A single word – “to” instead of “at,” “second” instead of “first,” “case” instead of “cover” – can decide a multimillion-dollar patent case. Courts have repeatedly shown that claim language is read exactly as written and the gap between what a drafter meant and what the patent says is a gap litigants exploit relentlessly. For patent litigators and the paralegals supporting them, proofreading isn’t administrative housekeeping – it’s a substantive litigation strategy.
Why Claim Language Gets Read Literally
Patent claims serve a public notice function: they tell competitors exactly what is off-limits. Courts protect that function aggressively, even when it produces results the patentee clearly didn’t intend. The canonical illustration is the “raise the temperature of the dough to” case, where a claim’s use of “to” instead of “at” made the claimed process literally impossible to perform – yet the Federal Circuit refused to correct it, emphasizing the public notice function and the principle that courts cannot redraft or rewrite clear claim language, even to save the claim from invalidity, holding that where a claim is susceptible to only one reasonable construction and that interpretation is nonsensical, the claim must still be invalidated rather than judicially repaired.
That single-word error cost the patentee the case. It’s the clearest possible demonstration that in patent litigation, “everyone knew what I meant” is not a defense.
Judicial Correction Is Available – But Only Barely
Courts can fix obvious drafting errors in narrow circumstances, but the bar is high and getting more precisely defined. In a November 2025 decision, the Federal Circuit reinforced that judicial correction of a patent claim error is available only where the error is evident on the face of the patent, the correction is not subject to reasonable debate based on the patent and its prosecution history and the correction is limited to an obvious minor typographical or clerical error. In that case – a dispute over a drilling tool patent – the district court had initially invalidated the asserted claims as indefinite because it read the disputed phrase as reflecting an intentional drafting choice rather than a mistake, partly because the patentee hadn’t sought a certificate of correction from the USPTO. The Federal Circuit reversed, finding the error correctable – but only after the case had already been fully litigated through invalidity on this exact question.
Compare that to a case where the Federal Circuit declined to help: a district court found that a claim’s disputed phrase was not evident as an error on the face of the patent and reasoned that the error’s presence throughout both the claims and the specification reflected an intentional drafting choice – not a slip – a conclusion reinforced by the patentee’s own failure to pursue a correction under 35 U.S.C. §255, the statute that lets the USPTO fix certain clerical and typographical errors. Courts read a party’s failure to seek prompt correction as evidence the error wasn’t actually an error at all.
The throughline for litigators: the same typo can either sink or survive a case depending entirely on how obvious, consistent and promptly addressed it appears in the intrinsic record. That record – the patent as issued and its prosecution history – is fixed by the time litigation starts. Proofreading has to happen before filing, because by litigation it’s almost always too late to change the words themselves.
Even Courts Will Rewrite Claims – When the Error Is Clean Enough
In one notable case, the Federal Circuit went further than merely correcting a typo – it rewrote a disputed claim term from “case” to “cover” where the patent’s own internal logic and prosecution history made clear that only “cover” made structural sense, applying the established rule that a district court may correct obvious minor typographical and clerical errors in patents. But this remains the exception, not something litigators can rely on. It required the error to be self-evidently the only sensible reading and consistent with everything else in the document – a bar most ambiguous or debatable drafting choices simply won’t clear.
Typos Don’t Just Live in the Patent – They Live in Litigation Documents Too
Proofreading discipline matters beyond the patent itself. In a 2023 decision, the Federal Circuit addressed a typographical error in an expert declaration, holding that the error should not have been treated as dispositive of the underlying technical dispute, particularly where neither party had ever briefed, argued, or flagged the typo as significant – meaning the tribunal’s reliance on it violated basic notice requirements. The case is a reminder that proofreading obligations extend to every document that enters the litigation record: expert reports, claim charts, infringement contentions and briefs are all read as carefully – and as literally – as the patent claims themselves.
Practical Implications for Litigation Teams
Before filing – the highest-leverage moment. Once a patent issues, the words are largely locked in absent the narrow certificate-of-correction or judicial-correction paths described above. Rigorous claim proofreading during prosecution – checking antecedent basis, verb tense, numbering consistency (first/second, upper/lower) and internal consistency between claims and specification – is dramatically cheaper and more effective than trying to fix a problem during litigation.
During prosecution history review, before asserting a patent. Before filing an infringement suit, litigation counsel should proofread the patent’s file history the way an accused infringer’s counsel will: looking for the same inconsistencies, unclear antecedents and internal contradictions courts have flagged in analogous cases. If an error exists, pursuing a certificate of correction under 35 U.S.C. §255 before asserting the patent measurably strengthens a later argument that any remaining ambiguity is a minor clerical slip rather than a substantive, intentional choice – courts have specifically read the absence of that step as evidence against the patentee.
Consistency across the patent family matters. Because courts assess whether an alleged error appears throughout both claims and specification and whether it reads as deliberate or accidental, proofreading needs to check not just the disputed claim term in isolation but every occurrence of that term and its logical relatives across the entire document – a single inconsistent usage can undercut an otherwise strong “obvious typo” argument.
Litigation documents need the same rigor as the patent. Expert reports, infringement contentions and claim construction briefs should go through the same level of proofreading scrutiny as the patent itself, since a stray error in these documents can become unexpectedly outcome-determinative, as the Corephotonics dispute demonstrated.
Relative and imprecise language is a standing indefiniteness risk. Courts have repeatedly found claims indefinite for relying on vague, relative terms – a golf club claim using “improved” was held indefinite, for example – which means proofreading during drafting should flag not just typos but imprecise language likely to invite an indefiniteness challenge years later in litigation.
A Working Checklist for Litigation-Facing Proofreading
| Check | Litigation risk if missed |
| Antecedent basis for every “the”/”said” reference | Indefiniteness challenge |
| Consistent use of “first”/”second,” “case”/”cover,” and similar paired terms across claims and spec | Ambiguity a court may or may not correct |
| Verb tense and prepositions in method/process claims (e.g., “to” vs. “at”) | Can render limitations literally impossible – fatal, uncorrectable error |
| Relative or subjective terms (“improved,” “substantially,” “high”) | Indefiniteness defense |
| Consistency between claim numbering and in-text cross-references | Confusion invites a “reasonable debate” finding, foreclosing correction |
| Prompt certificate-of-correction filings for known clerical errors | Delay can be read as evidence the error was intentional |
| Expert reports, contentions and briefs proofread with the same rigor as the patent | Even collateral documents can become case-dispositive |
Conclusion
Patent claim language is interpreted with unforgiving literalism and the cases above show both edges of that sword: a single mis-chosen preposition can invalidate an otherwise sound claim beyond any court’s power to fix, while a genuinely obvious typo – caught early, treated consistently and pursued through the right correction channel – can be preserved even after a district court gets it wrong. The dividing line between those outcomes is rarely legal sophistication; it’s whether someone caught the error before it calcified into the public record. For litigation teams, that makes rigorous proofreading – of the patent itself, its prosecution history and every document generated in support of enforcement – not a clerical afterthought but a frontline litigation strategy.