen out of favour with the Boards of Appeal. Such a requirement is now broadly considered to be more properly associated with an assessment of inventive step rather than novelty. There have been recent calls to refer the issue of whether the requirement is necessary to the Enlarged Board of Appeal (EBA). It will therefore be interesting to see whether the reasoning of T 1621/16 is applied by other Boards. If nothing else, the case is a reminder that the Boards of Appeal may not yet have reached a consensus on the appropriate criteria for assessing the novelty of selection inventions. 
Does the requirement for a "technical contribution" in a novel selection survive? (if selecting from convergent lists) (T 1621/16) Does the requirement for a "technical contribution" in a novel selection survive? (if selecting from convergent lists) (T 1621/16) Reviewed by Rose Hughes on Wednesday, March 25, 2020 Rating: 5


  1. I'm a fan of the "gold standard" and the notion of "derivable". What I don't yet see is convincing and uniform mapping of "derivable" onto the knowledge base of the imaginary skilled addressee.

    What is needed? Top class advocacy and top class membership of the Technical Boards. I am an optimist. Just as the Boards have sorted out (more or less) what is "technical" and what is not, with time they will converge on a methodology and logic that will inform everybody how to work out what the skilled person can and cannot "derive, directly and unambiguously" from any given document.

    Meanwhile, it is up to we advocates, to help the Rapporteurs and the Chairs of the TBA's, first to take a real world (and therefore necessarily complicated and nuanced) view of the knowledge and thought processes of the skilled reader of a new scientific or engineering disclosure how to solve a technical problem, and then the courage and energy to write a Decision in accordance with that view (as opposed to writing a quick, easy and formulistic Decision just to meet management productivity targets).

  2. MaxDrei, I hope that you don't mean by 'gold standard' that we take a strict line and simply decide on the basis of the most simple analysis. When selecting from 2 lists there is a genuine issue about how to balance fairness to the applicant and third parties. Taken to the extreme it means that any 'new combination' derived from features in the specification, such as A+B+C, adds matter if those 3 features are not specifically mentioned in combination. Examiners at the EPO do raise such objections even when the features are not disclosed as part of a traditional list, but are being discussed as different embodiments. The EPO seems to have taken the view that some type of inventive step analysis cannot be used to assess whether the combination is disclosed (i.e. using the rule that all obvious combinations are disclosed for the purpose of providing basis for claim amendment). There is the vague notion of implicit disclosure which does not lead to a workable test. After that we are left with 'pointers' to the combination having to be present in the specification, with little certainty as to what a pointer is, but it being related to preferred or exemplified embodiments. I predict none of these approaches will crystallise into the Decision that you desire which seems to be a concrete set of rules to decide on what is disclosed.

    My alternative approach, which I might as well suggest as coronavirus reshapes the economic landscape, is to ask whether the applicant 'deserves' the amendment they are trying to make in view of the contribution they are making over the prior art. If it is an epic contribution we should be lenient on added matter, and let them claim the invention as they wish. If it is a modest invention we should judge with a lot more caution what 'adds matter' in view of the fact they have contributed little. That does not take away the uncertainty but may make the system 'fairer'.

    PS. It is good to see that you seem to be well in these difficult times.

  3. Thank you, Holistic. If you have the time and the inclination, join us on the Kluwer blog, debating inter alia the EPO's approach to the problem of the undisclosed intermediate generalisation.

    Yes, I am glad we are both well. The times might already be, as you say, "difficult" but they are going to get a whole lot more "difficult" before they get better. Let us hope that debating patent law, while confined at home, achieves something useful to all of us.

  4. Dear Holistic,

    Whilst have a certain sympathy with the argument that the applicant might “deserve” certain amendments, the Case Law of the Boards of Appeal is nevertheless very clear in that respect. Inventive step is not what matters, and even if it might be obvious for a person skilled in the art, this is not the criterion adopted by the Boards.

    When an application is filed, nobody knows in advance whether it will represent a great contribution over the prior art or not. For this reason alone the notion of “deserving” is antonymic to the necessary safety of third parties. And that what the Boards have been saying for a long time.

    Whilst it is true that the whole reservoir of features originally disclosed can be used /when amending an application or patent, cf. T 81/03, it can however not be considered to be a reservoir from which individual features pertaining to separate sections can be combined in order artificially to create a particular combination, cf. T 1334/09, T 296/96 or T 1306/12.

    In T 1724/08 it was held that the criterion of "direct and unambiguous disclosure" is quite different from the criterion submitted by the appellant, namely whether a person skilled in the art in view of the disclosure is enabled to achieve the envisaged result, i.e. the claimed functionality, without undue difficulty.
    This latter criterion would allow many different new definitions of subject-matter, none of which would need to be derivable "directly" and "unambiguously" from the original specification.

    In T 583/09 it was held that technical information which is clearly and unambiguously derivable from the original application can be explicit or implicit.
    The term "implicit disclosure" should not be construed to mean matter that does not belong to the content of the technical information provided by a document but may be rendered obvious on the basis of the content of the disclosure.
    The term "implicit disclosure" relates solely to matter which is not explicitly mentioned, but is a clear and unambiguous consequence of what is explicitly mentioned

    But that is exactly what you suggest. If you have disclosed three separate embodiments you have not disclosed a combination of all three embodiments, and the general rubber sentences one can find in general at the end of the description will not help whatsoever.

    It is hard, but we should not forget that added subject-matter and novelty are the two sides of the same coin. What is lost on one side is gained on the other and vice versa.

  5. As ever, Attentive sets out clearly the established jurisprudence of the TBA's. I agree with him that the boilerplate paragraph at the end of a typical US-drafted specification, to the effect that the skilled reader can combine everything with everything, is superfluous and adds nothing to the "disclosure".

    But that's the whole point. It is superfluous precisely because the skilled reader knows that anyway. The skilled reader, scouring the text and drawings, hungry for useful teaching, is doing that anyway and (unlike litigation lawyers) does not need the paragraph at the end to remind him to do that. Who remembers the toy for little girls, with a paper figure accompanied by cut-out items of clothing, hat, blouse, skirt, shoes. My sister could play for hours, artfully combining one skirt with a different hat, gloves and shoes. To say that the only thing "derivable" in the toy was the combination on the top of the box is silly. A bit more effort please, from the Boards of Appeal, to put themselves in the shoes of the real world skilled addressee, not only for the purposes of Art 123(2) but also for Art 54 and Art 87 EPC.

    Attentive, yes of course the same standard should apply, for all of these, what is "disclosed". As I say, it's much harder 8and takes more time and mental effort) to write a reasoned decision explaining why the intermediate generalisation is implicit (or that D1 does indeed destroy novelty) than it is to cover one eye, say "I don't see it" and find that the claim offends Art 123(2) or that the claim is novel over the disclosure of D1.


All comments must be moderated by a member of the IPKat team before they appear on the blog. Comments will not be allowed if the contravene the IPKat policy that readers' comments should not be obscene or defamatory; they should not consist of ad hominem attacks on members of the blog team or other comment-posters and they should make a constructive contribution to the discussion of the post on which they purport to comment.

It is also the IPKat policy that comments should not be made completely anonymously, and users should use a consistent name or pseudonym (which should not itself be defamatory or obscene, or that of another real person), either in the "identity" field, or at the beginning of the comment. Current practice is to, however, allow a limited number of comments that contravene this policy, provided that the comment has a high degree of relevance and the comment chain does not become too difficult to follow.

Learn more here:

Powered by Blogger.