The argument that does not yet exist

In Minority Report, a detective compares two visions of a woman being drowned. They appear to show the same event, and the system has a category for repeated visions: an echo, which technicians are trained to disregard. But he notices that the ripples on the water move in opposite directions. “Watch the ripples,” he says. A detail in the background makes it possible to question whether the images belong to the category that had made one of them disposable.¹

What interests me about the scene is the relationship between the detail and the explanation. A discrepancy becomes consequential because someone understands which assumption it could disturb. Without that connection, it is just another difference in a complicated record. With it, the investigation has a new direction.

In difficult legal work, the decisive obstacle can be a proposition that has become so familiar that everyone has started reasoning from it. A signed agreement is assumed to carry the consequences of an ordinary contract. A hearing is said to have ended. An earlier decision is treated as having settled the point. Each description may contain a conclusion that still needs to be established, and a considerable amount of apparently compelling argument can accumulate on top of it.

In CSFK v HWH (No 2) [2026] 3 HKLRD 841, the Court of Appeal had to decide a novel question about whether a signed settlement agreement produced through a protected, court-led negotiation could be used at a later trial after one party withdrew consent. The agreement had been signed shortly after adjournment, following the judge’s encouragement to continue negotiating that afternoon. Resolving its admissibility required examining whether familiar rules governing completed settlements could be applied within the distinct process that had produced it.²

An absence of a directly applicable precedent can make this work feel like an invitation to invent an answer. I think it demands a more exacting kind of invention. We need to identify what the existing law has already committed itself to, determine where its application remains open, and construct a conclusion that can survive scrutiny from someone who has good reasons to resist it. The contribution may lie in a new relationship between established propositions, or in recognizing that a supposed conflict disappears once the propositions are stated accurately.

A strong novel argument makes an unfamiliar conclusion reachable through reasons the decision-maker can inspect, challenge, and use again. Its originality matters because it reveals a defensible route through a problem, rather than merely a more ingenious way to insist on the desired result.

1. Locate the uncertainty precisely

Before looking for a new argument, I want to know what kind of uncertainty I am facing. A missing fact, an ambiguous provision, an unsettled boundary between two doctrines, and a decision that appears adverse but addressed another question require different work. Additional research helps only to the extent that it reaches the uncertainty that actually controls the result.

In this appeal, the question of admissibility had been left open by an earlier appellate decision. The lower court nevertheless understood a passage in that decision as requiring the document to be admitted. A sentence reserving a substantive issue for the trial judge had acquired a further meaning about the evidence that judge would be permitted to consider.³

That is a useful place to begin an inquiry because it forces attention to what a proposition does. A court may identify who must decide an issue without determining how it must be decided or what material may be used. To establish the reach of the earlier passage, it was necessary to recover the question the court had been answering, the argument it was rejecting, and the surrounding explanation. The authority was important; its importance made it more necessary to read it accurately.

This is also why searching for a favorable sentence can become counterproductive. A passage may resemble the proposition we need while carrying an assumption that the present dispute puts in issue. Another decision may use unfamiliar language while explaining the very relationship on which our argument depends. I would rather understand that relationship than collect a longer list of quotations that seem to point in the right direction.

The next step is to work backward from the order sought. Which propositions must the court accept to make that order, and which would merely make the story more satisfying? If a proposed route depends on three conditions, support for two of them leaves the third as an obstacle. If there are alternative routes, they should remain distinguishable, so that abandoning one does not silently remove a premise needed by another.

For an evidential dispute, this may require separating the source of a protection, the conditions that make it apply, the particular use being opposed, and any asserted exception or waiver. These questions interact, but they do not collapse into a single impression that the document ought to be protected. A persuasive account of its origin cannot answer an established exception; equally, debating exceptions before establishing the applicable regime may send the research toward the wrong body of law.

I find it useful to state the opposing route with equal precision. What has to be true for the other side to obtain its order? Where does its evidence end and its inference begin? Which of its propositions can be accepted without conceding the result? This last question can be particularly valuable. Agreeing with an accurate general rule removes a needless dispute and allows attention to concentrate on whether the rule does the work being claimed for it.

The result is a more useful description of the unknown. We may know the text, the events, and several related principles, while lacking an authoritative answer about how they fit together in these circumstances. That is already a substantial foundation. The task is to build the missing connection without pretending that it has been established in advance.

2. Take the framework apart

A legal category compresses information. Calling an exchange a negotiation, an agreement, or an admission helps us bring prior learning to it quickly. The danger is that we may import the consequences associated with the category before examining whether the conditions for those consequences are present.

The central comparison in this case was between ordinary without-prejudice protection and the rules governing Financial Dispute Resolution, or FDR, a court-led process used to attempt settlement of financial issues in family proceedings. Both encourage settlement, and both protect communications. But the Court identified important differences in where the protection came from, what the participants were required to do, and how agreed terms acquired legal effect.⁴

The analytical work becomes clearer when the comparison is conducted element by element. Does protection depend on the character and context of a voluntary communication, or arise from participation in a defined forum? What must the parties disclose within that forum? Who directs the process? Which subsequent uses of the material are prohibited? What further act is needed before agreed terms become enforceable?

In the FDR framework considered by the Court, parties were required to attend, use their best endeavours to reach agreement, and put their offers and responses before the FDR judge without withholding them on grounds of privilege. A specific evidential rule excluded evidence of things said or admissions made in the hearing, subject to its stated exception for an offence committed there. Continuing consent and judicial approval were also material to the conversion of agreed terms into an enforceable order.⁴

These features explain why the comparison matters. If a process requires people to disclose their negotiating positions within it, the protections governing subsequent use perform work that cannot be understood solely by examining a label on the resulting document. If agreed terms require an additional exercise of judicial authority, a signature does not answer every question about what follows. The features connect; they are parts of an arrangement that has to function coherently.

First-principles thinking in law starts with that arrangement. The governing text and binding decisions constrain what we can plausibly argue. Purpose helps us examine their application, but it cannot be chosen freely to produce whichever result we prefer. An argument gains strength when the proposed reading explains why several provisions exist together and how they continue to operate, including the provisions that initially seem inconvenient.

This discipline also improves the use of analogy. A commercial settlement decision may contain an important explanation of why completed settlements can be proved. The question is what justifies carrying that consequence into another setting. Sharing the word “settlement” establishes a point of comparison; it leaves the conditions of transfer to be demonstrated. A useful analogy identifies the relationship that matters and checks whether it survives the move.

We must apply the same standard to distinctions favorable to our own side. Declaring a process unique can become a way to avoid explaining it. The fact that two regimes differ somewhere does not show that they differ on the issue before the court. The relevant difference must alter a premise needed for the proposed conclusion. Otherwise, the comparison produces an impressive table and very little legal progress.

One revealing test is to remove the disputed label and describe the events. A draft was produced during a judge-led negotiation. The judge adjourned and encouraged the parties to continue that afternoon. They remained in the building and signed the document after further discussion. Those facts establish a sequence. Calling its final stage “outside the process” adds an interpretation of the sequence, which must be justified against the rule governing it.⁵

The wording of the protection creates a further distinction. It addresses evidence of things said or admissions made in the course of the hearing. The time at which a document is completed and the time of the communications it records are therefore different questions. A signature added later cannot, by itself, establish that the document contains no evidence of exchanges that occurred earlier. The Court’s treatment of the typed terms and handwritten changes examines precisely that relationship between the instrument and the negotiation it records.⁶

This is elemental analysis at the level of a sentence. We identify the object of the rule, the event that brings it into operation, and the consequence it directs, then ask whether the proposed interpretation has substituted another condition. The inquiry still requires evidence connecting the document to the protected exchanges. It does, however, expose why concentrating on the moment of signature can leave the actual language of the prohibition unanswered.

Once the label is unpacked, the question changes. We can examine both the continuity of the protected process and what the document would reveal about it. That formulation preserves the facts while exposing the premise that the earlier formulation had hidden.

3. Follow public policy into a person’s decisions

Public policy is most useful when it explains a mechanism. Saying that an interpretation would undermine settlement identifies a concern, but leaves the reader to supply the steps between the interpretation and the harm. A serious argument should make those steps available for examination.

The familiar policy behind protection for settlement discussions concerns what people are willing to say while an outcome remains uncertain. As the Supreme Court explained in Oceanbulk, the prospect of subsequent reliance on an admission affects the willingness to speak frankly and, through that, the prospect of reaching agreement. The rule influences the negotiation before a later court ever decides whether to admit a particular statement.⁷

Apply that reasoning to a court-directed break for further discussion. If protection were to disappear whenever the judge left the room, a participant would have to reassess the status of each statement while continuing the very activity the court had encouraged. Their adviser would need to guard against language that might later be separated from its negotiating context. A concession made to discover whether agreement is possible would carry another possible use at trial. The rational response could be to say less, qualify more, or refuse to explore a position openly.

The Court’s treatment of ordinary stand-downs made that consequence concrete. FDR hearings commonly pause while parties negotiate outside the courtroom without the judge present. An interpretation that made physical presence decisive would place protection in doubt during a normal part of the process itself. In the circumstances of the appeal, the Court rejected a mechanistic cutoff at the time of adjournment and examined the continuity of the negotiations instead.⁵

This is the reasoning behind the sword of Damocles image. In the old account, a sword suspended by a horsehair prevents Damocles from enjoying the apparent comfort around him. Its effect does not wait for it to fall. Awareness of the danger changes his experience while the danger remains unrealized.⁸

For a negotiating party, the relevant threat is that a provisional concession could later be used against them if negotiations fail. The argument is about how that possibility shapes conduct now. A participant may withhold an offer that could have led to agreement because making it introduces a risk the process had appeared to remove. The damage can consist of conversations that never happen, rather than a measurable collection of concessions later admitted in evidence.

That gives the metaphor a precise job. It makes a prospective behavioral effect intelligible. It does not establish the legal protection, prove the extent of any empirical effect, or answer the exception in the governing text. Those parts of the argument need their own foundations. The image should make the causal explanation easier to see after the relevant circumstances have been established.

I think this is a general method for constructing public-policy arguments. Identify the person whose behavior matters, the choice they face, the information and risks available to them at that moment, and how the competing interpretations would change that choice. Then connect the predicted response to an objective the legal framework actually recognizes. Where evidence about behavior is available, use it. Where the argument is inferential, make the inference concrete enough to challenge.

The strongest counterargument also deserves this treatment. Confidentiality can restrict access to material that another party regards as important, while the ability to rely on a completed agreement can serve certainty and settlement. Those interests do not become frivolous because our client benefits from protection. The question is how the relevant framework accommodates them and whether the proposed use preserves that accommodation.

In this case, the Court distinguished the route by which FDR terms could become a consent order from the attempt to place this purported full settlement before the trial judge after continuing consent had disappeared. Its conclusion depended on the particular regime and document. That reasoning is more useful than an unrestricted proposition that every document lacking enforceability must also be inadmissible, which would travel far beyond the question decided.⁹

The policy argument therefore becomes strongest when it explains the system’s own commitments. We are asking what must remain true for people to participate in the way the process requires, and whether the competing interpretation would make ordinary participation self-defeating.

4. Give the best objection a real chance

The greatest danger after finding a promising explanation is that every subsequent investigation becomes an attempt to preserve it. The argument becomes more polished while the underlying question receives less attention.

In unfamiliar territory, I want a challenger who can reject the premise, not simply improve the answer to a predictable objection. That can be another member of the team or a deliberate change of position in my own analysis. Either way, the challenger needs the actual sources and difficult facts. Asking someone to assess a summary from which the awkward material has already been removed tests the summary’s persuasiveness more than the argument’s strength.

A serious objection in this appeal concerned duration. If protection continued after adjournment, did it follow that every communication between successive FDR hearings remained protected, regardless of the interval? A theory broad enough to cover the entire period until the process formally failed would raise questions that the same afternoon’s negotiations did not require the Court to resolve.¹⁰

The Court’s reasoning supplied a more bounded account. The draft already existed; the judge encouraged the continuation; the parties remained in the building; agreement followed within hours; and there had been no material change altering the nature of the negotiations. Those circumstances allowed the issue before the Court to be decided without announcing unlimited protection between hearings.¹⁰

There is a substantial lesson in that limitation. A narrower proposition can demand more intellectual work because we have to identify why the present case falls within it. Simply asserting that everything remains protected avoids that task. Explaining continuity requires an account of the events that sustain it and an openness to circumstances that would require another analysis.

Counterfactuals help test the account. What if fresh information materially changed the negotiations? What if a separate exchange began much later? What if the document recorded only a limited agreement and reserved identified matters? These questions should locate the boundary of the reasoning, rather than produce invented answers to every possible future case. We need to know which change disturbs the present argument and why.

A useful challenge can also turn an apparently adverse fact into a more informative one. Handwritten changes after adjournment might initially appear to separate a final document from the draft made in the hearing. But further changes may also show that bargaining continued. Their legal significance depends on the account of the process, not on whether they help or hurt the story at first glance. The Court rejected the attempted division between typed terms and later amendments, observing that additions could disclose admissions and deletions could disclose terms a party had refused to accept.¹¹

This is more exacting than finding a favorable interpretation of everything. The alternative reading must explain the evidence, and it must survive the possibility that the facts really do support a rupture. The discipline is to keep the observation stable while comparing the inferences available from it.

I also want to know whether the rule remains defensible when the parties exchange positions. Would I accept the same protection if it prevented my client from relying on the other side’s concession? A policy argument asks the court to adopt a reason that extends beyond the immediate beneficiary. It becomes difficult to defend when its logic changes with the name at the top of the brief.

The point of adversarial preparation is to learn which formulation deserves to reach the court. An objection that narrows the claim, changes the sequence, or causes a route to be abandoned has done useful work. A session that ends with every original proposition intact may have produced reassurance when we needed discovery.

5. Make the decisive connection easy to inspect

Once an argument has survived that examination, presentation becomes a problem of directing attention. The reader needs to see which proposition changes the outcome, what supports it, and how it answers the difficulty they are likely to encounter next.

This is where surgical persuasion differs from brevity for its own sake. Removing words can conceal the connection the court needs. Adding authorities can bury it. The appropriate level of detail depends on the inference being asked of the reader; a short proposition about a familiar rule may need little explanation, while a disputed transfer between two regimes may deserve several paragraphs.

The order of presentation should follow the dependencies of the argument. Establish the governing framework before importing consequences associated with another one. Explain the status of the material before relying on what could be done with it. Show the use being challenged before debating whether the protection would prevent every conceivable use. That sequence reduces the opportunity for a disputed conclusion to enter unnoticed as a premise.

The problem is especially visible in an assertion that an agreement must be admissible because it must be capable of enforcement. Before that inference can assist, the relevant legal status and route to enforcement have to be established. Otherwise, an expected consequence is being used to prove the condition on which the consequence depends. The reasoning should expose that dependency rather than answer the assertion solely with another broad assertion about privilege.

Purpose can perform equally important work when a fact is undisputed. The party had referred to the document in earlier proceedings. What followed from that reference depended partly on why it had been made. Challenging an order founded on a document may require identifying the document; treating that necessary reference as surrender of the protection could make the challenge defeat itself. In addressing waiver, the Court examined that purpose alongside the express assertions of privilege in the record.¹²

Notice what this form of persuasion asks the reader to do. It accepts the observable act and investigates the legal inference said to follow. It does not require denying a reference everyone can see. The argument becomes more credible because it identifies the missing connection directly.

The same principle governs documentary precision. A sentence in an earlier decision may determine a substantive issue, reserve it, summarize a party’s submission, or state an assumption for a limited purpose. Quoting the sentence accurately while changing its role can be as misleading as quoting it inaccurately. A useful citation should let the reader verify both the words and the work they performed.

This is why I care about exact locations and surrounding passages. The reference is a means of reopening the analysis, especially when someone disagrees with it. If the critical proposition changes during revision, the sources supporting the earlier formulation must be checked again. A citation that was accurate for a narrower sentence does not automatically support a more ambitious replacement.

Clarity also requires restraint about the opponent. An argument may depend on a category error or an unsupported inference without its author being dishonest. Explaining the defect gives the court something it can decide. Speculating about motives adds another allegation that needs proof and may distract from the issue that already matters.

The sword of Damocles belongs in this account of precision. Its value lies in gathering an already demonstrated relationship into an image the reader can retain: anticipated exposure changes present bargaining. When a metaphor introduces a new allegation, exaggerates the consequence, or substitutes emotional force for the missing legal connection, it has stopped helping the argument.

6. Give the court a conclusion it can adopt

A novel argument is unfinished until it explains what the court should do with it. A broad account of institutional purpose can be persuasive while leaving the actual order uncertain. An elegant distinction can fail to resolve the independent ground on which the decision below also rests.

I therefore want to examine the completed argument from the disposition backward as well as from the sources forward. What is the smallest set of propositions that would justify the relief sought? Which alternative grounds require an answer even if the central point succeeds? What remains for another hearing, another decision-maker, or a case with different facts?

In CSFK, the Court held that the first ground alone was sufficient to allow the appeal, then addressed the remaining grounds and directed that the Memorandum be inadmissible at trial. The result was substantial, but its precision matters as much as its breadth. It resolved the evidential question; it did not resolve the parties’ underlying financial dispute.¹³

The relationship between those grounds is more instructive than the number of them. Some depended on the same analysis of the FDR framework. Others concerned the reach of the earlier decision and the alleged waiver. Treating every successful ground as a separate validation of an entirely independent theory would obscure how the reasoning actually fitted together.

A winning result also requires discipline afterward. The court may accept a narrower route than the one advanced, decline to rely on a passage we considered important, or explain the point differently. In this judgment, for example, the Court observed that the procedural provision cited beneath a broad textbook statement about communications before, during, and after FDR did not itself address communications after the hearing. That caution belongs in what we learn from the decision, alongside the favorable result.¹⁴

I regard that as an essential part of developing craft. If success is taken to endorse every proposition advanced, the next argument may inherit the weakest parts of the last one. What should survive is the reasoning that the sources and decision actually support, together with an understanding of the questions still open.

There is a similar obligation to the client before the outcome is known. An unfamiliar route may deserve serious pursuit while remaining uncertain. We should be able to explain its dependencies, the strongest reason it might fail, and the practical value of the order it seeks. The desire to establish an important point cannot make the client’s time, cost, or alternatives disappear. The Court’s closing observation about the long course of this dispute and the value of proportionate resolution is a reminder that an important evidential victory remains one part of a person’s life.¹⁵

More room to investigate

I am interested in how technology could make this discipline available to more people. A lawyer working through a difficult classification may need to compare several regimes, reconstruct a document’s history, trace the context of an authority, and test a sequence of counterarguments. Assistance with those tasks could increase the number of plausible routes a small team has the capacity to examine.

The demanding part is deciding what counts as useful assistance. A system can return genuine authorities and still apply them through the wrong framework. It can produce a convincing objection that attacks a proposition no one needs to establish. It can make two drafts agree by removing the qualification that made one of them accurate. Evaluating the polish of the final text would miss these failures.

Research on reflective legal argument generation offers a bounded example of a better approach. Li Zhang and Kevin Ashley separate scrutiny of factual support from rhetorical refinement, including whether an argument should be abandoned when the supplied factors cannot support it. Their experiments use predefined factors. Discovering which distinctions matter in an unfamiliar record remains a further challenge, and it is precisely where much of the work described here begins.¹⁶

For that harder task, I would test a system on whether it can identify the assumption beneath the question, recover the relevant rule without losing its conditions, and explain what evidence would distinguish a promising interpretation from a convenient one. A useful challenger must be able to change the proposed answer. A useful research assistant must preserve the disagreement when the sources do not resolve it.

There is no reason to make the aim merely faster drafting. Better assistance could let a lawyer investigate a route that would otherwise remain unexplored, or identify early that an attractive argument cannot support the relief the client needs. Both outcomes can improve the work. The more ambitious possibility is that unfamiliar cases receive deeper examination because that examination becomes less expensive.

What we should carry into our own thinking

Before committing to a difficult decision, I would write down the conclusion I am leaning toward and the few things that must be true for it to hold. Beside each, I would identify the supporting evidence and anything I have supplied through interpretation. The point is to find where the conclusion is most vulnerable. Ten well-supported observations may still depend on one assumption that none of them establishes, and investigating that assumption could be more valuable than collecting ten more observations.

We should also decide what would change our minds before we become too invested in defending the answer. What would we expect to find if the strongest alternative explanation were correct? Which fact, if different, would require us to abandon or narrow our position? Those questions give the next investigation a purpose. They also make it harder to dismiss inconvenient evidence simply because it arrived after we had settled on a persuasive account.

When a premise changes, we need to follow that change through the rest of the argument. A corrected classification may alter which rule applies, which evidence matters, and what action is available. It is easy to repair the sentence where the mistake was found while leaving its consequences elsewhere untouched. I would keep the connections between the important propositions visible enough that a challenge to one prompts us to reopen the others that rely on it. This is especially useful when several people, or several AI agents, are developing different parts of the same analysis.

Having taken the problem apart, we should put it back together from the perspective of the people who have to live within it. What does the proposed rule encourage them to reveal, conceal, attempt, or avoid? Who benefits from the outcome we are measuring, and who bears a cost that the measurement leaves out? Their responses may change the information available to us next time. A process that discourages people from reporting a difficulty can eventually appear successful because the evidence of its failure has become harder to obtain. We should examine whether our own arrangement could produce that result.

Consider a hypothetical product team building an AI assistant to reduce customer-support escalations. Its success measure assumes that fewer referrals to a person mean more problems have been resolved. A customer who gives up trying to reach support would make the measure look better, too. The first test should therefore examine what happened to the underlying problem, including repeat contact and unresolved cases, and whether the customer could still obtain help when the assistant failed. The investigation might reveal that contradictory return instructions generate many of the requests, making a change to the original purchase journey more valuable than further investment in the assistant.

The technical design follows from the same inquiry. If the assistant says a refund has been issued, the team needs evidence of the corresponding payment action. A pending transaction needs a visible status, a failed one needs a recovery path, and a retry must not create a second refund. The promise now reaches across the interface, payment service, and support operation. Systems thinking helps us identify the relationships that must hold for the customer’s problem to be resolved, including failures that would remain invisible if each component were assessed separately.

We should then choose an action proportionate to what remains uncertain. A small, reversible test may be enough to distinguish two explanations; a consequential commitment may require further evidence or a narrower scope. Someone needs responsibility for making that choice and for changing the work when the answer arrives. An objection has little practical value if the requirements, resources, and deadline remain fixed regardless of what it reveals. Leaders have to make revision possible, including when the original plan was their own.

Over time, I would preserve a record of the important corrections and why they changed the decision. That record should tell the next person what to investigate when a similar problem appears, while keeping the conditions of the earlier conclusion visible. We can build on experience without requiring an unfamiliar situation to fit the last answer that succeeded.

The detective in Minority Report notices the ripples because he is prepared to reconsider what the images represent. The corresponding discovery may be a distinction within an established framework, a consequence no one has followed through, or an assumption whose support has been overstated. The opportunity begins in that uncertainty. A new argument earns its place when we can turn the discovery into a reason another person can examine and, if it survives, act upon.

---

Notes and sources

Footnotes

1. Scott Frank, Minority Report, revised screenplay, 16 May 2001, the scene in which Witwer compares the two visions of Anne Lively’s drowning. Screenplay. The quoted words are Witwer’s. Back to reference 1

2. CSFK v HWH (No 2) [2026] 3 HKLRD 841; [2026] HKCA 1036, CACV 62/2025, judgment delivered 6 July 2026. Reported judgment, §§11, 20–22, 51–54 and 60. The signed agreement’s admissibility was a novel issue left unresolved by the earlier appeal. The Court distinguished an agreement reached within FDR from a contract independently enforceable at law. Back to reference 2

3. CSFK, §§51–54. The earlier decision had not determined the Memorandum’s admissibility. The Court explained the scope of the earlier passage in the context of the issue then being addressed. Back to reference 3

4. CSFK, §§14–22 and 23–33. The Court distinguishes the FDR regime from ordinary without-prejudice negotiations, including the source of confidentiality, compulsory obligations, the evidential bar and its stated exception, and the role of judicial approval. These are propositions about the framework examined in the judgment, not a universal description of all settlement procedures. Back to reference 4 Back to reference 4-2

5. CSFK, §§4 and 35–41. The Court examined the existing draft, the judge’s encouragement, same-afternoon continuation, absence of material intervening change, and ordinary stand-down practice. Back to reference 5 Back to reference 5-2

6. CSFK, §§15, 35, and 42. The evidential bar concerns things said or admissions made in the protected hearing. The distinction between when a document is completed and what earlier exchanges it evidences is an analytical reading of those provisions and findings. Back to reference 6

7. Oceanbulk Shipping & Trading SA v TMT Asia Ltd [2010] UKSC 44; [2011] 1 AC 662, especially §§19 and 23–29. Supreme Court case page · Judgment. The decision also recognizes a defined interpretation exception, §§40–48. Its policy reasoning is relevant here; the exception is not transplanted into the distinct FDR regime. Back to reference 7

8. Cicero, Tusculan Disputations, Book V, the Damocles passage traditionally numbered §§61–62, appearing in chapter XXI in the linked translation. Public-domain text. The discussion of anticipated litigation exposure is the author’s application of the image. The image is not attributed to the Court in CSFK. Back to reference 8

9. CSFK, §§20–22 and 47–50. The Court specifically noted that the Memorandum purported to be a full agreement, rather than a partial agreement leaving identified issues for trial. Back to reference 9

10. CSFK, §§19 and 36–40. The judgment does not establish that every communication between FDR hearings, however remote or different in nature, is protected. The alternative factual situations posed in the essay are questions for testing the reasoning, not additional holdings. Back to reference 10 Back to reference 10-2

11. CSFK, §42. The Court rejected the attempted separation between typed terms and handwritten amendments and explained the possible significance of additions and deletions. Back to reference 11

12. CSFK, §§55–59. The Court considered express assertions of privilege and the purposes for which the Memorandum had been referred to in the identified proceedings and correspondence. Back to reference 12

13. CSFK, §§43–60. Ground 1 was sufficient to allow the appeal; the Court went on to address the other grounds and directed that the Memorandum be inadmissible at the ancillary-relief trial. Back to reference 13

14. CSFK, §19. The Court examined the procedural source cited for the textbook proposition rather than treating that proposition as resolving the disputed temporal boundary. Back to reference 14

15. CSFK, §63 and its continuation, urging pragmatism, reasonable proportion, procedural economy, and consideration of an amicable resolution. Back to reference 15

16. Li Zhang and Kevin D. Ashley, “Mitigating Manipulation and Enhancing Persuasion: A Reflective Multi-Agent Approach for Legal Argument Generation” (2025), arXiv:2506.02992v2. Paper. The study’s three-part argument task uses predefined factors and evaluates grounding, factor use, and abstention. The proposed tests for unfamiliar arguments in this essay extend beyond those experiments. Back to reference 16