The Examiner Is Not Opposing Counsel: Patent Prosecution as Mediative Advocacy

Executive Summary: This article argues that examiner-level patent prosecution is better understood as a form of mediative advocacy than as the adversarial process for which lawyers are traditionally trained. Unlike litigation, prosecution involves no opposing claimant, the asserted legal right remains mutable, and both the examiner and applicant ultimately seek a correct determination of patentability, although they approach that objective from different institutional roles. Effective prosecution therefore relies on skills associated with mediation: identifying interests beneath stated positions, diagnosing the actual source of disagreement, testing assumptions against the specification, prior art, governing law, and commercial objectives, generating supported claim alternatives, and recognizing when continued discussion should give way to petition or appeal. Grounded in current USPTO interview guidance and empirical research associating examiner interviews with fewer Office Actions to disposition, the article translates this theory into practical recommendations for preparing for, conducting, and memorializing interviews. It also acknowledges the model’s limits, including examiner production pressures, the duty of candor, prosecution-history risks, and the danger of unnecessary narrowing. Ultimately, the article contends that the best patent prosecutors know both how to solve claim-drafting problems collaboratively and when a dispute requires formal adjudication.

I. Introduction

The most revealing instruction in the United States Patent and Trademark Office’s May 2026 Interview Best Practices is not procedural. It is professional. Examiners and applicants “should not conduct interviews as adversaries,” the guidance explains, because “[c]ollaboration is more effective than argument.” U.S. Pat. & Trademark Off., Interview Best Practices 5 (May 2026). Those sentences challenge an instinct built into most lawyers long before they prosecute their first patent application.

Legal education is organized largely around disputes. The facts have occurred, the parties’ interests have diverged, and the operative legal instruments are fixed. The lawyer learns to identify the opponent’s error, defend a position, preserve every argument, and persuade a neutral decisionmaker to declare a winner. It is natural to carry that training into patent prosecution. An Office Action can look like an opposing brief. A rejection can feel like an adverse ruling. A response can become an effort to defeat every sentence the examiner wrote.

But ordinary examination is not litigation in miniature. There is no adverse claimant before the examiner. The claims are not fixed; they are still being formed. The examiner is not opposing counsel but a public decisionmaker charged with determining whether the applicant is entitled to a patent. And the central work of prosecution is often not winning an argument about an existing legal right. It is identifying language that will create a lawful, defensible, and commercially useful right.

Patent prosecution therefore is better understood as an adjudicative system conducted, in its ordinary pre-appeal phase, through mediative techniques. The analogy is functional rather than institutional. An examiner is not the impartial third-party facilitator contemplated by conventional mediation and cannot compromise the statutory requirements for patentability. See Model Standards of Conduct for Mediators pmbl. & stds. I–II (Am. Arb. Ass’n, Am. Bar Ass’n & Ass’n for Conflict Resol. 2005). Yet effective prosecution depends on many of the same skills that make mediation effective: distinguishing interests from positions, diagnosing the real source of disagreement, testing assumptions against objective criteria, generating alternatives, and knowing when continued negotiation is inferior to adjudication. This approach may be called mediative advocacy. It does not ask the patent prosecutor to advocate less forcefully. It asks the prosecutor to choose the form of advocacy that fits the process.

II. The Wrong Mental Model

The adversarial model begins with opposing parties. Each side benefits if the other loses, and a separate decisionmaker resolves their dispute. Examiner-level prosecution lacks that structure. The examiner owns no competing right, represents no accused infringer, and obtains no institutional victory merely because an application is rejected. Congress instead directs the USPTO to examine the application and to issue a patent if “the applicant is entitled to a patent under the law.” 35 U.S.C. § 131 (2024). The examiner’s responsibility is thus neither to secure allowance nor to prevent it, but to make an independent determination of patentability.

The USPTO describes that responsibility in deliberately balanced terms. The Office has “both an obligation not to unjustly issue patents and an obligation not to unjustly deny patents,” and innovation is best served when an inventor receives “the scope of protection that is deserved.” U.S. Pat. & Trademark Off., Manual of Patent Examining Procedure § 2001.04 (9th ed., Rev. 01.2024, Nov. 2024) [hereinafter MPEP]. That is not the mandate of an adversary. It is the mandate of a gatekeeper.

The Court of Customs and Patent Appeals made the point more directly more than fifty years ago. “The ex parte prosecution and examination of a patent application must not be considered as an adversary proceeding,” the court explained in Norton v. Curtiss. 433 F.2d 779, 794 (C.C.P.A. 1970). Norton arose from allegations of misconduct rather than an ordinary disagreement over patentability, but its description of the institutional relationship remains important. Examination requires both fact-finding and adjudication, often under substantial time and informational constraints. Id. The Office therefore depends upon the applicant for an accurate account of the invention, relevant technical facts, and known information material to patentability.

That dependence reinforces why the ordinary-opponent analogy is incomplete. Although litigation counsel also owe duties of candor and truthfulness, Rule 56 imposes a prosecution-specific duty of candor and good faith on those associated with the application. The rule begins from the premise that a patent “is affected with a public interest.” 37 C.F.R. § 1.56(a) (2025). The Supreme Court similarly has emphasized the public’s paramount interest in patents that arise free from fraud and remain confined to their legitimate scope. Precision Instrument Mfg. Co. v. Auto. Maint. Mach. Co., 324 U.S. 806, 816 (1945). And the Court has said that the relationship between patent practitioners and the Office requires the “highest degree of candor and good faith.” Kingsland v. Dorsey, 338 U.S. 318, 319 (1949) (per curiam).

Those authorities are invoked here for the professional and institutional relationship they describe, not as statements of the modern litigation standard for inequitable conduct. The Federal Circuit’s en banc decision in Therasense substantially tightened that doctrine. In ordinary nondisclosure cases, an accused infringer generally must prove both specific intent to deceive and but-for materiality, and the two elements must be established separately. Therasense, Inc. v. Becton, Dickinson & Co., 649 F.3d 1276, 1287, 1290–91 (Fed. Cir. 2011) (en banc).

None of this eliminates advocacy. Applicants may—and should—challenge unsupported findings, incorrect claim constructions, legally deficient rationales, and improper procedural shortcuts. But it does reject the idea that the examiner should be approached as an opponent whose concessions are extracted and whose errors are accumulated for later use. The better starting point is a shared, though not identical, objective: a correct determination of what the applicant may lawfully claim.

III. Two Responses to the Same Rejection

Consider an application claiming a battery controller that changes a charging protocol based on a “predicted degradation state.” The specification explains that the controller derives the prediction from historical thermal-cycle data and uses the result to reduce long-term battery damage. The examiner rejects the claim as obvious over two references. The first estimates a battery’s present state of health from current impedance and adjusts charging accordingly. The second predicts future degradation of industrial equipment from historical operating data. The examiner reasons that the second reference’s forecasting technique would improve the first reference’s battery controller.

A litigation-trained response may attack every component of the rejection. It may argue that the first reference does not make a prediction, that the second reference concerns a different problem, that the combination changes the first reference’s principle of operation, that the examiner has used hindsight, and that there would have been no reasonable expectation of success. Some or all of those arguments may be correct. The problem is not their legal availability. The problem is treating the written rationale as if it necessarily revealed the entire disagreement.

Suppose the examiner maps “degradation state” to the first reference’s state-of-health estimate and “predicted” to the second reference’s forecasting method. The examiner’s underlying construction is broader still: “predicted degradation state,” in the examiner’s view, encompasses an estimated condition derived from a present measurement and does not necessarily require a future state derived from historical battery behavior. A twenty-page response challenging the combination may leave that interpretive disagreement untouched. The next Office Action may retain the rejection, perhaps with a different reference or a more detailed explanation, and both participants may conclude that the other simply failed to listen.

A mediative approach begins elsewhere. Counsel first determines what the client actually needs. If the commercial implementation depends on historical thermal-cycle data, then abstract coverage of every possible prediction may be less important than claims that competitors cannot avoid while using that data. Counsel then asks the examiner to identify precisely what in the first reference satisfies the disputed limitation and whether the rejection depends on reading a present-state estimate as the claimed prediction. That question may reveal the interpretive gap.

The participants can then test alternatives against objective criteria. The applicant may maintain that the existing language already excludes present-state measurement. It may also present a clarification requiring the prediction to be generated from a time-ordered history of thermal-cycle measurements. A second option might recite that the predicted state represents degradation during a future charging interval. A dependent claim may capture a narrower commercial embodiment. If the broader interpretation remains important, the applicant may preserve it through argument or a continuation rather than forcing the entire portfolio strategy into a single claim.

This is not surrender disguised as cooperation. Counsel should reject language that adds no patentable distinction, invites an easy design-around, creates a proof problem, lacks written-description support, or unnecessarily limits related applications. Nor should counsel amend merely because the examiner asks. The difference is that counsel evaluates possible amendments after identifying the actual obstacle and the client’s actual interests. The inquiry shifts from “How do I defeat this Office Action?” to “What issue prevents this claim from being both allowable and valuable, and what is the best lawful way to resolve it?”

The hypothetical also exposes why a nominal victory can be misleading. Persuading the examiner to withdraw a particular rationale is not success if the next search produces the same rejection in a different form. Conversely, accepting a narrow claim is not success merely because a Notice of Allowance follows. Prosecution outcomes are multidimensional. Claim scope, detectability, design-around risk, validity, speed, cost, patent term, foreign consequences, continuation options, and the clarity of the prosecution record all matter. Counsel using this model evaluates the package rather than counting argumentative wins.

IV. The Mutable Right

The deepest difference between litigation and prosecution is that the legal right under discussion is not yet fixed. Litigation ordinarily applies law to past conduct or interprets an instrument already in existence. The disputed contract has been signed. The accused product has been sold. The asserted patent has issued. Advocacy is primarily retrospective: the parties contest what happened and what an existing text means.

Patent prosecution is constitutive. The claims may be amended, canceled, rewritten, divided, or pursued through related applications. The process does not merely determine the meaning of an existing right; it helps create the instrument that will define the right. Claim language is therefore both the subject of disagreement and the principal tool for resolving it.

That mutability changes the nature of advocacy. If an examiner interprets a term more broadly than the applicant intended, the applicant can argue for the intended construction, revise the language to express it, or do both. If a reference discloses a broad functional result but not the sequence that makes the invention commercially useful, the applicant can focus the dispute on that sequence. If a claim combines several inventive concepts in a way that obscures the true distinction, the applicant can separate those concepts into different claims or applications. The available outcomes are not limited to “rejection sustained” or “rejection withdrawn.”

The Federal Circuit has recognized the interactive character of examination. In In re Oetiker, Judge Plager explained that the examiner must state objections clearly enough to give the applicant a fair opportunity to answer with evidence and argument. 977 F.2d 1443, 1449 (Fed. Cir. 1992) (Plager, J., concurring). The majority likewise emphasized that patentability ultimately must be determined on the record as a whole, not by treating the examiner’s prima facie showing as the final inquiry. Id. at 1445–46. The Office’s indefiniteness practice reflects the same developmental premise: because the record is still being formed and the applicant can amend, examination can require the applicant to clarify language before the uncertainty becomes embedded in an issued patent. See In re Packard, 751 F.3d 1307, 1311–13 (Fed. Cir. 2014); id. at 1325 (Plager, J., concurring).

The MPEP does not describe the examiner as a passive adjudicator waiting to accept or reject whatever positions the applicant chooses to present. When patentable subject matter appears to have been disclosed but imperfectly claimed, the examiner’s action should be constructive and, when possible, offer a definite suggestion for correction. MPEP §§ 706, 707.07(j). The examiner should identify presently allowable claims and may suggest how rejected claims could be amended to become allowable. MPEP § 707.07(d). An examiner’s identification of allowable subject matter may also justify an interview aimed at early agreement on allowable claims. MPEP § 707.07(j).

These instructions do not authorize bargaining around the statute. They recognize an interactive effort to identify claim language that accurately captures the disclosed invention while satisfying legal constraints. Responsibility for selecting and drafting claims that serve the applicant’s interests remains with applicant’s counsel. The process nevertheless resembles principled mediation because the participants can generate options, test them against neutral criteria, and refine the instrument that resolves the disagreement.

V. The Mediative Core of Prosecution

The analogy is most useful at the level of professional method. Four mediation practices map particularly well onto effective examiner-level advocacy: identifying interests beneath positions, reality testing, generating options, and evaluating the alternative to agreement.

A. Interests Rather Than Positions

A position is the outcome a participant states: “the reference discloses the limitation,” “the claim is patentable,” or “this amendment is necessary.” The underlying interest explains why that position matters. The examiner may be concerned that a functional term reads on every implementation found during the search, that the alleged distinction cannot be identified from the claim language, or that a proposed amendment will require a new search. The applicant may care principally about a competitor’s implementation, rapid issuance, avoiding a difficult-to-prove limitation, or preserving a broader theory elsewhere in the family.

Until those interests are identified, the participants may argue past one another. The applicant may spend pages proving that a reference does not disclose a feature the examiner did not regard as dispositive. The examiner may propose language that overcomes the art but eliminates the claim’s commercial value. An interview can expose both errors. Understanding the examiner’s concern does not concede that it is legally correct, just as explaining the client’s commercial interest does not make that interest a patentability criterion. It reveals the problem that any useful resolution must address.

B. Reality Testing

Mediation asks parties to test positions against legal, factual, and practical constraints. Prosecution does the same through the specification, prior art, claim language, technical evidence, and governing law. The examiner tests whether the proposed breadth can survive the art and statutory requirements. The applicant tests whether the examiner’s mapping and reasoning withstand careful examination. Each participant can expose an assumption that is difficult to see in serial written exchanges.

Reality testing also includes commercial consequences. An amendment that distinguishes the cited references may still fail if it describes an internal operation that cannot be detected in an accused product. A limitation may cover the present product but exclude the client’s planned architecture. A quick allowance may be less valuable than a narrower amendment coupled with a continuation preserving the broader dispute. Mediative advocacy insists that patentability and value be evaluated together, while recognizing that only patentability controls the examiner’s decision.

C. Option Generation

Litigation briefing often narrows toward a binary decision. Prosecution can expand the set of available solutions. Counsel may offer an argument-only path, a clarification that preserves intended scope, a commercially acceptable fallback, a dependent claim directed to the preferred embodiment, supporting evidence, or a continuation strategy. The examiner may identify an allowable feature, suggest different wording, or explain why a proposed formulation would require additional searching.

The objective is a zone of possible allowance: claim formulations that are supported by the disclosure, distinguish the prior art, satisfy the Patent Act, and retain enough commercial value to justify issuance. The zone is not a midpoint between competing preferences. Patentability cannot be split like money in a settlement. It is bounded by objective criteria and public consequences.

D. The Alternative to Agreement

Effective negotiators evaluate any proposed resolution against the best available alternative. Patent prosecutors should do the same. Depending on the posture of the case, the relevant alternative may be another written response, an RCE, a continuation, a pre-appeal conference, appeal to the Patent Trial and Appeal Board, or abandonment. Counsel should know before an interview whether the client prefers a narrower patent now, a broader appellate fight, parallel paths, or a portfolio-level compromise.

Appeal should not be used as a threat designed to pressure an examiner. It is a different decision process. When collaboration cannot resolve a genuine legal or factual impasse, appeal moves the dispute from iterative examination to adjudication. The credibility of that alternative protects against improvident narrowing while giving both participants a reason to identify which disagreements actually require a Board decision.

VI. Putting Mediative Advocacy into Practice

USPTO interview guidance is consistent with this account of prosecution as structured problem-solving. The MPEP explains that examiner-applicant discussions are often indispensable, should improve mutual understanding, and can bridge substantive gaps while identifying patentable subject matter. MPEP §§ 713, 713.01. The May 2026 guidance similarly instructs both participants to maintain an open mind, discuss the overarching inventive concept, and look for opportunities to resolve as many outstanding issues as possible. U.S. Pat. & Trademark Off., Interview Best Practices 3–5 (May 2026). Although that document is nonbinding best-practice guidance, its language is unusually direct evidence of how the Office believes an interview should function.

Empirical evidence supports treating interviews as structurally important rather than a matter of etiquette. S. Sean Tu examined approximately 1.1 million utility applications containing at least one examiner interview and compared their Office-action counts with the overall grant and abandonment ratios of the same examiners. Interviewed applications that issued reached allowance after approximately two Office Actions, compared with 3.6 across those examiners’ overall dockets; interviewed applications that were abandoned reached that result after approximately 2.5 actions, compared with 5.9. The study found no disproportionate increase in allowance rates. Because the comparison was observational and interview selection was not random, the results establish association rather than causation. They nevertheless support the proposition that interview practice is a consequential part of prosecution, not merely professional etiquette. S. Sean Tu, Patent Examination and Examiner Interviews, 49 Fla. St. U. L. Rev. Online 1, 10–16 (2021).

The MPEP also seeks participation by representatives able to address proposed amendments and, on the Office side, an examiner with authority to advance prosecution. MPEP § 713.05. That structure supports a productive interview, but any understanding remains subject to formal action and, where necessary, further search.

Mediative advocacy is not a romantic account of examination. It operates under production pressure, limited interview time, variable examiner experience, search burdens, docket incentives, asymmetric information, and supervisory practices that may differ across art units. The Office’s May 2026 guidance notes that the FY 2026 Examination Performance Appraisal Plan allocates one attribute hour per application or RCE for the interview and related work, including preparation; additional time may be authorized with supervisory approval. Interview Best Practices, supra, at 4. Those constraints make disciplined preparation more important: counsel should arrive with focused issues, supported claim alternatives, client authority, and a clear judgment about when further discussion will no longer advance prosecution.

A. Before the Interview: Define the Client’s Interests

Preparation begins with the client, not the examiner. Counsel should identify the coverage that is commercially essential, limitations that would permit easy design-arounds, features whose infringement can be detected, the value of speed, effects on related domestic and foreign applications, and the scope that should remain available in a continuation. Without that analysis, counsel cannot distinguish a useful amendment from a merely allowable one.

Counsel should also bring a small set of evaluated alternatives—an argument-only path, a clarification, a commercially acceptable fallback, and, where appropriate, a continuation strategy—so that claim scope is not improvised under time pressure.

The interview agenda should facilitate that work. It should identify the claims and references to be discussed, isolate the dispositive issues, and direct the examiner to relevant figures or specification passages. If counsel expects to discuss an amendment, proposed language or at least a focused description should ordinarily be provided in advance. The Office’s current guidance emphasizes that a sufficiently detailed agenda improves preparation and the prospect of agreement, while an unfocused agenda can restrict the interview’s benefit. Interview Best Practices, supra, at 2–4. The agenda should be short enough for meaningful review within the examiner’s time constraints.

B. During the Interview: Diagnose Before Rebutting

The interview should begin with concrete diagnostic questions: Where does the examiner locate each disputed limitation? What construction does the rejection apply? Would a proposed clarification change the mapping? A focused technology explanation or annotated figure may reveal that the examiner and applicant have used the same term to describe different operations. Conversely, the discussion may confirm that the disagreement is legal and cannot be resolved through technical clarification.

In the battery-controller example, the decisive question is not whether counsel can recite every defect in the obviousness rationale, but whether the examiner is treating a present state-of-health estimate as a future degradation prediction.

Questions often advance the inquiry better than declarations. Which passage supplies the limitation? Does the rejection depend on a particular construction? Would the mapping change if the claim expressly recited the disputed sequence? Does the examiner regard a dependent-claim feature as absent from the completed search? Such questions can expose the decisive issue without forcing either participant to defend an unnecessarily absolute position. The MPEP instructs examiners to identify issues and resolve differences as much as possible, while recognizing that proposed language may require further searching and consideration. MPEP § 713.01(IV).

Once the issue is understood, counsel can test exact alternatives. One proposal may preserve the claim and rely on argument. Another may clarify the intended interpretation without materially narrowing scope. A third may incorporate a commercially meaningful limitation from a dependent claim. Proposed language should be tested not only against the cited art but also against § 112 support, antecedent basis, interactions with other limitations, detectability, and likely construction after issuance.

The examiner also should be invited, where appropriate, to identify allowable subject matter. The MPEP instructs examiners to identify allowable claims and, when possible, suggest corrections where patentable subject matter has been imperfectly claimed. MPEP §§ 706, 707.07(d), 707.07(j). If the remaining changes are limited and authorized, the agreement may be implemented through an examiner’s amendment. MPEP § 1302.04. Counsel must nevertheless evaluate examiner-proposed language independently. An examiner’s willingness to allow a formulation does not establish that the resulting claim is worth obtaining.

C. After the Interview: Create a Reliable Record

A productive discussion must be converted into the formal record. If agreement was reached, the amendment should reflect its exact terms, identify support, and avoid unintended construction consequences. If no agreement was reached, the interview may still have narrowed several disputes to one, revealed a need for evidence, identified an RCE path, or confirmed that appeal is the appropriate next step.

Patent interviews do not create an off-the-record, mediation-style privilege. An unpublished application generally remains confidential under 35 U.S.C. § 122(a) (2024), but the confidentiality of the file depends on the application’s status—not on treating the interview as a private settlement discussion. USPTO action rests on the written record, and the substance of any merits discussion must be recorded whether or not the participants reached agreement. 37 C.F.R. §§ 1.2, 1.133(b) (2025); MPEP § 713.04. The summary should identify the claims, references, proposed amendments, principal arguments, and outcome accurately but without unnecessary characterizations. A casual statement about what the invention “requires,” what the prior art “concedes,” or why a feature was added may later influence claim construction or estoppel. See Phillips v. AWH Corp., 415 F.3d 1303, 1317 (Fed. Cir. 2005) (en banc); Festo Corp. v. Shoketsu Kinzoku Kogyo Kabushiki Co., 535 U.S. 722, 736–40 (2002). Mediative advocacy requires openness in the discussion and discipline in memorializing it.

VII. Knowing When to Change Modes

Collaboration is a default method, not an obligation to negotiate indefinitely. Appeal becomes appropriate when the material dispute has crystallized, the record is sufficiently developed, and further amendment would surrender commercially important scope without resolving the underlying error. The trigger may be a claim construction inconsistent with the specification, an obviousness rationale lacking adequate factual support, a refusal to credit material evidence, or an eligibility analysis that remains legally deficient after the relevant technology and authorities have been presented.

Changing modes does not always mean going to the Board. Rejections involving prior art or other patentability issues are appealable, while many procedural disputes are petitionable; the MPEP identifies restriction or election requirements, finality, nonentry of amendments, and abandonment as matters ordinarily resolved by petition. Interview-practice disputes likewise may call for the Director’s supervisory authority rather than an appeal brief. 37 C.F.R. § 1.181(a) (2025); MPEP § 1201. The prosecutor must identify not merely that discussion has failed, but which decisionmaker has authority to resolve the remaining issue.

Transitional procedures permit one final test before the applicant changes modes. An after-final interview or amendment may place the application in condition for allowance or sharpen the issues for appeal. 37 C.F.R. § 1.116 (2025); MPEP § 713.09. Unless the submission itself places the application in condition for allowance, however, it does not suspend the response period; the applicant still must file a timely notice of appeal or another sufficient response to avoid abandonment. 37 C.F.R. § 1.116(c) (2025); MPEP § 714.13. A pre-appeal brief request is different: it triggers an internal panel review, not another negotiation session. The panel may leave the application under appeal, reopen prosecution, or allow the existing claims. Because the request is limited to five pages and must present succinct, concise, focused arguments, it is best used when a legal or factual error is concise and facially identifiable from the existing record. MPEP § 1204.02. Using those procedures is a disciplined effort to determine whether adjudication is genuinely necessary.

Once appeal is chosen, the form of advocacy appropriately changes. The appeal brief should identify reversible error, develop the factual and legal record, preserve the client’s position, and request a decision from the Board. Mediative advocacy improves that brief by stripping away peripheral disputes and presenting the Board with the issue that could not be resolved before the examiner.

VIII. The Limits and Risks of the Analogy

Patent examination is not literally mediation. A mediator ordinarily facilitates a resolution but does not decide the merits. An examiner does. The examiner must independently determine whether every statutory requirement is satisfied and may finally reject claims despite the applicant’s disagreement. 35 U.S.C. §§ 131–132 (2024); 37 C.F.R. § 1.113(a) (2025). Nor does the examiner serve as a neutral between two private disputants. The examiner acts for the Office and must account for the public interest that an ordinary settlement mediator does not represent.

Patentability also is not negotiable in the sense that private litigation claims are negotiable. An examiner cannot allow an unpatentable claim because the applicant accepted a limitation elsewhere, and an applicant cannot treat an interview agreement as an enforceable settlement. Even apparent agreement over the art of record remains subject to an updated search that may identify new prior art. Interview Best Practices, supra, at 6.

The analogy also creates practical risks. Lawyers who equate cooperation with compromise may narrow claims too quickly. Counsel who arrives without a clear account of the client’s commercial interests may accept language that secures allowance but eliminates value. A conversational interview may encourage imprecise descriptions that later become prosecution-history evidence. And efforts to maintain rapport must never displace the duty to challenge unsupported rejections or preserve an issue for appeal.

These limitations sharpen rather than defeat the thesis. Mediative advocacy describes a professional method, not the examiner’s legal identity. It explains why curiosity, technical translation, option generation, and calibrated concession often work better than reflexive combat. It does not diminish the examiner’s decisionmaking responsibility, the applicant’s duty of candor, or counsel’s obligation to protect the client.

IX. What Patent-Law Training Should Teach

Because mediative advocacy describes a professional method rather than the examiner’s institutional identity, its most direct implications concern how prosecutors are trained. Adversarial skills remain essential, but they do not prepare counsel for an interaction that resolves many applications: a short technical discussion in which counsel must listen, diagnose a claim-construction problem, evaluate amendments in real time, and protect a portfolio that the examiner may never see.

Training should therefore include competencies more commonly associated with negotiation and mediation. Students and junior lawyers should learn to distinguish interests from positions, ask diagnostic questions, explain technology visually, generate multiple supported formulations, obtain client authority, evaluate design-arounds and proof, and summarize oral discussions without creating avoidable estoppel. The American Bar Association’s own description of professional roles recognizes the difference: an advocate asserts the client’s position within the adversary system, while a negotiator seeks an advantageous result consistent with honest dealings. Model Rules of Pro. Conduct pmbl. ¶ 2 (Am. Bar Ass’n 2020). Effective prosecution requires knowing when each role should predominate.

A realistic training exercise would therefore include more than drafting a written response. The student should receive an Office Action and confidential information about the client’s commercial priorities, prepare a short interview agenda, conduct a thirty-minute examiner interview, revise claims, and prepare the interview summary. Performance should be evaluated not only by doctrinal accuracy or oral persuasiveness, but by whether counsel discovered the examiner’s actual concern, protected valuable scope, generated lawful alternatives, and created a clean record.

X. Conclusion: Know When to Sit at the Table and When to Stand at the Podium

Ordinary prosecution is often closer to structured problem-solving than courtroom combat because the claim remains mutable and the examiner is not opposing counsel. Mediative advocacy therefore provides the better default: identify the client’s interests, diagnose the examiner’s actual concern, test assumptions against the record, and generate precise language within nonnegotiable legal limits. When that process cannot produce an acceptable result, counsel should preserve the dispute and invoke adjudication. The best prosecutor knows which problems can be drafted through—and which disputes must be decided.