G 1/24 requires that the description must always be “consulted” to interpret a claim. The question remains whether a definition in the description can override an otherwise clear meaning in the claims. In T 0715/24, Bristol-Myers Squibb (BMS) sought to persuade the Board of Appeal that a negative feature, read together with a paragraph of the description, gave the word “monotherapy” for a cancer treatment an unusual meaning that found basis in the application as filed. The Board of Appeal was not persuaded that the description of the granted patent could be used to alter the definition of monotherapy as it is normally understood in the art. On that ordinary reading, the claim was found to add matter over the application as filed.
This decision comes as we await the Enlarged Board of Appeal (EBA) response to G 1/26 on claim interpretation and added matter. The Board of Appeal in this case appeared to have no issues with how G 1/24 should be applied, and that a claim can add matter in view of the application as filed when read in the context of the description of the granted patent.
 |
| Monotherapy |
Legal background on claim construction, added matter and G 1/24
Article 123(2)
EPC prohibits amendments that introduce subject-matter extending beyond the
content of the application as filed. The test is the “gold standard” confirmed
by the EBA in G 2/10. An amendment is allowable only within the limits of what
a skilled person would derive “directly and unambiguously, using common general
knowledge, and seen objectively and relative to the date of filing, from the
whole of the application as filed” (r. 9). It is, infamously, an unforgiving
standard (IPKat). Before it can be applied, however, the
claim must first be construed. The question of added matter, as the Board of
Appeal put it in the present case, “can only be answered on the basis of a
proper construction of the claim” (r. 1).
In G 1/24, the EBA held that the description and drawings shall always be consulted to interpret the claims, and not only where a claim is unclear (IPKat). What the EBA arguably did not settle in G 1/24 was what “consulting” the description actually entails, and in particular whether a definition in the description can be used to give a claim term a meaning it would not otherwise bear (IPKat). That purportedly unresolved question has now generated its own referral. In G 1/26, the EBA has been asked how far description definitions can influence an assessment of added matter (IPKat).
Case background: Checkpoint inhibitors
Nivolumab (BMS’s OPDIVO)
is one of the anti-PD-1 checkpoint inhibitors that has reshaped the treatment of
melanoma and other solid tumours. The commercial stakes are large. The global net
sales for OPDIVO were $10 billion in 2025. OPDIVO is now also marketed in combination
with the anti-LAG-3 antibody relatlimab (OPDUALAG),
approved in 2022, which reached $1.2 billion in global net sales in 2025. The 480
mg dose of nivolumab is indicated in the approved labels for both OPDIVO and OPDUALAG.
The case in T 0715/24 concerned EP 3288980, owned by BMS, relating to the treatment of PD-L1-positive melanoma using an anti-PD-1 antibody. Opposed by five Opponents, inter alia for added matter, the patent was revoked by the Opposition Division. Three Opponents (respondents III, IV and V) took part in the appeal.
Claim 1 of the
main request was directed to “a composition comprising nivolumab for use in
treating a melanoma” in a PD-L1-positive patient, “to be administered a flat
dose of 480 mg of nivolumab as monotherapy once every four weeks”, with a
negative feature stating that the patient is not administered a combination
of nivolumab and an anti-CTLA-4 antibody (such as ipilimumab).
The negative limitation
introduced into the claim makes commercial sense. “Monotherapy” might be
considered to exclude the newer OPDUALAG relatlimab
+ nivolumab combination. However, a look at the approved label for OPDIVO reveals
how the monotherapy language found its way into the claim. The label makes a
clear distinction between nivolumab “monotherapy” and nivolumab combined with the
CTLA-4 inhibitor ipilimumab. A combination with relatlimab is not contemplated
in the patent or the OPDIVO label. With this patent, BMS may have wished to cover both
of their marketed nivolumab products, each of which uses the 480 mg nivolumab
dose, either alone (OPDIVO) or in combination with relatlimab (OPDUALAG).
Interpretation
of the claim in view of the granted patent
Both sides
accepted the ordinary meaning of monotherapy as treatment with a single drug
(r. 4). The Patentee argued that the term in the claim should be understood as excluding
only the combination with an anti-CTLA-4 antibody, whilst permitting other
drugs (e.g. relatlimab). This was how the claim was originally drafted:
“1. A method for
treating a melanoma comprising:
(i) identifying a
patient having a PD-L1 positive melanoma tumor; and
(ii) administering to the patient an anti-PD-1 antibody or an antigen-binding portion thereof that binds specifically to a human PD-1 ("anti-PDl antibody monotherapy"), wherein the patient is not administered a combination of an anti-PD-1 antibody or an antigen-binding portion thereof and an anti-CTLA-4 antibody or an antigen-binding portion thereof that binds specifically to a human CTLA-4 ("combination therapy").”
The granted claim
narrowed the disclaimer. Where the claim as filed excluded a combination of
any anti-PD-1 antibody with an anti-CTLA-4 antibody, the granted claim excluded
only a combination of “nivolumab and an anti-CTLA-4 antibody”.
On claim construction, the Patentee argued
that, read as a whole, the patent pointed to a special definition of
“monotherapy” in paragraph [0040]. Paragraph [0040] stated: “The term ‘anti-PD-1
antibody monotherapy’ as used herein includes a therapy of an anti-PD-1
antibody without an anti-CTLA-4 antibody therapy.” The Patentee also argued that the negative feature in the claim also modified the definition of monotherapy.
The Board of
Appeal was, however, not persuaded. Interpreting the claim
“with a mind willing to understand” (r. 5), the Board of Appeal found the
negative feature did no defining work. In the words of the Board of Appeal,
“The negative feature is not drafted as a definition ... against the
well-recognised meaning of ‘monotherapy’ it merely introduces a fundamental
ambiguity” (r. 7).
Nor did paragraph [0040] define the term, thought the Board of Appeal. This paragraph merely listed embodiments and did so for “anti-PD-1 antibody monotherapy” in general, not “nivolumab as a monotherapy”. The term “monotherapy” was not mentioned specifically in the paragraph.
For the Board of Appeal, “monotherapy” therefore
kept its usual meaning, and the skilled person would understand it in the
context of the claim as meaning nivolumab used as a single drug. The original
claim as filed was, critically, not used to interpret the claim of the granted
patent.
Added matter in view of the application as filed
Given the construction of monotherapy according to its ordinary meaning, the question became whether that interpretation added matter. The Opponents first attacked the wording of the negative feature itself as added matter. However, the Board of Appeal expressly declined to decide that point. The objection that succeeded was aimed instead at the combination of features produced by the construction of monotherapy, i.e. monotherapy plus the specified dose and selected anti-PD-1 antibody. The question instead was whether the application as filed disclosed, directly and unambiguously, nivolumab for treating PD-L1-positive melanoma as an ordinary-sense monotherapy at the claimed flat dose.
Notably, the
Board of Appeal considered how claim 1 as originally filed had defined
“anti-PD-1 antibody monotherapy” in contradistinction to “combination therapy”,
in quotation marks: “The use of quotation marks and parenthesis indicates that
monotherapy is being assigned a specific definition in claim 1 as filed” (r.
13). Importantly for the Board of Appeal, this special meaning was not the
usual one now in the granted claim, a reading reinforced by claim 49 as filed,
which contemplated administering a further anti-cancer agent (r. 14). In
contrast to the granted claim, the original claim was therefore considered to
have an altered definition of monotherapy.
“[Claim 49] is
thus consistent with the special definition of monotherapy set out in claim 1
as filed, according to which the administration of additional therapeutic
agents is not excluded, provided that the patient is not administered the
combination of nivolumab and an anti-CTLA-4 antibody.” (r. 14)
Crucially, this
was not considered to change how the granted claim should be interpreted. Instead,
it was found to support the argument that the granted claims added matter.
The key issue for
the Patentee was therefore that monotherapy according to its ordinary meaning
had not been unambiguously disclosed, in the view of the Board of Appeal, in
combination with the selection of nivolumab and the specifically claimed dose. The
Board of Appeal accepted that the claims as originally filed used an altered
definition of monotherapy. However, the claim was considered to have been so
amended that a skilled person, reading it in view of the granted patent, would
consider monotherapy to have its ordinary meaning. Auxiliary request 1, which
redefined monotherapy positively as nivolumab “without another anti-cancer
agent”, fell for the same missing link (r. 21). The appeal was dismissed and
the patent revoked.
Final
thoughts
The problem for the
Patentee in this case was that they used an unusual definition of a term in the
original claims that they were unable to persuade the Board of Appeal could be
combined with specific limitations later added to the claim. Another warning
against using “unusual definitions”, and the particularly unforgiving nature of
the EPO’s added matter test. The present case thus drew a key distinction,
namely interpretation in view of the description as granted versus the
application as filed. The claim was construed in the light of the description
of the granted patent. Basis was tested against the application as filed,
including the original claims.
G 1/26 asks the
EBA whether the order that the description must always be consulted also
applies when a claim is tested for added matter (IPKat). In effect, the question is whether a
patentee may invoke a description definition to construe a claim into a shape
that has basis as filed and so escape an objection the prima facie wording
would attract. T 0715/24 is tangentially relevant. The Board of Appeal
consulted the description, as G 1/24 requires, and held that consultation
neither defined the term nor cured the ambiguity. T 0715/24 nonetheless leaves
plenty else that may be considered in G 1/26. In particular, the Board of
Appeal never had to decide whether a description definition can rescue a claim
because it found that there was no definition to apply. G 1/26 concerns the
harder case in which the description is found to explicitly define the disputed
term.
It is a welcome change to have a G 1/24 related case that can be mapped onto the commercial consequences (does anyone in the real world care if the tobacco in cigarettes is gathered or spiralled…?). The shifting definitions at play in this case make perfect sense when the on-market products are considered. Had the patent survived, it would have been interesting to see how courts might have chosen to interpret “monotherapy” in infringement proceedings.
Further reading
- Second medical use dosage regimen claim successfully traverses both insufficiency and “obvious-to-try” attacks (T 0799/16) (March 2021)
- The UPC’s approach to added matter and functional claims (Abbott v. Sibio Technology, UPC_CoA_382/2024) (March 2025)
- EBA decides G 1/24 on claim interpretation: The description should always be consulted (June 2025)
- First interpretation of G 1/24 (the description must always be consulted) (T 1561/23) (July 2025)
- G 1/24 in the spotlight: Description definitions do not override clear claim language (T 1999/23) (July 2025)
- No claim interpretation summersaults for the EPO (T 2027/23) (September 2025)
- The unforgiving gold standard: Why deleting a feature can add matter at the EPO (T 0422/23) (September 2025)
- First use of G 1/24 to broaden clear claim language (T 1849/23) (November 2025)
- BREAKING: Imminent new referral to the EBA on the relevance of G 1/24 to added matter (March 2026)
- New referral on claim interpretation and relevance of G 1/24 to added matter confirmed (G 1/26) (June 2026)
Reviewed by Dr Rose Hughes
on
Friday, August 07, 2026
Rating:
Follow the IPKat on LinkedIn 
![[Guest post] Standard Essential Patent Litigation in Brazil: A Short Story of Sharp Practices](https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjPoJ4p6Z_6wJ1RH5j15nBeM9q2Xq5vPk8DW8XL65QX87s3ApEQ6VduyOZmqFAQ2vttMw8bAx9zzvPePg1u43xvaEtS5YoHfnQw-ZLF9qh8is9NBDROjg6FCjH6yfvNbZjgz7UC6FKwngy-X4GCcvDw8utCtEFHr0OcBZwmzNqBGI8BxNITvI1DTQ/s72-c/Brazil%20Kat.png)
![[Book Review] Patent Claims: Interpretation, validity and infringement](https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhEEGu1bAwVHnVi9mEsTcoamOfyE29VHJJqdX6PDTtnB6PH_KNNNlVOG9hPWM3VVj5MY1c39tG-5wyJePJQ8HlkJj2fqNMYMfd9jrGy3ZNphsF7TULsBHTVo1WMbACCQBOAkU_nmGx3h26PTKAWhyuiVAG7sDgH4Heh-Q4YzsZMex1OCJPKMFCl0Q/s72-c/Book%20Review%20Kat.png)
Not a chemist and no longer in practice but nevertheless, on reading this excellent case report, I still have two questions.
ReplyDeleteFirst, tlet us circle back to foundational principles, and the disclosure content of the A publication, to a person skilled in the art, when we recall that the EPO's prohibition on added matter is to stop an Applicant "improving its position" (after establishing a filing date in a "First-to-File" jurisdiction). Here, is it just, is it fair to the inventor, that the patent owner here went down under the EPO's added matter jurisprudence. In other words, after filing, as a consequence of amending claim 1, was the content of disclosure to a skilled person in the B publication higher than that of the A publication? Or was it the same. what does the skilled person have to say about that?
Second, I see that the US family member went to grant in May. Was the Applicant in that jurisdiction able to secure the desired prosecution amendment? Will this case serve as a model case for comparing "added matter" vs "written description" case law in the two jurisdictions
The Board’s reasoning at point 7 of the decision is ... interesting.
ReplyDeleteTwo reasons are advanced for rejecting the patentee’s submissions, the first of which is:
“The negative feature is not drafted as a definition: the wording "wherein the patient is not administered ..." gives no indication that the feature serves to define or redefine the term, and against the well-recognised meaning of "monotherapy" it merely introduces a fundamental ambiguity”.
This is absolute nonsense.
The negative feature might not be a definition per se but its presence indirectly affects the skilled person’s understanding of the term “monotherapy”. This indirect effect is no different in principle to the repercussive effect that (at Reasons 14) the Board acknowledged PCT Claim 49 had on the interpretation of PCT Claim 1.
Also, it simply will not do for a Board to conclude that a feature “introduces a fundamental ambiguity” without then even attempting to determine how the skilled person might resolve that ambiguity (or at least interpret the claim in view of it). Unlike the Board, a skilled person presented with such an ambiguity would not simply throw up their hands and abandon the task of trying to make sense of the negative feature. They would instead do their best to arrive at an interpretation which was consistent with that feature ... as opposed to an interpretation which renders it entirely otiose.
The second reason advanced by the Board is:
“Nor does paragraph [0040] of the patent provide a definition; instead it lists various embodiments of the invention, and these relate to "anti-PD-1 antibody monotherapy" in general, whereas claim 1 concerns "nivolumab as a monotherapy" specifically”.
This reason fares little better when assessed against the facts. Para [40] of the patent reads as follows:
“The term "anti-PD-1 antibody monotherapy" as used herein includes a therapy of an anti-PD-1 antibody without an anti-CTLA-4 antibody therapy. The anti-PD-1 antibody monotherapy comprises, consists essentially of, or consists of administering one or more doses of an anti-PD-1 antibody to a patient in need thereof, but does not include administering an anti-CTLA-4 antibody. In one embodiment, the anti-PD-1 antibody monotherapy comprises administering one or more doses of nivolumab to a patient in need thereof, but does not include administering an anti-CTLA-4 antibody. In another embodiment, the anti-PD-1 antibody monotherapy comprises administering one or more doses of nivolumab to a patient in need thereof, but does not include administering an antibody specifically targeting a protein other than PD-1. In other embodiments, the anti-PD-1 antibody monotherapy comprises administering one or more doses of nivolumab to a patient in need thereof, but does not include administering another anti-cancer agent”.
The repeated references to embodiments using nivolumab directly challenge the Board’s assertion that the embodiments of para [40] relate (solely) to "anti-PD-1 antibody monotherapy".
The Board’s argument could perhaps be more subtle than it seems, namely an argument that the patent does not define “monotherapy” in connection with nivolumab treatment. However, this would require not only turning a blind eye to the context provided by the description but also the reliance upon the completely unjustifiable assumption that “monotherapy” suddenly means something different when applied to a specifically identified embodiment of a generically defined antibody.
Frankly, all that the Board has done is apply one set of (claim) interpretation principles to the disclosures of the application as filed and a completely different set of principles to the claims as granted. The Board’s decision therefore teaches us that using different interpretation principles leads to different interpretations. Who knew?
I think what the Board of Appeal is getting at here is:
Delete- The application as filed only discloses administering an anti-PD-1 antibody without an anti-CTLA-4 antibody therapy. It does not disclose an anti-PD-1 antibody monotherapy generally, in the sense that only an anti-PD-1 antibody and no other therapy is administered.
- The claim as granted is restricted to monotherapy. This is to be interpreted in the normal sense as no other therapy is administered.
- The claim as granted also states that anti-PD-1 antibody therapy is not administered. But this is redundant, as the monotherapy limitation is already there.
- Since the claim is restricted to monotherapy and the original application does not disclose the absence of any other therapy, just the absence of an anti-PD-1 antibody therapy, there is necessarily added subject matter.
They are not applying one set of principles to the application and another to the patent. They are noting that the claim of the patent is different from what is offered as basis for it in the application.
Having said that, this is all I have gathered from the case report and I have not delved into the case itself. The reference to the "insolubl[e] ambigu[ity]" seems on the face of it to be a red herring - what they are really saying is that the claim includes a narrow feature (monotherapy) and a broader feature (absence of anti-PD-1 antibody therapy), the latter being redundant in view of the former, but the latter being the only one that was properly disclosed in the original application.
Thank you PotP for that analysis. It prompted in me the thought of the test in the USA (Nautilus) for when a degree of ambiguity rises to a level high enough to render a claim indefinite, that test being "insolubly ambiguous". I suppose that, also in the USA, the intelligence that strives to solve the apparent ambiguity is a mind at least "willing to understand". Perhaps then, the mind that is open to persuasion is more open in the USA than the one at TBA level in the EPO, a mind not only willing to understand but also wanting to understand.
ReplyDeleteI wonder, is the EPO is on a mission to toughen up on patent drafting standards, taking advantage of every discrepancy between the A and B publications to crimp back the extent of protection which the EPO confers when granting a patent. A worthy mission, wouldn't you say?
No, I would not say it is a worthy mission - as what the EPO is doing is not looking for discrepancies between the A and B publications but instead looking for any opportunity to interpret the claims completely out of context (of the rest of the application as filed).
DeleteAs for "insolubly ambiguous", I think that would be the correct standard for the EPO to apply in those rare cases where it is simply impossible to work out what meaning was intended for the word(s) in question. However, as illustrated by the ease with which the Board interpreted the corresponding (broader) disclaimer in Claim 1 of the A publication, the problem in this case is that the negative feature in Claim 1 as granted was far from "insolubly ambiguous".
Finding ways to encourage better drafting might be a worthwhile aim. However, using inconsistent claim interpretation standards is no way to go about achieving that aim. I therefore suggest that the first order of business for the EPO ought to be ensuring that they apply consistent standards for document interpretation, not only between A and B publications but also for the claims vs. the disclosures of priority documents and (patent) prior art document. The EPO may also consider the EBA's comments in Reasons 16 of G 1/24 as providing an incentive to go even further - by ensuring at least some kind of consistency with the interpretation standards applied by the national courts and the UPC.
That doesn't sound like something that would be particularly difficult for the EPO to achieve. However, that would require the EPO to acknowledge that their document interpretation standards are not only internally inconsistent but are also out of step with the courts. I therefore will not hold my breath waiting for that to happen, as it seems far more likely that the EPO will instead find an endless stream of nonsense reasoning to revoke any patent which gives rise to inconvenient questions about claim interpretation.
I am surprised that in the Board’s decision, including its mentions of the parties’ arguments, it is never referred to para 0019 of the description of the patent as granted, which reads as follows :
ReplyDelete" Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure is related. For example, the Concise Dictionary of Biomedicine and Molecular Biology, Juo, Pei-Show, 2nd ed., 2002, CRC Press ; The Dictionary of Cell and Molecular Biology, 3rd ed., 1999, Academic Press ; and the Oxford Dictionary Of Biochemistry And Molecular Biology, Revised, 2000, Oxford University Press , provide one of skill with a general dictionary of many of the terms used in this disclosure."
This paragraph is the same as para 0022 of the original PCT application.
It is clear from para 0019 that (i) all the terms are to be understood in their ordinary meaning as evidenced by the specialised dictionaries familiar to the skilled person, (ii) except where a term is “defined otherwise”.
Such is the case for the term "anti-PD-1 antibody monotherapy" in para 0040, which provides a very narrow definition of monotherapy, limited to the exclusion of an anti-CTLA-4 antibody therapy.
The content of the application as filed, to be considered in the assessment of new matter, must be considered as a whole. This is especially relevant in this case, since it offers a list of definitions and the rule of para 0019 : the ordinary meaning does apply in general but the definitions “the term ... as used herein”, such as that of para 0040, prevail over the ordinary meaning.
This structured set of definitions is comparable to the list of definitions typically included in an agreement. Such definitions are binding on the parties. I do not see why the same principle could not apply to a patent. The discussion did not take place in the proceedings since para 0022 was utterly ignored.
Francis, the application as filed is littered with disclosures which point to "unusual" meanings for the terms "monotherapy" and "combination therapy".
DeleteFor example:
"[0006] The present disclosure provides a method for treating a melanoma comprising (i) identifying a patient having a PD-L1 positive melanoma tumor; and (ii) administering to the patient an anti-PD-1 antibody or an antigen-binding portion thereof that binds specifically to a human PD-1 ("anti-PD1 antibody monotherapy"), wherein the patient is not administered a combination of an anti-PD-1 antibody or an antigen-binding portion thereof and an anti-CTLA-4 antibody or an antigen-binding portion thereof that binds specifically to a human CTLA-4 ("combination therapy")".
See also paras [0018] and [0058]. It is a different question whether the specific (narrower) embodiment of Claim 1 as granted finds clear and unambiguous basis in the application as filed. However, there are at least several pointers to that subject matter. For example:
"[0088] In one embodiment, the anti-PD-1 antibody is nivolumab"; and
"[0106] An exemplary clinical anti-CTLA-4 antibody is the human mAb 10D1 (now known as ipilimumab and marketed as YERVOY® ) as disclosed in U.S. Patent No. 6,984,720.
If there is added matter, however, it is most certainly not for the reasons set out in the Board's decision.
I see that this case has its origins in a priority filing at the USPTO. With Rose's thoughts on "unusual definitions" at ther back of my mind, and having just read the learned contributions from PotP and F. Hagel I assume that the case was drafted in line with best US drafting practice and I take the aim of that to bestow on the prosecutor maximum flexibility to adopt, after filing, as the need arises, any one of a variety of different definitions of what is the "subject matter" sought to be protected,
ReplyDeleteIn other words, best practice involves having one's cake, and eating it too, by breeding as many Angora cats as the case will tolerate. As opposed to saying what you mean and meaning what you say.
Does this bode well, however, for prosecution at the EPO? Is this case a nice illustration of what happens when an unstoppable force (US-style patent application drafting) comes up against a steadfastly and implacably immovable object (the EPO's Gold Standard)?
Only asking.
Max, this is a case where the outcome hinges (almost) entirely upon the manner in which the claims are interpreted. Whether there is a clash between the EPO's gold standard and US-style patent drafting is rather beside the point.
DeleteCould the drafting of the specification have been improved (for prosecution at the EPO)? Certainly. Might improved drafting have eliminated the added matter issue? Very likely. However, this would only be because better drafting would have made the scope of the claim crystal-clear and basis in the description undeniable.
The primary effect of the less than perfect drafting was to cast doubt upon the meaning of the claim, and therefore to ensure that the outcome turned on how the EPO interpreted claim wording which did not have a crystal-clear meaning. Whilst better drafting would have provided a better outcome for the patent proprietor, it would have deprived the rest of us with a chance to see how the EPO deals with a claim that only makes sense when interpreted using a holistic approach. What we have learned from this case is that holistic interpretations are verboten at the EPO ... even where all other options make no sense at all.
I read all the comments with great interest and wish to comment as well.
ReplyDeleteI want to start first with claim 1 as filed. There is no doubt that bringing in the claim expressions in brackets introduces up front an ambiguity. I refer here to T 1481/05.
Like Max Drei, I am not a chemist, but to me, the board’s decision is coherent and convincing. The board looked at the claim and the original disclosure with a mind willing to understand and certainly not to misunderstand the lot.
The content of disclosure to a skilled person in the B publication is not valued higher than that of the A publication. The EPO is also not on a mission to toughen up on patent drafting standards. Whether a patent has been granted in the US is actually irrelevant for the EPO, as the legal provisions and their interpretation is different on both sides of the Atlantic.
One reproach which can be made to the board is that it referred to § of the patent and not to § of the PCT pamphlet. The content of the § quoted in the decision correspond to the content of § as filed.
The situation can be summarised as follows:
The question to be addressed was indeed to assess whether the application as filed disclosed nivolumab for use in treating PD-L1 positive melanoma as a monotherapy, in the usual meaning of the term, in the dosage regimen set out in the claim.
By excluding the combined administration of an anti-PD-1 antibody and an anti-CTLA-4 antibody, there was a first dent in the usual meaning of monotherapy. This could be acceptable at a pinch.
The second dent in the usual meaning of monotherapy is to be found in claim 49 as filed: “the method of claims 1-48 further comprising administering an anti-cancer agent”.
Claim 49 as filed makes it clear that the “monotherapy” in claim 1 as granted is definitely not a monotherapy in the usual meaning of this term.
The problem in the present case is that the applicant/ proprietor tried to have a foot in every door. With claim 49 as filed, he wanted to have the possibility to claim the administration of a further anti-cancer agent. With §43 as filed, §40 in the patent, the applicant/proprietor wanted to exclude administering another anti-cancer agent. As Max Drei said, you cannot have your cake and eat it.
At the EPO, the basis for assessing added matter is the originally filed application as a whole, here the PCT pamphlet. The applicant/proprietor is actually paying its shoddy original drafting. A holistic claim interpretation cannot compensate for such a shoddy drafting.
By the way, G 1/24 has not introduced at the EPO a holistic interpretation of the claims. The only decision which favours a holistic approach is T 439/22-2 (gathered sheet). All other decisions applying G 1/24 have not adopted this position.
As far as G 1/24 and added matter are concerned, there are plenty of decisions in which the boards have applied G 1/24 when it comes to added matter. I refer here to T 2488/22, T 0981/23, T 0873/23, T 1164/23 and T 0405/24.
In the latter decision, the board added that “there is no indication in G 1/24 that "consulting" or "referring to" the description and drawings could translate to adopting a claim interpretation which ensures that the disputed feature is originally disclosed and thus necessarily complies with Article 123(2) EPC. Such an approach which inherently assumes that there may be only one "correct" interpretation of a claim feature, namely the one derivable from the original description as its intended meaning, would not lead to an objective assessment of compliance with Article 123(2) EPC and thus jeopardise legal certainty. It would be tantamount to interpreting a claim feature such that, in the end, virtually no violation of Article 123(2) EPC within the meaning of the well-established "gold standard" could arise”.In view of the above decisions, the admissibility of the referral G 1/26 is thus questionable, and especially that of question 1.
@Proof of the pudding & MaxDrei
ReplyDeleteProof, thanks for your response. I simply wanted to point out that the description was to be consulted as a whole and para 0040 was not be considered in isolation but in combination with the entire set of definitions beginning with para 0019 (which to my surprise was not mentioned in the decision). Your comment implies that the definitions must be consistent.
Concerning the US drafting practice, the principle is laid out in MPEP 2111.01.III : “The presumption that a term is given its ordinary and customary meaning may be rebutted by the applicant by clearly setting forth a different definition of the term in the specification.”
Under US law, the applicant is entitled to act as its own lexicographer (in EPO parlance, the patent may be its own dictionary) but subject to clearly showing its intent to impart to a claimed term a meaning different from the ordinary meaning. The claimed term must be defined in the description by way of a “definitional statement”, which calls for a number of formal conditions (term between hyphens, use of “ as used herein” language, etc).
Such a definition is binding on the applicant/patentee : see Alnylam v Moderna, CAFC 4 June 2025, comments by Rose Hughes in IPKat 10 June 2025. It is taken seriously, leaving no room for playing angora cat.
Mr Hagel,
DeleteThanks for your explanations about US practice. In Europe, the notion of the patent being its own dictionary has been established by the German Federal Court (BGH).
G 1/24 cannot be said to have introduced this notion in European practice. I would rather say, the contrary is true in view of the case law following G 1/24. Up to now, all decisions having applied G 1/24, but T 439/22-2 (gathered sheet), have not gone along this line.
As a pure matter of fact, and without any judgement of value, it is interesting to note that the LQM which was at the origin of G 1/24 is also at the origin of G 1/26. I am not persuaded that following G 1/24, the EBA will be minded to reopen the discussion on the result of the “consultation” of the description. This is the more so since there are already a series of decisions linking G 1/24 with added matter.
Mr Thomas,
DeleteYour understanding of the EPO case law fittingly suits your personal disagreement with the patent being « its own dictionary » but it is at loggerheads with the EPO’s official presentation of its case law, see the CLBA 2026 edition, I.C.4.1 which explicitly endorses this principle of interpretation. And this is nothing new, it has been laid out with the same words since the 2013 edition, in which the relevant section was numbered I.C.3.1.
The qualification of « settled cas law » for this principle is thus fully justified. It owes nothing to the influence of the LQM you refer to, If I am correct in assuming that this LQM was formely a German judge : he joined the EPO well after 2013. There may have been as you suggest an influence of German patent law at the outset, but why should this be a relevant factor today ? The EPO has drawn on the practice of other member states. The « Swiss claims » relating to second medical use is an example.
G 1/24’s requirement to always consult the description is fully consistent with this principle since it implies that definitions of claimed terms in the description different from the ordinary meaning of the terms must be paid attention, but keeping in mind that as pointed out in the CLBA, the description has to be considered as a whole. My view is that such definitions can only be allowed to prevail over the ordinary meaning of claimed terms subject to very restrictive conditions, and I concur here with the conditions set out in US case law. In any event, what G 1/24 clearly rejects is a method of interpretation which looks at the claims alone.
Mr Hagel, my view is not just that the description should be consulted as a whole, but that the method used for interpreting the claims should be consistent with the method used for interpreting the disclosure of the application as filed.
DeleteThe Board's decision in T 0715/24 is a perfect illustration of how the use of different interpretative methods (claims vs. original disclosure) leads to what it essentially the same language being interpreted to mean different things.
The negative limitation in Claim 1 as granted is not disclosed as a stand-alone disclaimer in the application as filed. It instead derives from the negative limitation in Claim 1 as filed, by replacement of the generic definitions of the active agents with the specific definitions of paras [0088] and [0106]. Thus, if Claim 1 of the application as filed is interpreted such that the disclaimer makes technical sense - due to an unusual meaning afforded to the term "monotherapy" - logic dictates that the mere narrowing of that disclaimer (to include preferred definitions of generic terms) should not lead to any different interpretation or technical understanding.
The Board's logic for interpreting "monotherapy" differently in Claim 1 as originally filed rests solely upon the use of quotation marks and parentheses (to suggest a different interpretation that is then further supported by a contrast to "combination therapy" mentioned in original Claim 1, and by the subject matter of original Claim 49). But this simply means that the Board has turned a blind eye to the fact that the unusual definition for "monotherapy" is also suggested by the presence of a disclaimer which makes no sense if a standard definition is applied.
Essentially, this represents a method by which a claim is interpreted entirely out of context (of the rest of the specification) and with a mind aiming to misunderstand. This represents the complete opposite of the (whole document) interpretative method applied to the application as filed.
Mr Hagel-16.08.2026 at 14:27:00 GMT+1,
DeleteReply part 1
I explained to you before that the “White book” (CLBA) is issued by a department part of the boards of appeal, but its content is not vetted by the boards, and that it is always better to check the actual decision. I.C.4.1 has not been amended in the 2026 edition.
The passage you quote, relates, not to claim interpretation as such, when interpreting a claim under Art 52-57, i.e. under G 1/24, but within the framework of novelty with respect to a piece of prior art. I.C.4.1,deals with “General rules of interpretation of prior art documents”. In this part one indeed finds the following sentence: “The patent document might be its own dictionary (T 523/00, T 311/93)”. I note that all the decisions quoted in I.C.4.1 are decisions which predate G 1/24 by many years, and their relevance is thus rather not immediately apparent
.
It is correct that a claim in a piece of prior art might say something different in the description.
If, in a piece of prior art, the wording of a claim as such, allows a broad interpretation, would you as opponent, limit the teaching of this prior art to the limiting interpretation found in the description? Conversely, if, in piece of prior art, the wording of a claim as such, allows a limiting interpretation, would you as opponent, limit the teaching of this document to a broadening interpretation found in the description? Let’s face reality, you would not, as you would deprive yourself from a very relevant piece of prior art.
.
G 1/24 has never said that the result of the required “consultation” of the description is that the patent is its own dictionary. If this had been the intention of the EBA, it would have said so. Anything else is pure guess work.
Your quotation of the CLBA is thus as biased as your allegation that my understanding of the EPO case law fittingly suits my personal disagreement with the patent being its own dictionary.
I do therefore maintain that G 1/24 has not introduced the notion of the patent being its own dictionary, and your quotation of I.C.4.1 is not representative of its whole content.
Mr Hagel-16.08.2026 at 14:27:00 GMT+1,
DeleteReply part 1
I explained to you before that the “White book” (CLBA) is issued by a department part of the boards of appeal, but its content is not vetted by the boards, and that it is always better to check the actual decision. I.C.4.1 has not been amended in the 2026 edition.
The passage you quote, relates, not to claim interpretation as such, when interpreting a claim under Art 52-57, i.e. under G 1/24, but within the framework of novelty with respect to a piece of prior art. I.C.4.1,deals with “General rules of interpretation of prior art documents”. In this part one indeed finds the following sentence: “The patent document might be its own dictionary (T 523/00, T 311/93)”. I note that all the decisions quoted in I.C.4.1 are decisions which predate G 1/24 by many years, and their relevance is thus rather not immediately apparent
.
It is correct that a claim in a piece of prior art might say something different in the description.
If, in a piece of prior art, the wording of a claim as such, allows a broad interpretation, would you as opponent, limit the teaching of this prior art to the limiting interpretation found in the description? Conversely, if, in piece of prior art, the wording of a claim as such, allows a limiting interpretation, would you as opponent, limit the teaching of this document to a broadening interpretation found in the description? Let’s face reality, you would not, as you would deprive yourself from a very relevant piece of prior art.
.
G 1/24 has never said that the result of the required “consultation” of the description is that the patent is its own dictionary. If this had been the intention of the EBA, it would have said so. Anything else is pure guess work.
Your quotation of the CLBA is thus as biased as your allegation that my understanding of the EPO case law fittingly suits my personal disagreement with the patent being its own dictionary.
I do therefore maintain that G 1/24 has not introduced the notion of the patent being its own dictionary, and your quotation of I.C.4.1 is not representative of its whole content.
Mr Hagel-16.08.2026 at 14:27:00 GMT+1,
DeleteReply part 2
I understand “consulting” the description in the meaning of G 1/24, as I have explained above for a piece of prior art. Why should the interpretation differ? We have to remain coherent.
It is however a matter of fact, that all decisions applying G 1/24, but T 439/22-2, go along the lines explained above, and do not consider that the patent is its own dictionary. By today, there are 137 decisions applying G 1/24. T 439/22-2 is the only one considering that the patent is its own dictionary. If this not an “established” line of case, what is then to understand under this definition?
A broad claim cannot be interpreted in a limited way because the applicant/proprietor has hidden in the description some limiting interpretation. Conversely, a limited claim cannot be used to fence off broad piece of prior art, when at the same time, the applicant/proprietor has hidden in the description a broader interpretation of the limited claim. This is what the case law presently says. My view that the patent is not its own dictionary is thus in no way at loggerheads with the EPO’s official presentation of its case law, and certainly not the one following G 1/24 besides one exception.
I.C.4.1, can be resumed as follows: the skilled person reading a document does not construe individual points of the disclosure in isolation, but rather in the context of the document as a whole, cf. T 546/07, T 312/94 or, T 860/06. In other words, when determining what is made available to the public within the meaning of Art. 54(2) by a prior art patent document, it must be borne in mind that it is the description which chiefly serves to disclose the invention in a manner that it may be carried out, whereas the chief function of the claims is to define the subject-matter for which protection is sought, cf. T 1658/12. If you quote the White book, please quote it correctly and do not pick out a sentence of its context.
You and some others representatives, as well as the LQM in T 439/22-2, would have liked the EBA to have said that the patent is its own dictionary. If claims have a minimum of raison d’être, it is difficult to follow this position. At least there is then an inconsistency between claim and description if the matter for which protection is sought can be interpreted differently when looking at the description or at the claim. As the EBA said in G 1/24, the claim has then to be amended, cf. G 1/24, Reasons 20.
By the way, I know personally the LQM in question and have a good relationship with him. We might have divergent views on this topic, but this happens in mutual respect.
When you come up with Swiss-type claims, you are actually comparing apples and pears. Swiss type claims had been designed as a legal trick in order to allow the patenting of second/further medical indications. Since the EPC 2000 this trick is not any longer necessary, and even forbidden. This is to my knowledge, the only time the EPO has adopted a decision from a national judiciary. Do you know another case? I do not.
Some national judiciary claim following the case law of the EBA, but it actually does not, e.g. the German Federal Court (BGH) when it comes to added matter. All the legal tricks invented by the BGH to overcome added matter have been rejected by the boards of appeal of the EPO.
Mr Thomas -
ReplyDeleteLet's assume that (as you claim - I've not had time to check) T 439/22 is the outlier in an otherwise wholly consistent line of case law applying G 1/24.
Do you not then see it as a matter for great concern that the Board which referred the questions to the EBA in G 1/24, in the very case which led to G 1/24, has come to such a radically different interpretation of G 1/24 than all the other Boards?
Doesn't this illustrate that G 1/24 failed to set a clear and unambiguous test, concluding the matter once and for all?
At the very least, isn't it ironic?
@ Anon Y. Mouse-17.08.2026 at 21.08.00 GMT+1
DeleteI can assure you that T 439/22-2 is the outlier in an otherwise wholly consistent line of case law applying G 1/24.
For my blog, I am looking at all decisions published on the “Recent Decisions” page of the boards. Should I have found one which would have agreed with T 439/22-2, I would have published it immediately, be it only in due respect of the readership.
I would like to observe that in G 1/24, the EBA refused to admit Question 3, which reads: “May a definition or similar information on a term used in the claims which is explicitly given in the description be disregarded when interpreting the claims to assess patentability and, if so, under what conditions?”
If the EBA would have answered this question, it could have answered by saying that the patent is or is not its own dictionary. The fact is that the EBA did not give an answer about the result of the consultation. The EBA must have had good reasons for this when in the same breath it referred to the UPC decision NanoString Technologies -v- 10x Genomics. The EBA has only decided that the description has always to be consulted, nothing more.
Lookin at the description is certainly a kind of holistic approach, but not the holistic approach as defined in T 439/22-2 which has clearly considered that the patent is its own dictionary. This is not what the rest of the boards actually do.
As all other boards with the exception of that acting in T 439/22-2 have come to a logical and coherent conclusion, excluding that the patent is its own dictionary, G 1/24 has thus by no means failed to set a clear and unambiguous test.
I consider T 0873/24, not as an outlier, but as a desperate attempt to reopen the discussion on what should be the result of the consultation required in G 1/24. It is indeed ironical that the loud and clear message of the other boards has not come to the attention of the LQM at the origin of G 1/24 and G 1/26.
In T 0873/24, the board found three approaches when consulting the description:
- First approach (Reasons 2.1): consulting the description only to define the skilled person.
- Second approach (Reasons 2.2): no broadening or limitation of claims based on the patent specification. This is the bulk of decisions applying G 1/24 but not as in T 439/22-2.
- Third approach (Reasons 2.3): holistic approach permitting broadening and/or narrowing of the interpretation in view of the patent specification. In this part, among other decisions which simply refer to the general requirement of G 1/24, that the claims have to be looked at in the light of the description, the board heavily quotes T 439/22-2. What a surprise.
To sum it up, the boards, besides those having decided T 439/22-2 and T 0873/24 have clearly understood how G 1/24 has to be applied. It is thus incorrect to allege the G 1/24 has failed to set a clear and unambiguous test, concluding the matter once and for all. The contrary is true.
For the sake of completeness, the following represents the full (first) paragraph of CLBA I.C.4.1 to which which Mr Hagel referred:
ReplyDelete"According to the boards' settled case law, the technical disclosure in a prior art document must be considered as a whole (see T 56/87, OJ 1990, 188). The individual sections of a document cannot be considered in isolation from the others but must be seen in their overall context (T 312/94, T 452/05, T 456/10). In T 1321/04 the board found that the terms used in patent documents should be given their normal meaning in the relevant art, unless the description gave the terms a special meaning. The patent document might be its own dictionary (T 523/00, T 311/93). Thus, if a special meaning could be derived from the patent document, only this meaning was ultimately decisive".
From the last sub-clause of the last sentence, it is beyond me why anyone would think that an opponent would be free to pick and choose either a broad or a narrow definition of subject matter which the case law clearly indicates has only one meaning.
The concept of subject matter having a single meaning also follows from CLBA I.C.4:
"The consistent view in the case law is that for an invention to lack novelty, its subject-matter must be clearly and directly derivable from the prior art (see e.g. T 465/92, OJ 1996, 32; T 511/92) and all its features – not just the essential ones – must be known from the prior art (T 411/98)".
This is because Reasons 2.2.2 of G 1/03 states that:
"the European Patent System must be consistent and the concept of disclosure must be the same for the purposes of Articles 54, 87 and 123 EPC".
Applying the same concept of disclosure for both of Articles 54 and 123(2) EPC means that the "clearly and directly derivable" standard mentioned in CLBA I.C.4.1 is the same as the "clearly and unambiguously derivable" standard for added matter. A meaning will only be unambiguously derivable if it is a single meaning. Where there is more than one (equally) plausible alternative meaning, then there is no unambiguous disclosure.
Now, if only the EPO applied the same standards to claim interpretation.
But wait. Perhaps "directly" is not the same as "unambiguously". I mean, when it comes to Art 123(2), Art 123(3) and own prior use evidence, the standard to attain is "beyond reasonable doubt" whereas to prove other facts the normal civil "more likely than not" standard applies. Perhaps the skilled person, wanting to derive meaning, can clearly and directly (but not without ambiguity) derive more than one meaning from a document, for example when it is insolubly ambiguous. As to claim interpretation, I'm not sure that decisions about whether any given subject matter is or is not "disclosed" are relevant to the different task, of construing a claim.
DeleteWith respect to a published patent document, determining what is disclosed requires determining the meaning of the words in that document. Determining meaning is interpretation. Thus, one interprets the words of a published patent document to determine the subject matter that it discloses.
DeleteIndeed, CLBA I.C.4.1 makes this point explicit, by stating that "Determining the information content means interpreting what comprises the state of the art".
Whether for novelty, added matter or priority, interpreting the disclosure of a document starts with a mind willing to understand and proceeds by considering both context (provided by the specification as a whole) and the cgk (of the skilled person).
Why should the method of interpretation be any different simply because one is looking at words in the claims of a patent or application that is the subject of proceedings before the EPO (as opposed to words in any part of any other document, including the claims, that is relevant to but not the subject of those proceedings)?
If any interpretation concludes with a finding that the words in question have no clear meaning, this points to insoluble ambiguity. That ambiguity remains the same regardless of the status of the document in the proceedings before the EPO, eg whether it is the patent or application under examination or it is a prior art document cited against that patent or application.
Insoluble ambiguity in a prior art document leads to a lack of clear and unambiguous disclosure. In a patent or application, such insoluble ambiguity instead points to invalidity (eg due to non-compliance with Article 83 or 84 EPC). These are two sides of the same coin.
As for deriving more than one meaning from an insolubly ambiguous claim, what you really mean is deriving more than one possible meaning. However, patentability is, or at least should be, assessed based upon actual meanings, and not just hypothetical ones.
@ Proof of the pudding-18.08.2026 at 13.02.00 GMT+1
ReplyDeleteReply part 1
You seem to have overlooked the explanations I have been giving to Mr Hagel about the way the CLBA is drafted. The CLBA is actually drafted and issued by a department within the boards of appeal, but is not vetted by the boards. The comments in the CLBA are the opinions of the various members of this department and not that of the boards.
The quotation “Thus, if a special meaning could be derived from the patent document, only this meaning was ultimately decisive", is not the opinion of the boards and any conclusions drawn are not to be taken as exemplifying the board’s position.
It is your right to disagree, but I maintain that as opponent it would be foolish not to consider the broad interpretation found in the description if the claim is limited, or to ignore the broad claim when the description contains a limiting interpretation of a claimed feature. In as far as the description and the claims both represent reasonable and enabled teachings, the skilled person cannot ignore them.
Any clear teaching/information found in a document cannot be ignored. Let’s take the example of NPL. If the abstract and the document as such disclose each a reasonable and enabled technical teaching, the skilled person cannot ignore either, even if the technical teaching of the abstract and that of the document are different or even contradictory. This is applicable as well in case when different interpretations are possible when comparing claims and description.
I cannot but fully agree with Reasons 2.2.2 of G 1/03. When deciding that in a piece of prior art, any teaching derivable from it, "the European Patent System must be consistent and the concept of disclosure must be the same for the purposes of Articles 54, 87 and 123 EPC".
Your reading of Reasons 2.2.2 of G 1/03 appears as biased than that of Mr Hagel when quoting the CLBA I.C.4.1. It is interesting to note that Reasons 2.2.2 of G 1/03 deals with accidental anticipations when it comes to disclaimers. The sentence following the one you quoted reads: “However, this does not prejudice the question of what is to be considered as technical information disclosing the invention.” Further down one reads: “a mere disclaimer excluding the subject-matter of an accidental anticipation may be assumed not to change the technical information in the application as filed and, for this reason, also not to change the subject-matter of the application as filed, within the meaning of Article 123(2) EPC”.
It thus does not mean that a prior art document has to be read selectively by the skilled person. If a broadening or limiting interpretation has been introduced after filing, this new interpretation changes the subject-matter of the application as filed. When the broadening or limiting interpretation has been present ab initio, see Part 2.
@ Proof of the pudding-18.08.2026 at 13.02.00 GMT+1
ReplyDeleteReply part 2
What matters is thus the technical information which is directly and unambiguously derivable from a document, be it a piece of prior art or be it a patent application or a patent. When a claim gives to the skilled person, one possible technical information/teaching, and the description another one, more limited or broader, it is not possible to ignore this fact.
Your view, as well as that of Anonymous, Mr Hagel, or plenty of others, is that the patent has to be its own dictionary. This boils down to ignore this fact and to read into the claim, the broadening or limiting teaching/information found in the description. This is also the position adopted in T 439/22-2. This is clearly not what the EBA has decided in G 1/24.
The consequence of not reading in a claim a broadening or limiting teaching/information found in the description, it is that the application/patent as a whole, manifestly discloses two possible technically reasonable, but equally valid teachings/information to be given to the technical teaching/information directly and unambiguously derivable from the claims and the description.
In this situation, the matter for which protection is sought is then not properly defined, and, as required in G 1/24, Reasons 20, an objection under Art 84 has to be raised during examination. This exactly what you say in your last but one paragraph.
It is only by entertaining the fiction that patent has to be its own dictionary that you come to the conclusion that there can only be one technical teaching/information directly and unambiguously derivable from a claim. This is not what has been decided in G 1/24, and the interpretation given to G 1/24 by all boards applying it, but T 439/22-2.
In your last entry of 19.08.2026 at 18.13.00 GMT+1 you say that “patentability is, or at least should be, assessed based upon actual meanings, and not just hypothetical ones”. I can agree with this statement, however with a massive proviso. If the claim and the description allow to attribute meanings different from one another, one cannot be held to be the true one and the other cannot be held to be a merely hypothetical one. There are the two faces of the same coin.
This ambiguity has to be resolved by raising an objection under Art 84 during examination. As in opposition, Art 84 is not applicable to what has been granted, the proprietor will have to accept to be confronted with two approaches when it comes to decide upon patentability. A broad claim cannot be defended by referring to limiting statements in the description, and a limited claim cannot fence off broad prior art, when the description is allowing a broad interpretation of the claim.
I.C.4.1 might say that "Determining the information content means interpreting what comprises the state of the art". However, the information content of a prior art patent document might allow two possible interpretations, one derivable from the claim and one from the description. The same applies to NPL and an abstract. Why would the skilled person take one and disregard the other one? There is no legal basis in the EPC for this distinction, and certainly not in Art 54(2). .
The skilled person interprets the information content of a prior art document as a whole and derives from it any possible reasonable and enabling teachings. In order to remain coherent, this applies as well to a European patent application or a European patent. Interpreting a document as a whole does not mean selecting a specific interpretation, but accepting any teaching/information disclosed in one and the same document. This is the plain and immediate result of the “consultation” required in G 1/24.
OK, so you at least believe that the EPO should apply consistent methods of interpretation. Where we appear to disagree is the question of whether the same words in the same document should, unless the context demands otherwise, have the same meaning.
DeleteMy view is that the document should be interpreted as a whole, and that location of the words in question (whether description or claims) does not change their meaning. Whilst you disagree, I would point out that my view is completely in alignment with the case law cited in the relevant section of CLBA I.C.4.1, in particular the following from T 1321/04:
"2. Interpretation of claim 1
2.1 Regarding the different interpretations of the term "predetermined value", it is necessary to establish which meaning the person skilled in the art attributes to the wording "predetermined value of nitrogen oxide concentration" in claim 1.
2.2 Terms used in patent documents should be given their normal meaning in the relevant art, unless the description gives the terms a special meaning. The patent document may be its own dictionary (T 523/00, T 311/93, neither published in the OJ EPO).
Thus, if a special meaning can be derived from the patent document, only this meaning is ultimately decisive.
2.3 For understanding the meaning of the terms used in a patent document, the person skilled in the art does not consider the terms in isolation from the remainder of the document, ie with their literal meaning. On the contrary, the terms are considered in the context of the contents of the document as a whole (T 312/94, T 969/92, neither published in the OJ EPO).
Therefore, terms must be construed as they would be by the person skilled in the art according to the whole content of the application, taking into account what is achieved by the invention.
These findings are in line with the principle laid down by the boards of appeal that the description and the drawings are used to interpret a claim when an objective assessment of its content has to be made (see "Case Law of the Boards of Appeal of the European Patent Office", 4th edition, II.B.4.3, 2nd paragraph)".
See also Reasons 1 of T 523/00, which includes the following:
"The documents of a patent are to be considered as a self-contained whole, giving the words used their normal meaning in the art, unless in particular cases the description gives the words a special meaning. There is nothing in the EPC, which requires that terms used in different patents should have the same meaning. According to Article 69 EPC, the description and drawings of the patent, and not of other patents, shall be used to interpret the claims, that is to say the patent itself is its own dictionary. In the present case, the meaning of the term "fiber" is given in the description of the patent in suit, page 2, lines 55 and 56, together with the passage on page 3, lines 42 to 45, when yarns are implied. It follows that a fibre in the meaning of the patent in suit is the individual smallest element".
Am I to presume that, despite cases such as T 1321/04 and T 523/00 clearly representing “established” case law, you do not agree with the (claim) interpretation principles that they espouse?
One more point: if one interprets the same words in the same document to have multiple meanings (eg one meaning in the description as filed and a different meaning the claims as filed), can it really be said that either of those meanings is clearly and unambiguously disclosed?
DeleteWould not the existence of two (possible) meanings cast doubt upon which meaning is disclosed in any one instance? Which rule of law or logic could be relied upon to clearly and unambiguously determine which meaning is disclosed for each occurrence of the same wording?
@ Proof of the pudding-20.2026 at 15.27.00 GMT+1 and 16.40.00 GMT+1
DeleteAt least you have consulted the actual decisions and not merely referred to the CLBA.
My view is also that the document should be interpreted as a whole, but that location of the words in question (whether description or claims) DOES and necessarily CHANGE their meaning.
The two decisions you quote, T 1321/04 and T 523/00 largely predate G 1/24 and two single decisions cannot, with due respect, represent an “established” line of case law. In view of G 1/24, I therefore fully disagree with the claim interpretation principles that they espouse.
I can however agree with the statement in T 523/00 that according to Art 69, the description and drawings of the patent, and not of other patents, shall be used to interpret the claims, but I squarely disagree that this is to say that the patent itself is its own dictionary. This is not what G 1/24 has said. In G 1/24, the EBA has also made clear that Art 69 is primarily concerned with infringement.
One meaning/interpretation can be given to features in the description, and one meaning/interpretation can be given to features in the claims. The meaning given by the skilled person of a feature in the description can be broader or more limited than the meaning given by the skilled person to a feature in the claim and vice versa. The condition is that both meanings/interpretations make technical sense for the skilled person.
In this situation, I cannot find in the EPC any legal requirement that one meaning should prevail over the other. This is your theory and that brought forward in T 439/22-2, but represents a very peculiar way of applying G 1/24.
This position is at loggerheads with the line of case law of the boards applying G 1/24, but T 439/22-2. All those decisions (>135), but T 439/22-2 represent manifestly a new established line of case law.
You may have your theory, but it is by no means compelling for the boards, besides T 439/22-2. There are presently many more decisions saying that, when the meaning in the description is limited with respect to the broader meaning in the claim, broad prior art cannot be fenced off in view of the description. Conversely, when the claim is limiting, a broader definition in the description cannot be ignored and in spite of the limited claim, broad prior art can be opposed to the claim. Even T 439/22-2 came to this conclusion.
You did not reply to my comment that when the description and the claims allow different interpretations/meanings, the matter for which protection is sought, which should be defined in the claim, cannot be properly assessed and an objection under Art 84 has to be raised. This is an important consequence also required in G 1/24.
The same words in the description and in the claim cannot have multiple meanings. However, when different words or a different explanation of a claimed feature is given in the description and in the claims, then multiple meanings can easily be determined by the skilled person. It can thus well happen that e.g. one meaning can be given in the description as filed and a different meaning can be given to the claims as filed.
As such, both meanings, in the description and in the claims as filed, can be said directly and unambiguously derivable by the skilled person from the original filing, but they do not tie up. Before going any further in examination, the applicant/proprietor has to be invited to decide which meaning it prefers, and bring this in the claim, if primacy of the claim should retain any meaning. The applicant/proprietor cannot be left to decide which meaning should prevail depending on the circumstances. This is manifestly detrimental to the legal safety of third parties.
In T 0741/25, published today, when applying G 1/24 and consulting the description, the board concluded that what was defined in the description and in the claim as “micro-lens”, was actually anything but a “micro-lens”. G 1/24 has not finished to surprise us.
I apologise to the readers for any repeat of comments. When attempting to post a comment, I often receive the message “An error has occurred”. I then continue to post until I get the message “Your comment will be visible after approval”. I consent to the deletion of any double publication of a comment by the IP Kat’s webmaster. In another thread, I was not able at all to publish any comment.
ReplyDeleteI am beginning to think that the demise of the principle that what comes after and infringes will, if coming before, destroy novelty has a lot to answer for. Where we are now, within any given "disclosure" two different but equally legitimate claim "meanings" can be held to exist, side by side. Having to choose just one of them is perhaps a thing of the past. What say you?
ReplyDeleteAs someone who does not subscribe to the current fashion in (claim) interpretation principles at the EPO, I very much doubt that a skilled person would believe that the subject matter of a claim or disclosure resides in a quantum-like state of uncertainty. Instead, I believe that they would just get on with the task of determining what the claim or disclosure actually means.
DeleteReally, this seems to be a case of the EPO adopting principles that are suspiciously convenient for them: as if one can interpret any disclosure as if entirely isolated from the context of the document in which it appears, there will be no need for the EPO to take the time to read the entire specification carefully and consider the potentially thorny issue of how the context might affect meaning.
Regardless, it is not at all clear to me how this change in approach has come about. That is, I do not know why the EPO is now convinced that the Board in T 523/00 was completely wrong when it pronounced that "There is nothing in the EPC, which requires that terms used in different patents should have the same meaning". The relevant provisions of the law are the same, so what could have possibly have changed to make this statement any less true than when it was first written?
@ Max Drei-20.08.2026 at 14.15.00 GMT+1
DeleteWhen, within any given "disclosure" two different but equally legitimate claim "meanings" can be held to exist, side by side, it cannot be left to the applicant/proprietor to decide, depending on the circumstances, which one should prevail. The applicant/proprietor has to decide which one should prevail, stick to it and amend the claim or the description accordingly. This is easily possible in examination, but in opposition, care should be taken that the amendment is not infringing Art 123(3).
@ Proof of the pudding-21.08.2026 at 09.39.00 GMT+1
DeleteWhether you do not subscribe to the current fashion in (claim) interpretation principles at the EPO is actually irrelevant, as it is not for you to decide, since at the end of the day it is for the boards of appeal to decide.
When, after consulting the description, the skilled person comes to the conclusion that the meaning/interpretation derivable from the description is different from the meaning/interpretation derivable from the claim, the skilled parson has no reason to believe that the subject matter of a claim or disclosure resides in a quantum-like state of uncertainty.
In this situation, there are two technical meaningful interpretations possible, and it is for the applicant/proprietor to decide which one should prevail. He cannot have both at the same time. Furthermore, it is not for the skilled person, or the deciding body, to guess what the claim or disclosure actually means in presence of two technically reasonable meanings.
To come to the above conclusion, it is absolutely necessary to “consult” the entire description when interpreting the claims. G 1/24 requires the EPO to take the time to read the entire specification carefully and consider the potentially thorny issue of how the context might affect meaning. In case of any doubts about the definition of the matter for which protection is sought, which has be present in the claim and not in the description, the EPO will ask for clarification.
As explained before, T 523/00 has been superseded by G 1/24 and the decisions applying it, besides T 439/22-2. I can agree with T 523/00 that there is indeed nothing in the EPC, which requires that terms used in different patents should have the same meaning.
Actually is it extremely rare that the same terminology is found in a claim and in a piece of prior art. If the same terminology is found in the prior art and in the claim, it is sufficient to give the corresponding bibliographic data in the prior art to show identity of features. If the terminology is different in the prior art and in the claim, it has to be demonstrated why for the skilled person, there is identity of features.
If the claims says “tubular film” and the prior art “liner tube”, the two features are, a priori, not identical. If they have to be considered identical, some reasoning has to be brought forward, why for the skilled person, the “tubular film” and the “liner tube” are identical. This necessitates to consider the description of the patent at stake and to compare the technical information therein with the technical information in the prior art. This is exactly what G 1/24 requires.
The statement in T 523/00 is still true, but it does not allow to conclude that a patent is its own dictionary. Coherence has to apply when it comes to analysing the technical disclosure in a piece of prior art, when comparing this disclosure with that of the prior art. In NPL it is not possible to decide on the basis of claims vs disclosure, as NPL does not contain any claims, but a best an abstract.
Easy to amend? Not sure that is so, when the application as filed contains an ambiguity and the purpose of the amendment is to remove it.
ReplyDeleteThe original "disclosure" is that which is unambiguously derivable from the aplication as filed, and nothing more than that. If the result of the amendment of something ambiguous is to render it unambiguous then the amendment has added something to the "disclosure", that is to say, the amendment has added subject matter to the application as filed.
If there is a fault in my logic, please appraise me of it.
But, meanwhile, I'm thinking that it is no wonder that Gold Standard and "beyond any doubt" EPO practice on 123(2) is from time to time assessed as horribly strict.
@ Max Drei, 22.08.2026 at 23.00.00 GMT+1
DeleteNot every ambiguity is leading to added matter.
If, for example the clarification of a feature leads to adding “internal” to the position of a specific device with respect of a whole construction, it can lead to added matter as the specific device can also be “external”. I think here for example to a pitch modifying device in a wind turbine blade. Most of the pitch varying devices are internal, i.e. within the blade, but there are some which are external.
When the description allows the skilled person a different interpretation/meaning with respect to the interpretation/meaning derivable by the skilled person from the claim, there is an ambiguity as the matter for which protection is thought is not properly defined.
It is then for the applicant to make a choice not infringing Art 123(2). In opposition it is for the proprietor to make a choice not infringing Art 123(2) and Art 123(3). For the safety of third parties, neither the applicant nor the proprietor can be left to decide what they prefer depending on the circumstances.
A limiting interpretation in the description does not allow the applicant/proprietor to fence off prior art relevant for the broad claim. Conversely, in presence of a limited claim the presence of a broadening interpretation/meaning in the description cannot be ignored, and the limited claim can be attacked on the basis of a broad document. The latter case is typical for the well known Angora cat. A limited claim when it comes to validity, but a broad claim when it comes to infringement. Do you think this is acceptable? I do not. It is the duty of the EPO to nip such attempts in the bud.
Deciding which interpretation should prevail, i.e. description or claim, has, a priori, absolutely nothing to do with Art 123(2) but with Art 84.
The EPO is often criticised for its strict stance on added matter, whether expressed by the gold standard, “directly and unambiguously derivable…”, or “beyond any reasonable doubt”. There is a third criterion “the original disclosure is not a reservoir of features combinable at will”. Those three criteria represent the established case law of the boards, and it is good so.
By now applicants/proprietors know where they stand. What is gained on one side is lost on the other. You cannot have your cake and eat it. In other words, added matter and novelty are the two sides of the same coin.
With DXT's help I think I have found the flaw in my logic, as follows.
ReplyDeleteA term in a claim (for example, "gathered") can be ambiguous if from the description it has two distinctly different meanings. Both what if we can agree, that each of these two meanings is disclosed, in itself, directly and unambiguously. Is that imaginable?
If so, and if I then excise one (only one) of those clear and unambiguous meanings from the description, and leave the other undisturbed, the ambiguity is cured, without adding to the disclosure and without adding any matter.
Anon, the hypothetical scenario that you envisage is where a term has two possible meanings but it is not possible for the skilled person to determine which of those meanings is correct.
DeleteDoes not that make the meaning of that term, by definition, ambiguous? It would certainly fit the Wikipedia definition:
"Ambiguity is a state in which the meaning of a phrase, statement, situation, or resolution is not explicitly defined, making for several plausible interpretations. It arises when available information lacks sufficient context or a shared frame, so people cannot reliably determine what the problem is, what matters, what causes what, or what solution would count as correct".
The trouble with interpreting the claims out of context (of the description) is that doing so often gives rise to situations where the meaning of a term is ambiguous. Asserting that a term in a claim has an acontextual meaning that is clearly and unambiguously disclosed in the application as filed therefore becomes a stretch. This is because, to do so, it is necessary to combine the disclosures of the claims of the application as filed with a selected definition from the prior art / cgk, whilst at the same time ignoring relevant context and/or definitions provided in (the description of) the application itself.
In other words, it makes the unjustifiable assumptions that the skilled person would not read the description together with the claims, and would believe that meanings derived from sources external to the application clearly and unambiguously prevail over any different meaning derived from the application itself.
Also, even if we hold our noses and ignore the stink emanating from those assumptions, there is no avoiding further, thorny issues raised by those assumptions: such as what would the skilled person do if there is no commonly accepted meaning for a specific term in the context of the field of the invention? Also, if moving a term to the claims changes its meaning (relative to the description), is it ever possible to avoid adding subject matter when claims are amended to recite an embodiment disclosed solely in the description of the application as filed? In particular, is it possible to do this where the meaning of the term in the description is only implicit (from the context)?