Claim Construction

Claim construction — also called claim interpretation — is the process by which a court determines the meaning and scope of the terms used in a patent claim. Because a patent's patent claim language defines the legal boundaries of the invention, the words in the claims control nearly everything that follows in a dispute: whether an accused product infringes, whether prior art anticipates or renders the claim obvious, and whether the claim is even valid. Determining what those words mean is therefore the gateway question in most patent litigation.

In the United States, claim construction is a question of law decided by the judge rather than the jury. This allocation was settled by the US Supreme Court in Markman v. Westview Instruments, Inc., 517 U.S. 370 (1996), which is why the proceeding in which a court interprets disputed terms is commonly called a "Markman hearing." This article focuses on the US doctrine; claim interpretation exists in every patent system, but the procedures and standards differ substantially in other jurisdictions, and US case law should not be assumed to apply abroad.

Claim construction is frequently the most consequential ruling in a patent case. A single disputed term — what "a" means, whether "connected to" requires direct contact, whether "substantially" imposes a numerical limit — can determine the outcome of the entire dispute. If a court construes a term narrowly, an accused product may fall outside the claim and escape infringement; if the court construes the same term broadly, the claim may read on the accused product but also on prior art that invalidates it. The two sides of that vise — infringement and validity — mean that a construction favorable on one front is often unfavorable on the other, and skilled counsel argue constructions with both consequences in mind.

Because so much rides on the outcome, claim construction shapes settlement dynamics. Many cases resolve shortly after the Markman ruling, once the parties can see how the case will likely come out. For litigators, the quality of claim construction briefing and argument can be outcome-determinative in a way few other pretrial events are.

The doctrine also reaches beyond the courtroom. The way claims are drafted and amended during patent prosecution creates the record — the specification and the prosecution history — that a court will later mine for meaning. Prosecution counsel who understand how courts construe claims can draft specifications that support intended scope and avoid statements that later narrow the claims through prosecution-history estoppel.

How It Works

The legal framework. The foundational decision is Markman, which held that construing a patent claim is a matter of law reserved to the court. The substantive methodology was then clarified by the Federal Circuit, sitting en banc, in Phillips v. AWH Corp., 415 F.3d 1303 (Fed. Cir. 2005). Phillips established two governing principles. First, claim terms are given their "ordinary and customary meaning" — the meaning they would have to a person of ordinary skill in the art at the time of the invention, read in the context of the entire patent. Second, the court applies a hierarchy of evidence.

The intrinsic-evidence hierarchy. Under Phillips, intrinsic evidence is primary and is examined in a customary order:

1. The claims themselves, including how a term is used across different claims (claim differentiation). 2. The specification, which Phillips called the single best guide to the meaning of a disputed term, including any definition the patentee acted as a lexicographer to supply. 3. The prosecution history, the record of the back-and-forth with the USPTO, which can show how the applicant understood and limited the claims.

Extrinsic evidence. Dictionaries, treatises, and expert testimony may be consulted, but Phillips relegated them to secondary importance and held they cannot be used to contradict a meaning that is clear from the intrinsic record. Phillips specifically rejected the prior approach that had elevated dictionaries above the specification.

The standard of review on appeal. In Teva Pharmaceuticals USA, Inc. v. Sandoz, Inc., 574 U.S. 318 (2015), the Supreme Court refined how appellate courts review claim construction. When a district court construes a claim using only intrinsic evidence, the ultimate construction is a legal conclusion reviewed de novo. But when the court resolves a subsidiary dispute of fact about extrinsic evidence — for example, what a person of ordinary skill would have understood a technical term to mean — that factual finding is reviewed only for "clear error." Teva thus carved out a deferential standard for the factual underpinnings of construction.

The Markman hearing. Procedurally, the parties exchange proposed constructions for the disputed terms, brief their positions with citations to the intrinsic and extrinsic record, and present argument at the Markman hearing. The court then issues a claim construction order that the rest of the case — expert reports, summary judgment, trial — must follow.

Prosecution versus litigation standards. Historically the forum mattered. In ordinary examination the USPTO applies the "broadest reasonable interpretation" (BRI) standard, which reads claims more expansively than a court would. The PTAB also applied BRI in America Invents Act trials until the USPTO changed its rules effective November 13, 2018, adopting the Phillips standard for inter partes review, post-grant review, and covered business method proceedings. As a result, PTAB trial construction is now aligned with district court construction, although routine examination of pending applications still uses BRI.

Key Considerations for Law Firms

Identify the dispositive terms early. Not every term needs construction. Effective claim construction strategy isolates the handful of terms whose meaning will actually decide infringement or validity and concentrates resources there. AI patent tools can accelerate this triage by surfacing how a term is used across the claims, the specification, and the file history. &AI offers an agentic platform whose agent, "Andy," can work across both prosecution and litigation tasks — including assembling claim charts and invalidity contentions — which helps teams see how a proposed construction plays out against accused products and prior art.

Build the construction from the intrinsic record. Because Phillips prioritizes intrinsic evidence, the strongest constructions are anchored in the specification and prosecution history rather than in dictionaries or expert opinion. Patlytics supports element-by-element claim charting and infringement analysis, which can help counsel test how a candidate construction maps claim limitations onto an accused product before committing to it in a brief.

Coordinate prosecution and litigation positions. Statements made during prosecution can bind the patentee later through prosecution-history estoppel. Firms that handle both prosecution and enforcement should track the file history carefully so that arguments made to secure allowance do not inadvertently surrender claim scope needed for enforcement.

Use landscape context. Understanding how a term has been construed in related patents or asserted in a portfolio informs strategy. PatSnap provides patent analytics, landscape mapping, and competitive intelligence that can place a disputed patent in the context of a broader family or technology area.

Preserve the factual record for appeal. After Teva, the way a court resolves underlying factual questions affects the standard of review. Where extrinsic evidence such as expert testimony is needed, counsel should make a clear record so that favorable factual findings receive clear-error deference on appeal.

Limitations and Risks

Claim construction is unpredictable. Reasonable judges construe the same term differently, and reversal rates on appeal have historically been significant. No construction can be guaranteed, and litigation strategy should account for the possibility that a key term is construed unfavorably.

Forum and standard differences create traps. A construction adopted under BRI during examination or in a pre-2018 PTAB proceeding may not match the Phillips construction a district court reaches. Counsel should not assume that a construction from one forum or era transfers cleanly to another.

AI tools do not perform legal construction. AI patent platforms can organize the intrinsic record, generate claim charts, and surface candidate meanings, but they do not decide what a term legally means. Claim construction is an exercise of legal judgment that depends on the specific record and governing case law, and the determination ultimately rests with the court. Any AI-generated analysis must be reviewed by qualified counsel, and its characterizations of the specification, prosecution history, and case authorities verified against the underlying documents.

Confidentiality and accuracy. Patent disputes involve sensitive technical and strategic information. Firms using AI tools should confirm how each vendor handles confidential matter and should independently verify any AI output — including citations to cases such as Markman, Phillips, and Teva — before relying on it in a filing.

Jurisdictional limits. This doctrine is US-specific. Other patent systems interpret claims under their own rules and procedures, and the Markman/Phillips/Teva framework should not be applied to foreign proceedings without local advice.