For discussion: Has the requirement that claims be «supported by the description» been perverted over time?

Reading time: 13 minutes

Objections with respect to alleged violations of Art. 84 EPC have become more and more cumbersome in recent times. I cannot really tell why. The law has not changed ever since enactment of the EPC 1973 in this respect. But the EPO Guidelines were tightened substantially, purportedly merely reflecting long-standing practice. My experience is different. Back in the days, I had only been required to make features of independent claims also mandatory in the specification. A no-brainer that did no harm, so I complied. Nowadays, practitioners are regularly facing extensive objections to tidy up the specification to large extent, causing anything but minimally invasive changes.

The 2021 revision of the Guidelines caused an outcry amongst users, and the EPO made certain minor amendments in the next revision cycle of the Guidelines (see e.g. here). The most controversially discussed sections of the Guidelines are F-IV, 4.3 (inconsistencies), F-IV, 4.4 (claim-like clauses); don’t miss to tick the ‘show modifications’ checkbox on the right side.

However, the more general question is still unresolved:

Why the heck do we have to adapt the specification at all?

In my view, there is just no sound legal basis for it. I am fully aware that this position is controversial, running against the gist of some rather old decisions of Boards of Appeal of the EPO. My line of thinking is as follows. As always, I stand to be proven wrong. 

Article 84 at a glance

The mere wording of Art. 84 EPC is plain and simple (emphasis added):

Art. 84 — Claims
The claims shall define the matter for which protection is sought. They shall be clear and concise and be supported by the description.

The second sentence of Art. 84 EPC requires the claims to be

    1. clear;
    2. concise; and
    3. supported by the description.
the travaux préparatoires

It is evident from the TP that the first two criteria, i.e. clarity and conciseness were meant to reflect Art. 6 PCT in the new legal framework of the EPC:

The same holds true for the «support by the description» requirement, with an initially pretty different (but much clearer) wording:  

No patent claim shall contain subject matter that is not listed in the description.

The TP are silent about the reasons why the initially very clear wording of the third criterion had not been further pursued but rather been amended to exactly match Art. 6 PCT. In my view, the most likely reason is an attempt of harmonization of the wording. There is nothing in the TP that could suggest an intent of the legislator to change the substance of the initial draft, i.e. that the third criterion was simply meant to exclude subject-matter from the claims that is not listed in the description.

In the first place, all requirements of this provision had been included in the draft Implementing Regulations. Later, the sub-committee proposed to adopt the exact same wording of Art. 6 PCT, and to move it to the Convention itself:

In the first place, the draft article read as follows (exactly as Art. 6 PCT):

However, the Conference did not like the wording «fully supported by the description» and asked the Working Party to examine deletion of «fully» and the replacement by a less restrictive wording:

And, indeed, the term «fully» was deleted:

Even though the term «fully» was not replaced by a less restrictive wording, the TP leave no doubt about the legislator’s intent to prevent the third criterion from being interpreteded in the sense of the initial draft, i.e. that the claims shall be «fully supported by the description».

This is how we ended up with the wording of Art. 71a at the Conference:

Here is the final wording of the article, which has never been changed ever since:

For sake of completeness, I have reviewed the records of the Washington Conference on the PCT — which do not provide any further insight.

T 1989/18 — 3.3.04: Shaking the current practice to the very foundations

Applicants rarely risk a refusal merely for non-compliance with a formalistic objection concerning the «necessary» adaptation of the description once allowability of a set of claims has been indicated by the ED. Not so in the case underlying T 1989/18 — 3.3.04 (EP 2 794 651; DREX and file wrapper). Accordingly, the Board of Appeal had to investigate whether there was any legal basis for demanding an adaptation of the specification at all.

The Board took an approach similar to what I take from the TP, i.e. the «supported by the description» simply means:

[T]he subject-matter of the claim must be taken from the description, it being inadmissible to claim any subject-matter which is not described.

This must not be confused with the first requirement of clarity, in this Board’s view:

When assessing clarity, the description cannot be relied upon to resolve a clarity issue in a claim, nor can it give rise to any such isseue if the definition of the subject-matter in a claim is clear per se.

Accordingly, it is held (¶7):

Article 84 EPC cannot serve as a legal basis for the refusal.

The Board did not find any legal basis in R. 42(1) lit. c EPC (in the absence of a non-unity objection; ¶8) and R. 48(1) lit. c EPC (only pertaining to patent applications, not patents — opposed to R. 42 EPC; ¶9-12), either.

Accordingly, the Board allowed the appeal and remitted the case to the ED with an order to grant the patent — with an unamended description.

the aftermath of T 1989/18 — 3.3.04

To the best of my knowledge, one further Board of Appeal has explicitly adopted the approach of T 1989/18 — 3.3.04, i.e. in the matter T 1444/20 — 3.3.01 (¶2 et seqq.).

On the other hand, I am aware of four recent Board of Appeal decisions that dissent with T 1989/18 — 3.3.04. I will address them below.

T 1024/18 — 3.2.06

Decision T 1024/18 — 3.2.06 concurs with T 1989/18 — 3.3.04 in that the claims must be clear in themselves; criterion i) of Art. 84 EPC. However, with respect to the third criterion, this Board holds that «supported by the description» cannot mean support in only a part of the description, because this would be at odds with the wording «supported by the description» (apparently understood with an emphasis on the word ‘the’).

This is a somewhat circular reasoning, isn’t it? The claims are only supported by the description when they are supported throughout the description; otherwise, they are not supported(?).

What is more, the TP are clear in that a less restrictive understanding than «fully» supported by the description was intended by the legislator. Against this background, I am having a hard time to accept a reasoning according to which support by (only) «part» of the description is not sufficient (¶3.1.8). 

Further, the Board mentions that the provision (Art. 84) had been moved from the Implementing Regulations to the Convention itself, because of its importance for national infringement procedings; ¶3.1.9. The Board draws from this that also the requirement of «supported by the description» serves this purpose.

I cannot deduce this from the TP; see above. The TP refer to the importance of the provision, but they are silent about which criteria thereof were considered to be of so much importance for national infringement proceedings. In my view, there is no reason to assume all three criteria were equally meant. For instance, I fail to see that the legislator had conciseness of the claims in mind with respect to the outstanding importance in infringement litigation. In my experience, conciseness of a claim is no issue at all. So, why would one assume that «supported by the description» was meant to be of oustanding importance?

T 2293/18 — 3.5.02

This decision holds, with reference to T 409/91 (¶3.3-3.5), that «supported by the description» requires that the claimed subject-matter has a basis in the description, which is to make sure that the claims only extend to subject-matter which, after reading the description, are at the disposal of the skilled person; ¶3.3.5.

Further, with reference to T 659/93 (¶3.4), this decision holds that «supported by the description» is to assure that the scope corresponds with the actual contribution to the art, and that the claims can be worked over the entire scope; ¶3.3.5. .

All this is not really controversial. But does it help to answer the question?

The decision goes on to hold that «it follows» that claims and description must not contradict each other, as part of a single document:

I wonder: Why is this a necessary consequence? I feel that the claims may well be workable over their entire scope, and they may well correspond to the actual contribution to the art, without the need to perfectly match with the description («übereinstimmen»). In any event, I feel requiring a perfect match puts the bar at the wrong level: «Full» support was clearly not intended by the legislator.

T 0121/20 — 3.2.01

This decision concurs with the reasons of T 1024/18 — 3.2.06 (which I do not find very convincing; see above), without adding anything in substance; ¶10.2.

UPDATE 11 June 2022: Yet another decision of the same Board 3.2.01 has been published on 9 June 2022, i.e. T 1516/20. This decision, again, merely refers to and consents with T 1024/18 — 3.2.06, but does not add anything in substance; ¶5.

T 2766/17 — 3.2.02

This decision agrees with T 1989/18 in that the description cannot be relied upon to resolve a clarity issue in the claim, but it does not follow the a contrario reasoning:

The decision illustrates the unitary character of a patent with the example that terms may be given a special meaning in the description, i.e. that a patent specification can become its own dictionary. In conclusion, this decision holds that statements in the description contradicting the plain claim wording may cause doubts as to the intended meaning of this wording:

I am not convinced. Primacy of the claims is the overriding principle of any claim construction. If parts of the description really contradict the plain wording of the claim, I fail to see the clarity issue. These parts are irrelevant and can simply be ignored.

Practice under the PCT

The PCT International Search and Preliminary Examination Guidelines (PCT/GL/ISPE/11) are meant to reflect the practice under Art. 6 PCT, i.e. under the more restrictive wording «fully supported by the description». The following is noted therein:

Notably, not each and every inconsistency has to be removed even under Art. 6 PCT:

However, inconsistencies which do not cause doubt as to the meaning of the claims may be overlooked.

I feel the above examle in the PCT Guidelines is telling — because it illustrates how this practice will inevitably lead to undesired results. I cannot see how a semiconductor in one «embodiment» of the description could cause doubt as to the meaning of the plain wording «electronic tube» in a claim. A tube is a tube, not a semiconductor. Period. The disclosed (non-)embodiment is simply not literally covered by the claimed invention — and we all know what that means for a patentee when trying to assert infringement under the DoE. On the other hand, if this (non-)embodiment is deleted before grant in order to meet the «supported by the description» requirement, a semiconductor may easily be asserted as infringement under the DoE. The result is the exact opposite of legal certainty and serves no public interest at all.

Comparative study of national practices

WIPO has conducted a comparative study on  the relationship of the claims and the disclosure back in 2002; SCP/7/6.

The practice of the EPO is extensively reviewed, and summarized as follows:

No mentioning of amendments to the description at all. Essential features must be taken up in the claims, but that’s it. The same holds true for the other jurisdictions assessed.

In any event, no national practice in a major jurisdiction that I am aware of requires applicants to adapt the description, and in particular not in the way the EPO does. Against this background, I feel that it is somewhat far-fetched to assume that the PCT contracting states really intended to require substantial amendments to the description by way of Art. 6 PCT (and even less so under the less restrictive wording of Art. 84 EPC).

Yet another WIPO document gets it right, in my view. SCP/22/4 holds:

I would just leave it at this. No touching of the description is necessary towards this end.

Where are we going from here

Within a time frame of only a few months, we have seen six decisions on the issue of «supported by the description» and its implications under Art. 84 EPC. T 1989/18 — 3.3.04 cannot simply be ignored anymore since its approach is explicitly supported by yet another Board in T 1444/20 — 3.3.01. Four other Boards dissented. I feel it is only a matter of time until a Board decides to refer the issue to the Enlarged Board of Appeal; Art. 112(1) lit. a EPC. The sooner the better, to finally get things straight.

A referal by the President of the EPO under Art. 112(1) lit. b EPC is not to be expected, in my opinion.


T 1989/18 — 3.3.04 (HTML)

Chair: Bart CLAES
Member: Oskar LECHNER
Lukas BÜHLER (lm)

Decision of 16 December 2021:

DISSENTING WITH T 1989/18 — 3.3.04

T 1024/18 — 3.2.06 (HTML)

Chair: Michael HARRISON
Member: Martin HANNAM
Wilhelm UNGLER (lm)
Claes ALMBERG (lm)

Decision of 1 March 2022:

T 2293/18 — 3.5.02 (HTML)

Chair: Richard LORD
Member: Claudia VASSOILLE
Anna BACCHIN (lm)

Decision of 31 March 2022:

T 0121/20 — 3.2.01 (HTML)

Chair: Giovanni PRICOLO
Member: Andrea WAGNER
Peter GUNTZ (lm)

Decision of 11 March 2022:

T 2766/17 — 3.2.02 (HTML)

Member: David CECCARELLI
Christof SCHMIDT (lm)

Decision of 17 March 2022:

CONSENTING WITH T 1989/18 — 3.3.04 

T 1444/20 — 3.3.01 (HTML)

Chair: Albert LINDNER
Member: Regina HAUSS
Lukas BÜHLER (lm)

Decision of 28 April 2022:


Enter your name and email address below to get notified of new posts by email.

Print Friendly, PDF & Email

10 Replies to “For discussion: Has the requirement that claims be «supported by the description» been perverted over time?”

  1. T 1444/20 has already be commented on another blog:

    Just have a look. It could be quite helpful.

    It interesting to note that in both cases T 1989/18 and T 1244/20 we find the same legal member. It might thus not be a surprise when we find a similar reasoning.

    In T 1444/20 the board added a criticism of the Guidelines F-IV, 4.4. The board however sovereignly disregarded Art 20 RPBA20. As an identical article was in the RPBA07, the board cannot claim absence of knowledge. I would expect that a board which thinks fit to go against a long line of case law should at least be able to respect its own rules of procedure. This was manifestly not the case. People in glass houses….

    I also invite you to have a careful look at T 2766/17. In this decision the BA, being aware of T 1989/18 concurs with it, in that the claims should be clear by themselves, but only to the extent that the description should not contain statements contradicting the plain claim wording.

    In other words, the description has to be adapted!

    The case law of the BA has made it clear what it means to adapt the description to the claims. Any inference that the claims could be interpreted differently than their plain wording is to be deleted, whether the profession likes it or not. Just have a look at T 2677/17, it could not be clearer.

    None of the commenters so eager not to amend the description seem not to have grasped the idea that the job of the EPO is not to allow fancy interpretations of the claims by leaving statements which can give a different meaning to the claims. That is what is required by Art 84.

    Interpretation of the claims and the way the DoE is applied by national jurisdictions is not a matter for the EPO. But I maintain that it would have brought some harmonisation if the protocol would have defined what is to be understood by an equivalent. We are left with an Art 2 talking about equivalents, but nowhere there is a definition of what could be an equivalent.

    When part of the disclosure is not any longer part of the invention as defined in the claims, what does it bring to insist upon leaving it as being part of what is claimed? I do not like using the word cheating, but in my opinion it is what it boils down to. That the EPO wants to put all applicants on a par is just a legitimate aim.

    When I see the pemetrexed case, it becomes abundantly clear what has happened there. Thanks to a very broad statement in the introduction of the description, the proprietor was able to claim an invention he had never made at the time. The only example disclosed was pemetrexed disodium and nothing else. The applicant at the time had manifestly no other data at hand otherwise he would have submitted further information about other salts. But he bluntly refused to limit the disclosure to pemetrexed disodium.

    No surprise that after this case, the screws had to be tightened at the EPO. Please do not forget.

    I do not intend to continue a long discussion on this topic, be it on this thread or on any other. The arguments on both sides have been exchanged and there is nothing to add.

    Let’s agree that we disagree, but do not claim from the EPO something which it cannot deliver in view of the long lasting line of case law on this topic.

    The description has to be adapted and it is not a single legal member who will be able to change things, even he might be convinced of his point of view. Other boards have shown him how to look at Art 84 which does not limit itself to clarity but also mentions support by the description.

    1. I do agree that we disagree on this topic, Daniel. I have seen this coming, and I have no hard feelings about it. This is how discussions on aspects of law frequently end up – until the highest instance finally resolves it.

      Just a few thoughts in reply to the points that you raised:

      It is true that in both T1989/18 and T1444/20 the same legal member was involved. However, decisions are issued by Boards, not by individual members. As a matter of fact, we have two Boards now that question the legal basis for requiring adaptation of the description. What is more, both Boards are in the Chemistry / Pharma / Biotech field. In my experience, this is where the pain caused by the adaptation requirement is most severe. In my view, this is not a coincidence. In my view, it won’t take long until the EBoA will be involved (I hope so, at least); it cannot be that in a certain field of technology a different practice on an aspect of law is established than in other fields.

      The Board’s criticism of the Guidelines (F-IV, 4.4) in T 1444/20, para. 2.6, is straight to the point; don’t you agree?

      I analyzed T2677/17 in my post. But I disagree that this decision “could not be clearer”. The reasoning in this decision is leap-frogging from a non-controversial issue (that a description may comprise a binding definition of a term used in a claim) to a conclusion (that contradictions / inconsistencies need to be resolved). What is missing is a compelling reasoning in between. A definition in the spec on the one hand, and contradictions or inconsistencies on the other hand are totally different things; only the latter is at stake here. The “dictionary” element comes into play in case of definitions; this is what dictionaries are all about. A blunt statement in a decision (be it as clear as it may be) cannot replace a thorough legal assessment. This is what distinguishes an unfounded opinion from a well-reasoned decision, in my view.

      I wonder: What is your opinion on the TP, Daniel? I am not aware of any BoA decision that has ever assessed the deletion of “fully”, with the explicit intention of the legislator to get rid of this “restrictive” term? In this respect, I am really troubled by T1024/18. They did look into the TP, picked something that seemingly helped their case and – at least in my view – overinterpreted it; but missed the essential part, i.e. the deletion of “fully”. It is exactly this bit of the TP that clearly runs against their reasoning that a support in “part” is not sufficient.

      And what about the comparison with the even more restrictive wording of Art. 6 PCT, requiring the claims to be “fully supported by the description”? No signatory state of the PCT ever came up with an interpretation of what is commonly referred to as the support-requirement in their interpretation at national level. I find it highly unlikely that the signatory states of the PCT intended this wording to impose more severe requirements than what they all have implemented at national level. It is only the EPO that came up with such a consequence, nota bene based on a less restrictive wording. That comes across a bit shaky, doesn’t it?

      With respect to the long line of established case-law on the issue: You are right, this line of case-law exists. But their reasoning is leap-frogging over the essential issues in pretty much the same way as T2677/17. A long-established practice as such is not impressing me much. It may still be wrong. For instance, it took ages until we knew whether a patentee may oppose his own patent, and even the EBoA changed its mind on this issue.

      I tried to keep my post free from personal opinion, to the extent possible. I have my opinion, you know that. I focussed on the legal research instead, the travaux préparatoires, etc. I stand to be corrected and proven wrong on these aspects. And I will accept whatever the EBoA holds is the right approach.

      Finally, re Pemetrexed: This is a wholly different story. I know this case in and out since we were involved on behalf of the defendant / alleged infringer. The claims were limited to pemetrexed disodium, and the applicant did not “bluntly refuse to limit the disclosure to pemetrexed disodium”. Please see the DREX with marked-up amendments re “pemetrexed disodium” here.

  2. If I understand correctly, Daniel Xavier Thomas believes that the decisions of the national courts concerning the Pemetrexed case are an aberration, and that it is the duty of the EPO to try to ensure that such aberrations do not occur again.
    I wonder how widely this belief is shared at the EPO.

    1. The pemetrexed decisions are, to large extent, very classic infringement assessments under the DoE. Whatever kind of questionnaires the national courts applied. That’s no EPO business. The tricky thing was that, in view of the very strict approach of the EPO under Art. 123(2) EPC, the applicant was not allowed to claim pemetrexed, but rather only pemetrexed disodium. The applicant immediately complied, without hesitation or fighting for a broader claim on appeal. Later, when it came to infringement, other counterions than sodium were held to be still infringing. One may like it or not (I definitely don’t), but this is how the courts decided. This has nothing to do with adaptation of the description to the limited claim; this has been done.

    2. I never said that the decisions on the pemetrexed case where an aberration.
      I simply said that during prosecution the then applicant made no effort whatsoever to show that other salts could work.
      During prosecution when faced with the objection that the only example was with disodium, he never made the attempt to show that his invention was applicable to other salts. He could have tried to file further experimental data, but he was only interested in a very quick grant. His only attempt was to claim pemetrexed in general. He was then faced, rightly, with an objection of added matter.
      It was a disgrace that Lord Neuberger in his famous decision criticised the examiner for this. He missed an opportunity to enquire about the way added matter is handled at the EPO
      The then applicant insisted to leave I n the description a formulation implying a much broader contribution to the art than what he had actually done.
      To paraphrase Me Mathely, the Patentrecht discovered his patent on the date of a potential infringement!

  3. Dear Martin,
    I would simply come to the conclusion that we disagree on this topic.
    I doubt personally that we can expect the President of the EPO to file a referral on this topic.
    It might come from a board at best.
    It would indeed be the best solution provided the board does not rewrite the question or come up with a dynamic interpretation so as to please the management.
    It is perfectly possible for a board to criticize the Guidelines, but it should at least abide by its rules of procedure.
    As far as pemetrexed is concerned, I have also been involved in it deeply and I claim to know the case quite well.
    I have even found out that the district court of The Hague has refused to consider infringement in view of the behavior of the then applicant through prosecution. This judgement was set aside by the Court of Appeal.
    I would go as far as to claim that the fastening of the screws in matters of adaptation of the description at the EPO started with the pemetrexed case. The only example disclosed was disodium and nothing else, but the applicant bluntly refused to limit the description to what he had invented at the time. Just look at the register!
    I will not continue the discussion as we are turning in circles and everything what could be said has been said. A look at the TP is not really helpful in this matter.

    1. Same here in Switzerland, the first instance had held that the patentee’s conduct during prosecution prevented him from asserting infringement of anything but pemetrexed disodium. This decision was set aside by the Supreme Court, too. Mere compliance with a formal requirement (Art. 123(2) EPC) is no waiver of scope under the DoE. One may like it or not (I definitly don’t), but that’s what the highest courts held in so many countries. The EPO cannot do anything about it.

      Please, help me with the applicant’s blunt refusal to limit the description. The DREX (see deep link in my previous reply) shows that the description was amended. Here is a highlighted version of the most important passage that has been amended. What am I missing?

  4. Martin, you are sadly wasting your time attempting to engage in a discussion with DX Thomas on this. As shown by his comments on this topic on numerous recent IPKat posts, both under his own name and as an “Attentive Observer”, he is convinced of his view and will not enter into serious discussion of the reasons for it.

    It is rather sad and frustrating that someone so eminent, who has always contributed so thoughtfully and reasonably to the discussion of the case law elsewhere, should have inexplicably decided on this one particular topic to adopt such a blinkered stance scarcely differing from that of an internet troll.

    There is a very interesting discussion to be had, and your blogpost here is thought-provoking and (to this reader) convincing. Now, would anyone who is convinced of the EPO’s current approach care to engage in a real discussion of that alternative point of view?

    1. I have a point of view which might not be shared by many representatives which are all quite adamant that the description should not be amended.
      This position is based on two decisions which might have been taken by different boards, but both with the same legal member. It is probably just a coincidence.
      There remains a long lasting line of case law disagreeing with T 1989/18 (or T 1444/20).
      I would allow myself to say that all representatives, Martin included, hoping for an introduction of those two decisions into the Guidelines, will eventually feel like the famous Godot.
      Just look at Point 5 of the reasons of T 1516/20 not published later than June 9th

      1. Thank you, Daniel, for bringing T 1516/20 to my attention. I have included it in my summary above, properly marked-up.

        I note that this newest decision is from the same Board 3.2.01 as T 121/20 (already mentioned earlier in my post), and it comes along with the same “reasoning”: This Board does not consent with T 1989/18 and consents with T 1024/18, without adding _anything_in substance. I have addressed the flaws of T 1024/18 in detail in my post. In sum, the issue will not be resolved by way of votings.

        At some point, a Board or the Enlarged Board will (have to) take the long way and dive into the TP (in total, not only coming up with the convenient bits). There is a “long lasting line of case law”, no doubt about it. But it simply never placed the practice on solid legal ground. Be the line of case law as long as it may; it currently doesn’t impress me much. “We always did it like that” is no reason to continue. If that what has been presented by now – and in particular in T 1024/18 – is really all there is in support of the current practice, then the legal basis is more than shaky. As soon as it is supported by a proper reasoning, I will be silent. Not so any earlier.

Leave a Reply

Your email address will not be published.