Search First, Draft Deliberately: How a Reference-Specific Background Can Support Patent Eligibility and Preserve Claim Meaning

Executive Summary:  A deliberate prefiling prior-art search can serve as a specification-design tool, not merely a patentability screen. By identifying “framing references,” accurately describing their architectures and limitations, and tying each distinction to a claimed technical mechanism, an applicant can create an intrinsic record that supports patent eligibility under 35 U.S.C. § 101 and informs later claim construction. That record may allow the patent owner to demonstrate that a later-cited reference employs the same conventional architecture described in the Background and therefore lacks the claim’s distinguishing limitation. The article emphasizes, however, that these effects are forum specific: applicant-admitted prior art is treated differently in examination, IPR, and district-court litigation, while claim construction remains a separate intrinsic-record inquiry. It also distinguishes Rule 56 disclosure obligations from optional substantive characterization and prosecution disclaimer under Ekchian from specification disavowal under Thorner, SciMed, and Indivior. Ultimately, the article argues that disciplined, claim-tethered Background drafting can produce valuable and durable boundaries—provided applicants accept that those same boundaries will constrain both validity and infringement positions.

I. Introduction

Patent drafters are often taught to treat the Background as a liability. Every sentence may be characterized as an admission, every criticism of an earlier system may be invoked as a specification disavowal, and every description of “the invention” may later be invoked to narrow the claims. The conventional response is understandable: keep the Background short, generic, and noncommittal.

That approach avoids some risks, but it creates others. When an issued patent is later attacked under 35 U.S.C. § 101, the patent owner may need the specification to show that the claims concern a concrete technological improvement rather than a desired result implemented with generic tools. When a later prior-art challenge turns on the meaning of a claim term, the patent owner may need the intrinsic record to establish the technical context in which a skilled artisan would have understood that term at filing. A deliberately drafted Background can help in both settings. Indeed, an intentional and carefully bounded exclusion can become an affirmative asset: if a later challenger relies on a reference employing the excluded conventional architecture, the patent owner can invoke the filing-date boundary to show that the reference lacks the claim's distinguishing limitation.

The better practice is therefore neither maximal silence nor maximal criticism. It is to search before drafting, identify the references that actually frame the claimed advance, and address each such framing reference with technical precision. For purposes of this article, a “framing reference” is a reference whose architecture, operation, or limitation helps define the claimed technical advance; the term does not express a conclusion that the reference is “material to patentability” under 37 C.F.R. § 1.56(b). For each reference, the application should state what the reference affirmatively discloses, identify the particular limitation or relationship it does not describe, and connect that distinction to a mechanism carried by the claims and affirmatively supported in the Detailed Description. Other search results should be submitted when required by Rule 56 and may otherwise be listed without substantive characterization. The object is not to turn the Background into an advocacy brief. It is to create a technically coherent intrinsic record that the applicant is prepared to live with throughout the patent’s life.

II. The Search as a Specification-Design Tool

A prefiling search is usually justified as a way to evaluate novelty and nonobviousness under 35 U.S.C. §§ 102 and 103. That is only part of its value. A disciplined search also identifies the vocabulary of the field, the closest technical architectures, the functions already performed by individual components, and the distinctions likely to matter during examination or litigation. Those findings should influence more than the initial claim set. They should shape the Background, Summary, Detailed Description, drawings, alternative embodiments, and continuation strategy.

For each framing reference, the drafter should be able to answer five questions. What arrangement or process does the reference actually disclose? What technical limitation follows from that arrangement? What component relationship, data flow, control step, or sequence changes in the applicant’s solution? How does that change produce the asserted technical effect? Where does the claim recite the mechanism responsible for that effect? If the application cannot answer all five, the draft may be relying on an unclaimed advantage or an unsupported distinction.

Reference-by-reference treatment also prevents the prior art from collapsing into an imprecise collective. One reference may disclose a server, another may disclose peer nodes, and a third may use trace routes for network testing. Describing all three simply as “conventional systems” can obscure the claimed ordered combination and may inadvertently suggest that the combination itself was conventional. Separate treatment allows the application to distinguish what was known at the component level from the particular arrangement or use alleged to be inventive. The Background should not, however, supply an obviousness rationale by explaining why a skilled artisan would have combined the references. Its function is to preserve technical context, not to perform the challenger’s combination analysis.

Precision also matters because labeling subject matter “prior art” can have substantive consequences. The Patent Office treats an applicant’s identification of another’s work as prior art as an admission that may be used in patentability analysis. See In re Nomiya, 509 F.2d 566, 571 (C.C.P.A. 1975); MPEP § 2129. The Background should therefore avoid unsupported statements that every feature of a cited document was conventional, publicly known, or legally available as prior art. “Reference A describes” is usually more accurate than “the prior art establishes,” and “Reference A does not expressly describe X” is usually safer than “conventional systems cannot perform X.”

The distinction between disclosure and characterization is illustrated by Malvern Panalytical Inc. v. TA Instruments-Waters LLC, 85 F.4th 1365, 1375–78 (Fed. Cir. 2023). There, the Federal Circuit explained that the degree to which an applicant characterizes a cited reference affects how informative the reference may be during claim construction. A bare IDS listing did not amount to an admission that statements from an unrelated prosecution controlled the meaning of the patent’s claims. Id. at 1376–78. A substantive Background discussion is different: its purpose is to say something about the technology. That added usefulness carries added responsibility.

III. Applicant-Admitted Prior Art Depends on the Forum

An applicant’s substantive identification of another’s work as prior art can create applicant-admitted prior art (“AAPA”), but the consequences are forum specific. The same statement may support a rejection during original examination, have a restricted evidentiary role in an inter partes review, inform a district court’s validity analysis, and affect claim construction without operating as prior art at all. A Background admission therefore does not have one universal post-grant effect.

The applicant’s disclosure obligation is distinct from the choice whether to characterize a reference substantively. Rule 56 does not impose a general duty to conduct a patentability search. MPEP § 609. Once an individual covered by the rule knows information to be material to patentability, however, the information must be disclosed to the Office. 37 C.F.R. § 1.56(a); MPEP §§ 2001.04–2001.06. Information is material under Rule 56 when it is noncumulative and either establishes, alone or with other information, a prima facie case of unpatentability or refutes, or is inconsistent with, a position the applicant takes in opposing unpatentability or asserting patentability. 37 C.F.R. § 1.56(b).

The duty applies to each pending claim and continues until the patent is granted. MPEP § 2001.04. A bare IDS listing is not an admission that the cited information is prior art against the claims, MPEP § 2129(IV), and the IDS itself is not an admission that the information is material under Rule 56, 37 C.F.R. § 1.97(h). Disclosure and substantive characterization are therefore separate questions.

A.                 Examination: Admissions Can Support Rejection, but a Bare IDS Does Not

During examination, the USPTO treats an applicant’s statement identifying another’s work as “prior art” as an admission that may be used in anticipation and obviousness determinations, even if the admitted subject matter would not independently qualify under a statutory category of § 102. MPEP § 2129(I)–(II); In re Nomiya, 509 F.2d 566, 571 (C.C.P.A. 1975). That principle does not ordinarily convert the inventor’s own work into prior art without a statutory basis. Riverwood Int’l Corp. v. R.A. Jones & Co., 324 F.3d 1346, 1354–55 (Fed. Cir. 2003). Nor does a bare IDS listing create that admission. 37 C.F.R. § 1.97(h); MPEP § 2129(IV).

B.                 Inter Partes Review: Section 311(b), Qualcomm, and the 2025 USPTO Policy

IPR practice is narrower than original examination. Section 311(b) permits a petitioner to challenge claims only under §§ 102 or 103 and only on the basis of prior art consisting of patents or printed publications. 35 U.S.C. § 311(b). A description of prior art appearing in the challenged patent is not itself a prior-art patent or printed publication merely because the description is printed in the patent.

In Qualcomm Inc. v. Apple Inc. (“Qualcomm I”), the Federal Circuit accordingly held that AAPA cannot form the basis of an IPR ground. 24 F.4th 1367, 1373–77 (Fed. Cir. 2022). The court did not categorically exclude AAPA from the proceeding; it recognized that admissions may serve as evidence of a skilled artisan’s background knowledge. Id. at 1375–77.

On remand, the Board adopted an “in combination” rule under which AAPA did not form the basis of a ground whenever the petition combined it with a qualifying patent or printed publication. The Federal Circuit rejected that categorical rule in Qualcomm Inc. v. Apple Inc. (“Qualcomm II”), 134 F.4th 1355, 1364–68 (Fed. Cir. 2025). Pairing AAPA with qualifying documentary art does not automatically cure the problem; the question is whether AAPA actually forms the basis, or part of the basis, of the ground. Id. at 1365–68.

For IPR petitions filed on or after September 1, 2025, current USPTO policy enforces 37 C.F.R. § 42.104(b)(4), which requires the petition to identify where each claim element appears in the relied-upon prior-art patents or printed publications. AAPA, expert testimony, common sense, and other forms of general knowledge may not supply a missing claim limitation. They may still support a motivation to combine, demonstrate the knowledge of a person of ordinary skill, or explain how that person would understand a qualifying reference’s disclosure. Memorandum from Coke Morgan Stewart, Acting Under Secretary of Commerce for Intellectual Property & Acting Director, U.S. Patent & Trademark Office, to Members of the Patent Trial & Appeal Board, Enforcement and Non-Waiver of 37 C.F.R. § 42.104(b)(4) and Permissible Uses of General Knowledge in Inter Partes Reviews 1–3 (July 31, 2025).

These restrictions are specific to IPR. Post-grant review under 35 U.S.C. § 321(b) permits a broader set of grounds, so “post-grant challenge” should not be used as though every PTAB proceeding has the same evidentiary limits.

C.                 District-Court Validity Litigation: Section 311(b) Does Not Apply

District-court validity litigation is not governed by § 311(b)’s patents-and-printed-publications restriction or the 2025 USPTO policy. A defendant may invoke the grounds available under 35 U.S.C. § 282(b), subject to the presumption of validity and the clear-and-convincing-evidence burden. See Microsoft Corp. v. i4i Ltd. P’ship, 564 U.S. 91, 95 (2011). Admissions in a specification concerning the prior art may be evidence in, and may bind the patentee during, a later obviousness inquiry. PharmaStem Therapeutics, Inc. v. ViaCell, Inc., 491 F.3d 1342, 1362 (Fed. Cir. 2007). The IPR restrictions therefore do not make careless Background characterizations harmless in district court.

D.                 Claim Construction: Intrinsic Evidence Is a Separate Inquiry

Claim construction does not ask whether the Background itself qualifies as prior art. Claims are read in the context of the entire patent, and the specification is ordinarily the most important source of their meaning. Phillips v. AWH Corp., 415 F.3d 1303, 1312–17 (Fed. Cir. 2005) (en banc). A Background statement may therefore operate as claim-constraining intrinsic evidence even when § 311(b) prevents a petitioner from using it as the basis of an IPR ground.

The degree of characterization matters. A bare IDS citation did not admit the truth of statements made in an unrelated prosecution or make those statements controlling intrinsic evidence in Malvern Panalytical Inc. v. TA Instruments-Waters LLC, 85 F.4th 1365, 1375–78 (Fed. Cir. 2023). A substantive specification comparison, an express definition, and an unmistakable exclusion are different drafting acts. Specification lexicography or disavowal is also doctrinally distinct from prosecution disclaimer.

IV. Building a Claim-Tethered § 101 Record

A prefiling search can provide the factual foundation needed to explain why a claimed arrangement is a technological solution rather than merely a desired result performed on a computer. Under the two-step framework of Alice Corp. v. CLS Bank International, a court first asks whether the claim is directed to an ineligible concept and, if so, whether the claim contains an inventive concept sufficient to transform that concept into a patent-eligible application. 573 U.S. 208, 217–18, 221 (2014). The specification can help identify the focus of the claimed advance at step one and can help explain at step two why a particular arrangement of otherwise familiar components was not merely well-understood, routine, and conventional activity.

The search does not itself establish eligibility. Novelty and eligibility remain distinct inquiries. See Diamond v. Diehr, 450 U.S. 175, 188–89 (1981). Nor does the absence of a claimed feature from located references automatically supply an inventive concept. The search’s value is practical: it forces the applicant to define the technical baseline, isolate the mechanism that changes that baseline, and align the claims and disclosure around that mechanism before filing. The specification may illuminate the claims, but it cannot import an unclaimed technical solution merely because that solution appears in the written description. See ChargePoint, Inc. v. SemaConnect, Inc., 920 F.3d 759, 766–70 (Fed. Cir. 2019).

Federal Circuit decisions repeatedly show the importance of that alignment. In Enfish, LLC v. Microsoft Corp., the specification’s explanation of how a self-referential database table differed from conventional database structures and improved computer functionality helped establish that the claims were directed to a specific improvement rather than an abstract idea. 822 F.3d 1327, 1335–39 (Fed. Cir. 2016). In BASCOM Global Internet Services, Inc. v. AT&T Mobility LLC, the court explained that an inventive concept may reside in a nonconventional ordered combination of familiar components. 827 F.3d 1341, 1350–52 (Fed. Cir. 2016). And in McRO, Inc. v. Bandai Namco Games America Inc., the claims’ specific rules for producing automated animation mattered because they claimed a particular process rather than the result of producing accurate lip synchronization. 837 F.3d 1299, 1313–16 (Fed. Cir. 2016).

A.                 CosmoKey: Why Reference-by-Reference Precision Matters

CosmoKey Solutions GmbH & Co. KG v. Duo Security LLC directly illustrates the value of treating identified references separately and accurately. The patent’s Background described three earlier mobile-phone authentication techniques. The district court read that discussion as admitting that the later-recited authentication sequence was well understood, routine, and conventional. The Federal Circuit rejected that reading because none of the three references taught the final four claimed steps, including timed activation of a normally inactive authentication function, communication of its active status, and automatic deactivation. 15 F.4th 1091, 1095–99 (Fed. Cir. 2021).

Read in context, the Background distinguished the earlier techniques from the claimed sequence and explained how the sequence achieved increased security with reduced complexity. The Federal Circuit held that the ordered combination supplied a specific improvement to computer-implemented authentication and reversed the Rule 12(c) ineligibility judgment at Alice step two. Id. at 1097–99. The case does not establish that every search result belongs in the specification. It establishes the more disciplined point: when the Background discusses several references, separate and technically accurate treatment may prevent the discussion from being misread as an admission that the claimed combination was itself conventional.

B.                 Cooperative Entertainment as a Drafting Model

Cooperative Entertainment, Inc. v. Kollective Technology, Inc. provides a particularly useful example of an art-informed specification supporting an eligibility position. 50 F.4th 127, 129–36 (Fed. Cir. 2022) (“Cooperative I”). The patent concerned distribution of large files through a dynamic peer-to-peer network. Its specification described the earlier technical baseline: video streaming was ordinarily controlled by a content-distribution network, or CDN, that distributed content from the originating CDN server. The claimed system instead used peer nodes consuming the same content contemporaneously to distribute content outside controlled networks and CDNs. It also used trace routes in segmenting requested content. Id. at 129–31.

The specification did more than assert that this approach was better. It described the changed network architecture, explained how content moved through that architecture, and identified concrete effects, including smoother playback, less buffering, increased redundancy and efficiency, reduced CDN cost, and greater use of client-device capacity. Id. at 131–33. Those details supplied a technical explanation rather than an aspirational list of benefits.

That record became important after the district court dismissed the infringement complaint under Federal Rule of Civil Procedure 12(b)(6), holding the claims ineligible. The district court characterized the patent as implementing the abstract idea of preparing and transmitting data using generic computer components and conventional technology. The Federal Circuit did not resolve the parties’ dispute at Alice step one. Instead, it held that two plausibly alleged inventive concepts precluded dismissal at step two: the particular dynamic peer-to-peer architecture operating outside CDN control and the use of trace routes in content segmentation. Cooperative I, 50 F.4th at 131.

The result depended on alignment. The claim language recited the asserted mechanisms; the written description explained how those mechanisms differed from and improved the earlier architecture; the prosecution history preserved the relevant technical distinction; and the amended complaint plausibly alleged that the mechanisms were not well understood, routine, or conventional. Id. at 131–35. That coordinated record created factual disputes that could not be resolved against the patent owner at the pleading stage. Id. at 133, 135–36; see also Berkheimer v. HP Inc., 881 F.3d 1360, 1368–70 (Fed. Cir. 2018).

The decision answers a recurring objection: individual components may be conventional without the claimed arrangement being conventional. Section 101 does not require the applicant to invent every computer, server, or network node. If the asserted advance resides in their particular relationship or operation, the claim should recite that arrangement and the specification should explain how it improves the technology. Cooperative I rejected the contention that conventional P2P networks and CDNs ended the inquiry because useful improvements to networks can be patentable even when standard computing equipment is used. 50 F.4th at 135.

The procedural limit is equally important. Cooperative I did not finally hold the claims patent eligible. It held only that the complaint and intrinsic record contained plausible allegations of inventive concepts sufficient to defeat a Rule 12 motion. Id. at 135–36. The case likewise did not hold that conclusory phrases such as “improved efficiency” or “not conventional” automatically create a factual dispute. The asserted effects were tied to a particular claimed architecture, and the specification explained how that architecture operated. A later complaint could repeat and contextualize the filing-date account, but it could not safely substitute a litigation-created technical theory for an explanation missing from the patent.

The drafting lesson is therefore concrete. For each framing reference, the Background should identify the reference’s actual technical arrangement. The application should identify the operational constraint associated with that arrangement, explain the mechanism by which the disclosed invention changes it, and use the Detailed Description and drawings to show the relevant component relationship, timing, data flow, or control path. The independent claims should carry the relationship on which the eligibility position is expected to depend; dependent claims should preserve narrower implementations and alternative eligibility positions.

C.                 The Eligibility Position Must Remain the Infringement Position

The later history of Cooperative Entertainment exposes the cost of that specificity. After remand, Cooperative failed to plead that the accused product used trace routes to segment content. It then attempted to characterize the claims as using trace routes to segment peer networks rather than content. The Federal Circuit held that the new theory was waived and judicially estopped because Cooperative had previously prevailed by arguing that trace routes were used to segment content; the claim language and intrinsic record confirmed the same understanding. Cooperative Entertainment, Inc. v. Kollective Technology, Inc., No. 2024-1550, slip op. at 5–10 (Fed. Cir. Dec. 16, 2025) (nonprecedential) (“Cooperative II”). The distinction that helped the patent survive an eligibility dismissal thus constrained the patent owner’s later infringement theory.

That sequel captures the central principle of search-informed drafting: a mechanism invoked as the inventive concept must be treated as a genuine part of the invention’s scope. The applicant should not rely on a narrow technical story for eligibility while expecting to adopt a broader, inconsistent story against an accused product.

V. Preserving Claim Meaning Against Later Prior Art

The same reference-specific discussion may later matter when a challenger maps new prior art onto the claims. Claim terms are read in the context of the entire patent, and the specification is often the most important source of their meaning. Phillips v. AWH Corp., 415 F.3d 1303, 1312–17 (Fed. Cir. 2005) (en banc). A Background that accurately identifies the filing-date architecture and the role performed by a disputed component can help show that a superficially similar reference does not disclose the limitation as a skilled artisan would understand it.

The correct proposition is not that a patent owner may narrow the claims whenever later art appears. Courts do not rewrite claims merely to preserve validity. Rather, the filing-date intrinsic record may establish that a term already carried a particular technical meaning or required a particular relationship. If the asserted reference lacks that properly construed limitation, it does not anticipate the claim and may not support the proposed obviousness combination.

A.                 IBM v. Iancu: The Affirmative Claim-Construction Example

In International Business Machines Corp. v. Iancu, the Patent Trial and Appeal Board construed “federated computing environment” broadly enough to include two computer systems within a single enterprise. 759 F. App’x 1002, 1007–09 (Fed. Cir. 2019) (nonprecedential). The Federal Circuit rejected that reading. The patent’s Background explained the distinction between systems operating within one enterprise and a federation involving cooperative relationships among separate enterprises. The specification then defined and repeatedly described the federated environment in inter-enterprise terms. Id. at 1007–08. Read as a whole, the intrinsic record required a plurality of enterprises.

That construction materially affected the prior-art analysis. The Federal Circuit vacated and remanded one inter partes review because the Board had used an overbroad construction and reversed another anticipation determination because substantial evidence did not show the separate single-sign-on limitation. Id. at 1008–10. Although nonprecedential and decided under the then-applicable broadest-reasonable-interpretation standard, IBM demonstrates the affirmative potential of a technically precise Background: it may preserve the context needed to prevent a later tribunal from treating materially different architectures as interchangeable.

B.                 Prosecution Disclaimer: Ekchian and Deliberate Surrender

Prosecution disclaimer concerns representations made to the Patent Office during prosecution, including substantive arguments accompanying an IDS. As the Federal Circuit explained, “Since, by distinguishing the claimed invention over the prior art, an applicant is indicating what the claims do not cover, he is by implication surrendering such protection.” Ekchian v. Home Depot, Inc., 104 F.3d 1299, 1304, 41 U.S.P.Q.2d (BNA) 1364, 1368 (Fed. Cir. 1997). “Surrender” is therefore the appropriate description of the prosecution-history consequence in Ekchian; it should not be used as shorthand for a statement appearing only in the specification.

Specification language may constrain claim scope through a route distinct from prosecution disclaimer. A patentee may expressly define a term or clearly disavow otherwise available scope, but both standards are demanding. Thorner v. Sony Computer Entertainment America LLC, 669 F.3d 1362, 1365–67 (Fed. Cir. 2012). A Background comparison that merely describes one implementation or criticizes an alternative does not automatically exclude every alternative. Unless the specification supplies lexicography or a clear and unmistakable disavowal, its statements remain claim-constraining intrinsic evidence considered under ordinary claim-construction principles.

When the full intrinsic record nevertheless establishes that a claim requires a particular feature or relationship, that boundary can become useful when a challenger relies on a reference employing materially the same architecture as the conventional system described in the Background. The patent owner can seek a construction faithful to the intrinsic record and then demonstrate that the challenger’s reference shares the conventional system’s relevant structural or functional constraints—and therefore lacks the claim’s distinguishing limitation. IBM v. Iancu illustrates the point: an express specification definition, read with the Background, required distinct enterprises lacking unitary control, and the Federal Circuit vacated an anticipation determination resting on the Board’s broader construction. 759 F. App’x at 1007–09. The Background’s characterization is not itself proof that a later reference is equivalent, however. The patent owner must establish the relevant structural or functional correspondence through the reference’s disclosure and, where appropriate, expert evidence.

The strategy carries a symmetrical cost. The proposed construction must follow ordinary claim-construction principles and apply equally to validity and infringement; a patent owner cannot create a validity-only narrowing after a challenge is filed. See Phillips, 415 F.3d at 1327 (limiting the validity-preserving canon to residual ambiguity after application of ordinary interpretive tools). The usefulness of a prosecution disclaimer depends on its precision. Ekchian vacated summary judgment because the district court had read the applicant’s distinction more broadly than the record permitted: the applicant had distinguished the functional roles of the relevant liquids, not their conductivity levels generally. 104 F.3d at 1303–05. When a properly construed limitation excludes the relevant conventional architecture and the challenger’s reference is proven to fall on that side of the boundary, the limitation can defeat an anticipation mapping or expose a missing element in an obviousness combination—but that same boundary will also constrain literal infringement and potentially the doctrine of equivalents.

The contrast with Malvern Panalytical is instructive. A bare IDS citation may have little claim-construction weight, while a substantive characterization can be informative and potentially limiting. Malvern Panalytical, 85 F.4th at 1375–78. The drafter should make that tradeoff consciously. If a difference is important enough to support eligibility or distinguish later art, it should be stated accurately and tied to claim language. If no substantive characterization is needed, a bare disclosure may be preferable.

VI. The Value—and Price—of a Durable Intrinsic Record

The same construction must govern validity and infringement. A patent owner cannot use a narrow meaning to avoid prior art and a broader meaning to capture an accused product. The cases therefore do not support creating hidden limitations as insurance against unknown art. They support deliberate intrinsic-record design, with full awareness of the claim scope the intrinsic record may exclude. Ordinary meaning is not displaced merely because the specification describes a preferred implementation; the pertinent exceptions are lexicography and clear disavowal. See Thorner v. Sony Computer Entertainment America LLC, 669 F.3d 1362, 1365–67 (Fed. Cir. 2012).

A.                 SciMed: Specification Disavowal Applies to Literal Scope and Equivalents

In SciMed Life Systems, Inc. v. Advanced Cardiovascular Systems, Inc., the claims concerned balloon-dilation catheters with inflation and guide-wire lumens. 242 F.3d 1337, 1340–41 (Fed. Cir. 2001). Although the claim language could have encompassed both coaxial and side-by-side arrangements, the specification criticized the disadvantages of prior-art side-by-side lumens, repeatedly characterized coaxial lumens as part of the invention, and identified the coaxial arrangement as the structure for all embodiments. Id. at 1342–44.

The Federal Circuit treated that application-wide pattern as a clear specification disavowal. It construed the claims to exclude the side-by-side arrangement, affirmed noninfringement, and refused to permit the patentee to recapture the excluded structure through the doctrine of equivalents. Id. at 1344–47. If later prior art had disclosed only the side-by-side arrangement, the same construction might have helped preserve validity. But it also excluded an accused product using that arrangement. SciMed does not mean that every criticism narrows a claim; the disavowal arose from repeated, categorical statements, including language applying the coaxial structure to all embodiments. The case nonetheless illustrates why “the present invention,” “all embodiments,” “essential,” and “cannot” require special care.

B.                 Indivior: Specification Disavowal Can Follow the Shared Disclosure

Indivior Inc. v. Dr. Reddy’s Laboratories, S.A. adds a continuation-practice warning. In a nonprecedential preliminary-injunction appeal, the Federal Circuit concluded that repeated statements disparaging conventional top-air drying and contrasting that technique with the “present invention” excluded films made solely through conventional top-air drying, even though the continuation claims no longer expressly recited drying. 752 F. App’x 1024, 1029–35 (Fed. Cir. 2018) (nonprecedential). The divided decision should be described cautiously, but it illustrates that deleting terminology from later claims may not recover territory unmistakably excluded by the shared specification.

The later precedential appeal involving the related patent family reinforces the point. The Federal Circuit held that the specification repeatedly disparaged conventional top-air drying because it failed to produce the uniform films that were central to the asserted invention, and it affirmed a construction excluding that method. Indivior Inc. v. Dr. Reddy’s Laboratories, S.A., 930 F.3d 1325, 1336–40 (Fed. Cir. 2019). An eligibility narrative that describes one mechanism as the reason an invention works may be valuable. Repeatedly defining that mechanism as indispensable, however, can make the narrative a family-wide restriction.

C.                 Tronzo: Prior-Art Discussion Is Not Affirmative Written-Description Support

Tronzo v. Biomet, Inc. presents a separate problem under 35 U.S.C. §§ 112 and 120. The parent application described a conical acetabular cup, discussed other cup shapes only as inferior prior art, and emphasized the advantages of the conical configuration. 156 F.3d 1154, 1158–59 (Fed. Cir. 1998). That discussion did not demonstrate possession of a broader genus encompassing the other shapes. The later generic claims therefore lacked written-description support and lost the parent’s filing date. Id.

The lesson is easy to miss. A Background statement that another reference discloses an alternative does not affirmatively disclose that the applicant invented, possessed, or contemplated that alternative as part of the invention. The same passage may help establish that an alternative was excluded for claim-construction purposes while failing to support later claims seeking to encompass it. If the applicant may want broader continuation claims, the Detailed Description should affirmatively disclose the common inventive principle, intermediate levels of generality, and alternative implementations. Those alternatives should be presented as embodiments of the applicant’s solution, not left solely in a prior-art discussion that labels them defective.

VII. A Comparative Check: The EPO’s Requirement for Technical Context

The U.S. discussion establishes both the affirmative uses and the domestic costs of a reference-specific Background. European practice supplies a useful drafting analogue, although not doctrinal equivalence. The Implementing Regulations to the European Patent Convention provide that the description “shall” indicate known background art useful for understanding the invention, preparing the European search report, and examining the application, and should preferably cite the documents reflecting that art. The description must also make the claimed technical problem and solution understandable and state advantageous effects with reference to the background art. Implementing Regulations to the Convention on the Grant of European Patents r. 42(1)(b)–(c) (Eur. Pat. Off. 2026) (“EPC Implementing Regulations”). The rule concerns the description, not a separately captioned “Background,” and does not impose an exhaustive-search duty. See Case T 2321/08, Catchwords 1–2 (EPO Tech. Bd. App. May 11, 2009) (explaining that Rule 42(1)(b) imposes no “stringent obligation” to cite all known art at filing and permits later amendment).

The April 2026 EPO Guidelines state that an applicant should, in principle, cite the closest prior art known at filing. If examination identifies a more relevant document, the examining division may require a reference and brief factual summary to place the invention in proper perspective; new advantage statements remain constrained by the prohibition on added matter. European Patent Office, Guidelines for Examination in the European Patent Office pt. F, ch. II, § 4.3 (Apr. 2026) [hereinafter EPO Guidelines]; see also Convention on the Grant of European Patents art. 123(2), Oct. 5, 1973, 1065 U.N.T.S. 199. The Guidelines also caution against disparaging a particular prior product or process or presenting either the prior art or the invention misleadingly. EPO Guidelines, pt. F, ch. II, § 4.5. And when identifying a suitable starting point for inventive-step analysis, the EPO generally treats the applicant’s acknowledgement of known art as correct unless a mistake is identified. Id. pt. G, ch. VII, § 5.1.

The EPO analogy therefore confirms a drafting proposition rather than a shared legal rule. Both systems reward a description that accurately locates the claimed mechanism within its technical setting. The consequences differ: European practice uses that account principally to support search, examination, amendment, and problem-solution analysis, while U.S. law may also give it consequences for eligibility, AAPA, claim construction, prosecution disclaimer, specification disavowal, written description, and infringement.

VIII. Conclusion

A prior-art search should be part of specification design, not merely a prefiling patentability screen. Addressing each framing reference with precision can create a filing-date account of the technical problem, claimed mechanism, and resulting improvement. That account may support a § 101 position, preserve claim meaning against later art, and permit the patent owner to show in a later validity challenge that a petitioner's reference falls on the conventional side of an intentionally drawn boundary and therefore lacks the distinguishing claim limitation.

European practice makes the contextual function of the description explicit: useful known background art should be identified, the documents reflecting it should preferably be cited, and the technical problem, solution, and advantageous effects should be understandable against that background. That requirement reinforces rather than displaces the article’s central recommendation—search first, then draft the comparison with precision.

The benefit is inseparable from its price. The same words may narrow literal scope and equivalents, bind later infringement theories, follow a shared specification into continuations, or fail to support broader claims. The goal is not aggressive disparagement, but a durable record built from accurate comparisons, claim-tethered explanations, affirmative disclosure of alternatives, and deliberate acceptance of the resulting boundaries. The best Background says no more—and no less—than the applicant is prepared to defend for the life of the patent. Cooperative I shows the benefit of that discipline; Cooperative II shows its price.