Preliminary Issues and Unconscionability: Hybrid “Assumed Knowledge” Case Management Orders Are Unworkable

Case: Castleton v Post Office Ltd & Anor (Rev1) [2026] EWCA Civ 577
Court: England and Wales Court of Appeal (Civil Division)
Date: 12 May 2026
Judges: Peter Jackson LJ, Popplewell LJ, Zacaroli LJ (judgment of Zacaroli LJ)

1. Introduction

The appellant, a former sub-postmaster, brought a 2025 claim against Post Office Limited (“POL”) and Fujitsu Services Limited (“Fujitsu”) arising out of a 2007 civil judgment obtained against him after alleged branch account shortfalls recorded through the Horizon accounting system. The appellant alleges (among other matters) unlawful means conspiracy, abuse of process, and that the 2007 judgment was obtained by fraud; and he seeks to set aside that judgment and related bankruptcy consequences.

POL and Fujitsu rely on a 10 December 2019 settlement deed (the “Settlement Deed”) entered into in the Group Litigation Order proceedings (Bates & Ors v Post Office Limited) as a general release barring the “Historic Claims” (those relating to the earlier proceedings and alleged fraud).

At a directions hearing, the High Court ordered a separate preliminary issues trial to determine (i) construction of the Settlement Deed, (ii) whether it also releases claims against Fujitsu, (iii) whether POL is precluded from relying on it due to “unconscionability”, and (iv) an alleged fraudulent misrepresentation in 2019 relating to why a central Horizon witness (Gareth Jenkins) was not called. Critically, the High Court directed that the “unconscionability” issue would be tried on a hybrid basis: assuming (for the preliminary trial only) that the Historic Claims were viable and that POL knew that in 2019, while still requiring factual determinations on other elements.

Key issue on appeal (in substance)

Although permission to appeal was granted on a pleading point (CPR 16.5(1)), the Court of Appeal addressed the broader case-management concern identified by Lewison LJ: whether a preliminary issues trial may properly proceed on a hybrid basis—partly on assumed facts (including assumed corporate knowledge) and partly on judicial fact-finding—where the assumed matters may overlap with what must be proved.

2. Summary of the Judgment

The Court of Appeal held that the High Court’s hybrid approach—requiring the court to determine “unconscionability” while provisionally assuming that POL knew in 2019 that the Historic Claims were viable—creates too many practical difficulties to be workable. The assumed “knowledge” was described as a “legal assumption unpopulated by any facts”, yet the surrounding questions (including what POL knew about the appellant’s ignorance, and the overall “all the circumstances” unconscionability assessment) are likely to require examination of overlapping factual terrain.

The Court further noted a significant disagreement that emerged on appeal as to what “provisional assumption … for the purposes of the preliminary issues trial only” actually means—particularly whether POL could later litigate its actual knowledge if it lost the preliminary issues trial. That uncertainty itself reinforced the unworkability of the order as framed.

The Court concluded that the order “cannot stand in its present form”, and indicated that the parties should file short submissions on what should happen next (including whether the assumption should simply be removed, and whether the remainder of the preliminary issues structure remains appropriate). Because the order would need revisiting, the pleaded-defence postponement issue would also need reconsideration. The Court rejected a request to remit to a differently constituted first instance court.

3. Analysis

3.1 Precedents cited and their influence

(a) Bank of Credit and Commerce International SA v Ali [2001] UKHL 8; [2002] 1 AC 251

This House of Lords authority featured centrally in the Court of Appeal’s explanation of the legal nature of “unconscionability” (or “sharp practice”) as a potential route to prevent a party from relying on a general release.

The Court quoted Lord Nicholls at §32, distinguishing:

  • Mutual ignorance cases (both parties unaware of a claim later discovered), from
  • Asymmetric knowledge cases (the releasee knows the releasor “had or might have a claim” and also knows the releasor is ignorant of it), where seeking a general release without disclosure “could be unacceptable sharp practice” and the law should provide a remedy.

This informed the Court’s view that “unconscionability” is not a mechanistic checklist: even if certain elements are established (knowledge/ignorance), the court still must decide the overarching evaluative question—whether reliance on the release is unconscionable “in all the circumstances”. That evaluative question is fact-sensitive, making it problematic to “assume away” one major factual component (POL’s knowledge) while litigating others.

(b) The Horizon GLO judgments

  • Bates and ors v Post Office Ltd (No.3 Common Issues) [2019] EWHC 606 (QB)
  • Bates and ors v Post Office Ltd (No.6 Horizon Issues Rev.1) [2019] EWHC 3408 (QB)

These were not applied as binding “rules” for the case-management point, but they provided essential context: the Settlement Deed was executed after circulation of the draft Horizon Issues judgment and before hand-down, and it contained a “widely drafted” general release. The present action’s attempt to escape the Settlement Deed is therefore embedded in the procedural and evidential history of the GLO, including allegations about how the GLO was litigated (for example, the reasons given for not calling a central Horizon witness).

3.2 Legal reasoning

(a) Scope of the appeal: permission and the court’s own interest

Although permission to appeal was granted on a narrow pleading ground, Lewison LJ’s reason for granting permission framed a broader and more structural question: when ordering preliminary issues, the court should generally either (a) provide for all necessary facts to be found, or (b) direct trial on assumed truth of pleaded facts. The Court of Appeal treated that as within scope, noting that efficient use of court resources is a matter in which the court has its own interest.

(b) Hybrid preliminary issues: permissible in principle, but problematic here

The Court did not accept that a hybrid approach is always impermissible. It accepted, for example, that assuming the pleaded claim is “viable” can be a practical necessity when deciding whether a settlement deed is a complete answer.

The problem was narrower and sharper: assuming corporate knowledge of viable claims, while requiring fact-finding on related knowledge/ignorance questions and on the overarching unconscionability evaluation.

(c) Why assuming POL’s “knowledge” undermines workable adjudication

The Court identified multiple interlocking difficulties:

  • Knowledge is fact-laden. Determining whether POL “knew” in 2019 that the Historic Claims were viable would ordinarily entail findings about what particular individuals knew and when, and whether and how that knowledge is attributed to a corporation. Replacing that with an abstract “legal assumption” leaves a gap where facts may nonetheless be needed for adjacent issues.
  • Likely overlap with other required determinations. Whether POL knew the appellant was unaware of the Historic Claims may depend on when and how information about those claims existed within POL, whether it was suppressed, and its availability outside POL—all matters intertwined with what POL knew about the claims themselves.
  • The “all the circumstances” assessment requires granular knowledge findings. The precise extent and nature of POL’s knowledge may be central to whether reliance on a release is unconscionable: seniority of personnel with knowledge, what exactly they understood about viability, and whether information was actively suppressed.
  • Disclosure and cross-examination uncertainty. The assumption risked distorting practical trial management: Would disclosure be ordered on “knowledge” topics if knowledge is assumed? Could witnesses be cross-examined on knowledge facts relevant to credibility or to other elements? The Court considered that such uncertainty would likely cause unfairness or inefficiency.
  • Ambiguity about finality (“second bite”). A serious divergence emerged: POL argued the assumption would not stop it later litigating actual knowledge if it lost; others viewed the assumption as akin to an admission. If POL’s view were right, two trials could be needed with overlapping evidence and risk of inconsistent findings—precisely the kind of duplication preliminary issues are meant to avoid.

(d) Consequence: the order cannot stand as drafted

Given these practical and conceptual problems, the Court held the provisional assumption as to POL’s knowledge makes the preliminary issues trial structure unworkable. The order must therefore be revisited. The Court did not itself finally re-design the preliminary issues framework, instead inviting short further submissions on next steps (including whether removal of the assumption would leave the preliminary issues trial still appropriate).

(e) The CPR 16.5(1) pleading point

The appeal was formally permitted on the point that the respondents were relieved of pleading to certain pleaded passages (Annex 2) despite CPR 16.5(1). The Court’s answer was essentially pragmatic: because the structure of the preliminary issues (particularly the assumption) must be reconsidered, the identification of what parts of the pleading can properly be deferred must also be reconsidered. The Court observed that some deferred passages appeared to relate to matters intended to be tried in the preliminary issues.

3.3 Impact

The decision is a significant case-management marker for litigation seeking to escape broad settlement releases on “sharp practice/unconscionability” grounds, especially in complex institutional settings where “knowledge” is:

  • distributed across individuals, time periods, and corporate structures;
  • likely to be contested both factually (who knew what/when) and legally (attribution to the corporation); and
  • interwoven with other evaluative questions (such as whether reliance is unconscionable “in all the circumstances”).

Practically, the case cautions against preliminary issue designs that attempt to achieve efficiency by assuming a core corporate-knowledge element while still requiring the court to decide closely related knowledge and evaluative questions. Courts and parties designing preliminary issues will need to ensure either:

  • coherent fact-finding on the knowledge question within the preliminary trial; or
  • a clear assumed-facts model that does not create gaps, ambiguity, or duplicative later trials.

In the broader Horizon-related civil landscape, the decision underscores that challenges to the effect of the 2019 Settlement Deed may entail fact-intensive inquiry into what was known within POL, by whom, and how that bears on the “sharp practice” analysis. Attempts to shortcut those inquiries via partial assumptions risk procedural breakdown.

4. Complex Concepts Simplified

4.1 “Preliminary issues”

A preliminary issues trial is an early trial of selected issues (often legal or discrete factual points) intended to simplify or potentially dispose of the rest of the case. It saves costs only if it avoids duplicated evidence and does not create later “satellite” trials of overlapping matters.

4.2 “General release” in a settlement deed

A general release is a clause by which a claimant agrees they will not bring (or will give up) claims—sometimes broadly defined as claims “of whatsoever nature” whether known or unknown. Courts typically enforce clear releases, but there are limited doctrines that can prevent reliance in exceptional circumstances.

4.3 “Unconscionability” / “sharp practice” (in this context)

Here, “unconscionability” refers to the idea (drawn from Bank of Credit and Commerce International SA v Ali) that it may be wrongful for one party to take a general release when it knows the other party has (or might have) a claim and knows the other is ignorant of it—especially where non-disclosure makes taking the release unfair. Even then, the court must still decide whether, considering all circumstances, it is unconscionable to rely on the settlement.

4.4 “Corporate knowledge”

A company “knows” something only through individuals. Legal rules determine when and how individual knowledge is attributed to the company. That attribution can be complex where events span many years, personnel change, or knowledge is compartmentalised—making “assumptions” about corporate knowledge potentially unstable unless carefully defined.

5. Conclusion

Castleton v Post Office Ltd & Anor (Rev1) [2026] EWCA Civ 577 establishes that, while preliminary issues can legitimately proceed on assumed pleaded viability, a preliminary trial of “unconscionability/sharp practice” becomes unworkable where the court is asked to decide the issue while merely assuming a central and fact-laden component—corporate knowledge of viable claims—because that knowledge is likely to overlap with other necessary findings and with the overall evaluative judgment. The decision emphasises coherence and finality in preliminary issue design, and warns against hybrid structures that generate uncertainty on disclosure, cross-examination, duplication, and the risk of inconsistent findings.