You did not invent what you did not disclose

How to read the ledger below. Each numbered point carries the kind of evidence behind it. ESTABLISHED is stated in the authority's own text, with a locator. IMPLEMENTED is the rule as the authority actually applied it. OBSERVED is carried by a source outside this one. ASSUMED is the drafter's inference, flagged as such and not the authority speaking. Quotations are checked against a cached copy of the primary source by script.

In one paragraph

The Federal Circuit held in Enanta Pharmaceuticals v. Pfizer on 23 June 2026 that a provisional application disclosing a substituent as C2 to C12 alkyl does not provide written description support for the C1 alkyl disclosed in the later patent, even though the applicant said the C2 was an obvious typographical error that should have read C1, because applicants did not invent what they did not disclose. The patent therefore lost its claimed priority date and became exposed to an intervening disclosure.

Reader summary

Enanta's provisional disclosed a substituent as C2-C12-alkyl. The patent that later claimed priority from it disclosed C1-alkyl. Enanta said the C2 was a typographical error it noticed about a year later, that C1 was plainly what was meant, and that correcting it added no new matter. Pfizer said the provisional simply does not disclose C1-alkyl, so the priority claim fails and Pfizer's intervening disclosure of nirmatrelvir anticipates.

The court sided with Pfizer, in a sentence that carries the whole record: applicants are taken at their word that they invented what the provisional specifically disclosed, and by the same logic they did not invent what they did not disclose.

What makes this worth its own record is that the court insists this is not the usual written-description dispute. Ariad concerned whether a specification supports a broad genus claim; In re Ruschig concerned whether a broad genus supports a later-claimed species. Both ask how far a disclosure reaches. Here the question is narrower and harder to argue around: whether a disclosure covers a value it does not recite, where the argument for coverage is the applicant's own account of what it meant to write.

The rule about correcting obvious errors is not abolished. The opinion acknowledges that corrections adding no new matter are permitted where one of skill would consider the error obvious. What that route does not do is retroactively put the corrected text into the earlier document for priority purposes.

Why this matters when you are drafting

Provisionals are read literally, and a continuation inherits that reading. The entire priority chain rests on what the earliest document says. Claim 1 is the rule to carry: what was not disclosed was not invented, as of that date.

Correcting an error and preserving a date are different things. Claim 6 is the trap. The availability of a correction mechanism invites the assumption that the corrected text relates back. It does not. Ask separately whether the correction is permitted and whether the earlier date survives it.

Check the provisional's literal text against the claims you intend to pursue, early. A one-character discrepancy defeated a priority claim here. That is a mechanical review, it can be done at drafting time, and it is cheapest before an intervening publication exists.

This is the failure mode with the least argumentative room in Tier B. Claim 10. Where a blaze-marks case turns on characterising a disclosure, this turns on its absence, and there is correspondingly little to argue.

The exposure is always the intervening art. Claim 8, and the same pattern as univ-minnesota-v-gilead and purdue-v-collegium. When assessing a support defect in a priority document, the operative question is what published in the interval.

What the authority establishes

1. [ESTABLISHED, the holding in one sentence] "We respect applicants' statements in their specification that they invented what was specifically disclosed in the '048 provisional, but similarly we conclude that they did not invent what they did not disclose." (P. 10.)

1. [ESTABLISHED, the precise dispute] "the dispute is whether there is adequate written description support in the '048 provisional's disclosure of -NHC(O)-C2-C12-alkyl for the '953 patent's disclosure of -NHC(O)-C1-alkyl". (P. 5.) One carbon, and the priority chain turned on it.

1. [ESTABLISHED, the court distinguishes the usual cases] "This case does not present the type of written description disputes addressed in cases such as Ariad Pharms." (P. 5.) "The issue in Ariad was whether a patent specification provided adequate written description support for a broad genus claim" (P. 5.), and "the issue in In re Ruschig was whether a patent specification that disclosed a broad genus provided adequate written description support for a later-claimed species". (P. 5.)

1. [ESTABLISHED, the applicant's theory] "Enanta argued that the '953 patent is entitled to claim priority from the '048 provisional because Enanta's alteration to correct an obvious typographical error added no new matter not disclosed in the '048 provisional". (P. 4.) That theory failed.

1. [ESTABLISHED, when the error was noticed] Enanta says it realised the provisional "contained a typographical error" about a year after filing, and that the "C2" in the provisional "really should have been a" C1. (P. 3.) The correction was retrospective.

1. [ESTABLISHED, the error-correction route is not abolished] The opinion acknowledges corrections under the reissue statute "that add '[n]o new matter,' so long as one of skill in the art would consider the error to be 'obvious.'" (P. 6.) The route exists; it does not fix the earlier document's disclosure.

1. [ESTABLISHED, the possession standard applied] Written description requires "detail that one skilled in the art can clearly conclude that the inventor was 'in possession' of the claimed invention as of the filing date sought". (P. 8.) The filing date sought is the provisional's, which is what makes the omission decisive.

1. [ESTABLISHED, what the loss of priority cost] Pfizer argued that without the priority claim "Pfizer's intervening disclosure of nirmatrelvir anticipated" the asserted claims. (P. 4.) As in the rest of this tier, the priority holding is the whole case.

1. [IMPLEMENTED, intent is not evidence of disclosure] The court treats the applicant's account of what it meant as beside the point. Claim 1's symmetry does the work: the same deference that credits an applicant for what the document says withholds credit for what it does not.

1. [ASSUMED, drafter's inference, flagged as such] This holding is harder to argue around than the blaze-marks cases, because there is no disclosure to characterise. In a blaze-marks dispute a practitioner can argue about what the specification points at; here the value is simply absent. The opinion does not rank the difficulty; this follows from claim 3's distinction.

Quotations, verified against the source

All from the Federal Circuit's opinion. Page locators computed from the docket stamp.

• "contained a typographical error" (P. 3.)
• "Enanta argued that the '953 patent is entitled to claim priority from the '048 provisional because Enanta's alteration to correct an obvious typographical error added no new matter not disclosed in the '048 provisional" (P. 4.)
• "This case does not present the type of written description disputes addressed in cases such as Ariad Pharms." (P. 5.)
• "The issue in Ariad was whether a patent specification provided adequate written description support for a broad genus claim" (P. 5.)
• "the issue in In re Ruschig was whether a patent specification that disclosed a broad genus provided adequate written description support for a later-claimed species" (P. 5.)
• "detail that one skilled in the art can clearly conclude that the inventor was 'in possession' of the claimed invention as of the filing date sought" (P. 8.)
• "We respect applicants' statements in their specification that they invented what was specifically disclosed in the '048 provisional, but similarly we conclude that they did not invent what they did not disclose." (P. 10.)

Tensions and open questions

Where the line sits between an unsupported value and a genuine scrivener's error is not drawn. Claim 6 preserves the correction route while claim 1 denies it retroactive effect, and the opinion does not say whether any error is obvious enough to be read into the earlier document.

The court's distinction of Ariad and Ruschig is asserted rather than tested. Claim 3 sets this case apart as being about a value absent from the disclosure. Whether a chemical range missing one endpoint is truly a different kind of problem from a genus missing blaze marks is arguable, and the opinion states the distinction rather than defending it at length.

Unresolved by this record: whether a corrected provisional could ever support the corrected text from the original date. Everything here turns on the document as filed.

What this record does and does not cover

The written-description and priority analysis was read end to end, including the distinction of Ariad and In re Ruschig and the treatment of the typographical-error argument.

What was NOT read closely, and is therefore outside what this record can support: the chemical substance of the compounds, the anticipation analysis, and the reissue and correction authorities beyond the single proposition quoted. This record says nothing about whether Pfizer's disclosure in fact anticipates, only that anticipation was the consequence urged once priority failed.

Currency

This record was verified against its primary source on 2026-08-15. That is a fidelity check: the quotations, locators and attribution are accurate to that source. Whether the authority has since been narrowed, distinguished, abrogated or reversed has not been checked here. Confirm it still stands before relying on it.