Abstract
Keywords
Facts
During a 20th February 2018 search of Ms Perry's home, police found coded notes written on cigarette papers and hidden in a perfume box near her computer. The Crown advanced that the notes were part of a dissident-republican debrief of a 2015 police operation that led to the 2017 conviction of Kevin Nolan for firearms and explosives offences. The notes contained material likely to assist future terrorism by recording where items had been stored, suspected MI5 surveillance, and who could be trusted. Perry accepted the notes were in her handwriting and reflected a debrief but claimed unknown third parties pushed papers through her letterbox shortly before Christmas 2017; she copied them to protect source anonymity, then discarded the originals. She maintained she did not ‘collect’ the information, that copying was not ‘making a record’, and that any ‘currency’ the information once had had ‘long since dissipated’ (at [10]). Her defence statement set this out. Paragraph 4(n) added that the papers arrived ‘some considerable time’ after the Nolan events and sentence; she considered the information's ‘currency… had long since dissipated’. At trial, O’Hara J probed an apparent inconsistency: if the notes were largely meaningless to her and she did not know who Nolan was, how could she assert their obsolescence? At first instance, O’Hara J convicted after finding Ms Perry's account ‘not even possibly’ (at [16]) truthful, drawing together five strands: He read paragraph 4(n) of her defence statement – asserting the notes’ ‘currency… had long since dissipated’ – as showing knowledge of the 2015 arms find and Nolan's 2017 conviction, which squarely conflicted with her oral claim that the notes were largely meaningless to her; he rejected as inherently implausible her explanation that she copied the papers merely to protect a source, a conclusion reinforced by her decision to retain the notes despite claiming they had no value and by the absence of any comparable material saved to her laptop; he treated the fact the notes were hidden (albeit not expertly) as evidence of knowing possession rather than innocent handling; he noted that, despite her assertion that she frequently wrote on cigarette papers, none were found in the search and none produced at trial; and he concluded she had shifted to ‘a new and different’ (at [16]) narrative at trial to avoid the difficult questions her defence statement invited about her knowledge and the supposed obsolescence of the material. The Court of Appeal later observed that these five particulars were not exhaustive of the credibility concerns available on the evidence. The Court of Appeal dismissed her appeal, describing the trial judge's reading of para 4(n) as ‘obvious’ (or, if a fact question, ‘irresistible’) and emphasising that credibility rested on multiple strands beyond that paragraph. It nevertheless certified a question of general public importance about whether construing a defence statement is a question of law for the judge in a jury trial. The Supreme Court (Lord Hamblen JSC) held that the answer depends on purpose and context: the legal effect of a document is for the judge (law), but the meaning intended or understood by maker/reader is a question of fact for the fact-finder. In this case, the interpretation of para 4(n) was a question of fact because it concerned what the accused meant or understood, and because the defence statement had been put in evidence.
Appellate control over findings of fact in Northern Ireland non-jury criminal trials follows the four points summarised in Northern Ireland Railways v Tweed NIJB No.15 and restated by Lord Lowry LCJ in R v Thain [1985] NI 45. The Supreme Court quoted them in full: (1) primary facts – especially credibility, observation, memory – ‘can rarely be disturbed’ if supported by evidence; (2) inferences from documents or clear facts place an appellate court ‘in as good a position’ as the trial judge, though the latter's conclusions still command weight; (3) reversal is more readily justified where the judge misdirected in law, or misunderstood/misused facts and thereby reached a wrong conclusion; and (4) appellate courts must not replace the trial judge's view with conjecture about a balanced probability. In this case they can be applied as such:
Category 1 (primary facts/credibility): The trial judge's central task was to assess Perry's credibility against the totality of evidence. His reasons included: the defence statement's phrasing versus her oral account; the implausibility of copying and then keeping notes she said were useless; their concealment; and the absence of other tobacco-paper notes. These are classic credibility and primary fact assessments, strongly insulated on appeal if supported by evidence. The Court of Appeal confirmed further cogent reasons existed beyond para 4(n), reinforcing deference under category 1. Category 2 (inferences from documents): Perry argued that construing para 4(n) was a document-based inference, putting the appellate court in as good a position as the trial judge. The Supreme Court rejected that characterisation: the meaning of para 4(n) could not be divorced from the evidential context, including her inability in evidence to give any other reason for believing the notes were obsolete. Accordingly, this was not a pure category 2 case. Even if document inferences are sometimes equally open on appeal, the trial judge's overall evaluation still attracts weight. Category 3 (misdirection or misuse of facts): The Supreme Court located the case ‘most obviously within category 3’. There was no legal misdirection, and the judge's reading of para 4(n) – that the second sentence's ‘dissipated currency’ reason was the temporal remoteness in the first sentence – was a natural, nonperverse inference. Critically, Perry offered no alternative reason in her defence statement or oral evidence for that conclusion. Absent misdirection or misuse of facts, appellate interference was unwarranted. Category 4 (no conjecture/substitution): The Supreme Court cautioned against substituting its own view in a ‘balanced’ case. Here there were concurrent findings (trial and Court of Appeal) on credibility and meaning, and second-tier interference would run contrary to settled practice promoting finality and efficient use of appellate resources.
Perry's submission that an adverse interpretation of her defence statement could only be used if the judge was sure, it was the only proper interpretation (by analogy with directions on confessions) failed. Once the defence statement was in evidence, relevant statements could be weighed with the rest of the case; while overall guilt must be proved beyond reasonable doubt, individual evidential building blocks need not each meet that threshold.
The case clarifies that where a defence statement is put in evidence, disputes about what the accused meant typically raise questions of fact for the fact-finder; appellate courts will intervene only within the Tweed/Thain guardrails, most pertinently category 3. Practically, defence statements must anticipate judicial scrutiny: if asserting that information was ‘obsolete’ or its ‘currency… long since dissipated’, embed the reason – for example, lapse of time since a specified event, intervening operational developments, prior public-domain disclosure, or later documents that superseded earlier details. Ambiguity invites adverse inferences folded into credibility, an area where category 1 deference will usually be decisive. The decision underscores that defence statements, once in evidence, may found ordinary inferential reasoning about credibility, and that disputes about their meaning are quintessentially for the fact-finder, not for second-appeal re-calibration.
Commentary
For those of us practising or researching from the defence side of the aisle, the Supreme Court's organising move in Perry is welcome because it resists turning everyday credibility fights into pseudo-legal points. The Court restores an orthodox distinction: questions about a document's legal effect belonging to the judge; questions about its meaning – what the maker intended or how an ordinary reader would understand it in context – belong to the fact-finder (at [24]). That is not a clever technicality; it allocates decision making to the actor best placed to assess it. Meaning-in-context is the bread and butter of trial practice, and it is right that the tribunal hearing the evidence owns it. Importantly, the Court refuses to answer the ‘law or fact?’ question in the abstract: everything ‘depends on the nature of the statement and the purpose for which that statement is being relied upon’ (at [22]). That nuance matters, because defence statements perform two functions under the CPIA regime: they have a procedural life (are we compliant with s 6A?) and an evidential life (what did the client say, know, or mean?). Only the former reliably generates questions of effect for the judge; the latter is a classic question of meaning for the tribunal.
The decisive move below was prosaic but powerful: the judge read two adjacent sentences together and drew a ‘natural’ link – namely, that the assertion that ‘the currency in the information had long since dissipated’ was connected to the statement that the material was received ‘some considerable time’ after Nolan's conviction and sentence. The Supreme Court endorsed that as an orthodox inference: ‘“One does not need the inclusion of the word therefore”’ to read the sentences as connected, especially where no alternative reason was offered in the statement or in testimony (at [31]). From the defence side, this could be uncomfortable because adjacency can do heavy evidential lifting when our drafting is conclusory. But two guardrails matter. First, the Court is explicit that the inference contributed to – rather than determined – the adverse credibility finding; the judge had other cogent reasons, and the Court of Appeal concurred (at [32]). Second, Perry treats the reading as rooted in context, not in textual formalism: the absence of another stated basis for ‘dissipated currency’ and the inability to supply one credibly at trial made the inference reasonable. The lesson is stark: if we leave a vacuum around the ‘because’, the tribunal is entitled to fill it with ordinary-language logic. The antidote is to draft in a way that either supplies the connective tissue (because X) or breaks any unintended linkage (however/not because/although).
A central feature of Perry might be under-appreciated: ‘It is common ground that the defence statement and its factual contents were thereby put in evidence’ (at [12]). Once that step is taken – here, by the defence – the document becomes a live evidential exhibit: it can corroborate, contradict, and be mined for inference like any other prior statement. The appellant argued for confession style safeguards; the Court's response is blunt: ‘This is not … analogous to a confession’ (at [29]). With that, Perry also rejects the idea that each evidential stepping-stone must be proved beyond reasonable doubt: ‘Whilst the accused's guilt has to be proved beyond reasonable doubt, that does not mean that each evidential matter relied upon has to be proved to that standard’ (at [72]). For practitioners, two consequences follow. First, be intentional about whether to put the defence statement in evidence at all. Sometimes you will want to (e.g. to prove reasonable excuse or to contextualise conduct); if so, build a corroborative scaffold – other notes evidencing the same practice, third-party communications fixing timelines, digital artefacts anchoring recollections. If you do not need it as evidence, let it do its CPIA job and no more. Second, if meaning is contested, meet it with facts. Do not rely on doctrinal incantations to protect vague drafting; lead the reasons in chief and support them with material that makes the reading you invite feel inevitable rather than ex postrationalisation.
The appellate posture in Perry is orthodox but unforgiving for defendants. The Court reprises the familiar hierarchy: primary findings of fact – especially those steeped in credibility assessments – ‘can rarely be disturbed’ where supported by evidence; document-based inferences may in theory be more open, but the appellate court still gives due weight to the trial judge's conclusions; and speculation has no place (at [26]). Layered onto this is the doctrine of concurrent findings: where both the trial court and the intermediate appellate court agree on the facts, it will be ‘only in very rare cases’ that the Supreme Court will disturb those findings given its ‘prime function’ is to decide points of law of general public importance (at [33]). For the defence, this confirms that the trial is where meaning battles must be won. Attempts to relabel a meaning issue as a question of legal ‘construction’ will not carry you over the high threshold for second appeals. Strategically, therefore, lawyers will have to build your transcript with appellate restraint in mind: identify the meaning/effect fork expressly; invite the tribunal to treat meaning as a factual issue; and ensure your ‘because’ evidence is clearly led and recorded. Two statutory features explain why Perry lands where it does. First, s 6A CPIA prescribes what a defence statement must do and contain. That teleology helps classify disputes: an argument about whether a statement satisfies s.6A or triggers a procedural consequence is about effect (judge territory), whereas an argument about what a sentence meant to its maker or an ordinary reader – particularly where it narrates facts or beliefs – is a meaning question (fact-finder territory) (at [24]). Second, s.6E CPIA deems a defence statement to be given ‘with the authority of the accused’ unless the contrary is proved – so, absent contrary proof, its contents are the defendant's words for evidential purposes (at [25]). Together, these provisions turn defence statements into double-edged tools. They are indispensable for focusing issues and stimulating disclosure – both of which satisfy the efficiency measures so often sought by the justice system. However, once you deploy them in evidence they can – and will – be read as the client's own account. The practical response is to draft them with the same forensic care you would bring to a witness statement: avoid conclusory slogans; record the basis for belief or knowledge; and update or amend where disclosure changes the picture.
This case does not invent a special rule for non-jury trials, but the texture changes when the judge is the fact-finder. Relying Northern Ireland Railways v Tweed (1982) NIJB No.15 and R v Thain [1985] NI 457 the Court underscores why an appellate court should be slow to second-guess a trial judge's evaluation of a defendant's answers (at [26]). In England and Wales, the meaning/effect divide applies identically in jury trials – the jury resolves meaning, the judge polices legal effect. But in a non-jury setting, the judge's questions will often crystallise the meaning dispute in real time. For the defence, they treat these moments as chances to fill the gaps: if the court sees an adjacency-based inference, they need to be ready with the contemporaneous reason that breaks the causal link; if the court queries why ‘dissipated currency’ was asserted, they need to have the documentary or experiential basis to hand. The objective is two-fold: persuade the fact-finder now and create a transcript that shows the meaning issue was squarely confronted on the evidence. It bears emphasis that Perry does not license courts to treat every document question as ‘fact’. The judgement refuses to decide the allocation ‘in abstract terms’ (at [22]). There will be cases where the quarrel is genuinely about legal effect. For example, whether the statement sufficiently particularised an alibi to forestall adverse comment, or whether a statutory consequence was engaged; those are judge questions. Nor does Perry dilute the Crown's ultimate burden. It simply clarifies that not every intermediate inference must be proved to the criminal standard of beyond reasonable doubt (at [72]), and that confession jurisprudence does not govern the use of defence statements (at [29]). Fairness remains protected by ordinary trial safeguards: the right to give a full explanation, judicial directions on how to treat prior statements, and reasoned findings that expose the route to verdict. For the defence, that means our safeguards are earned through precision and proof, not through importing analogies that Perry squarely rejects.
After Perry, the takeaway for defence lawyers is simple: draft a defence statement on the assumption that a judge will read each sentence to your client and ask, ‘What did you mean by this?’ build the answer into the text. If you say the information was obsolete or its ‘currency … had long since dissipated’, explain why – for example, (i) the lapse of time since a specified event, (ii) intervening operational developments, (iii) prior disclosure into the public domain or (iv) later documents that superseded and made the earlier details useless. Where two ideas appear side-by-side but are not meant to be causally linked, say so explicitly (‘however’, ‘not because’, and ‘although’), or restructure the paragraph entirely. Those of us who work on disclosure and managerialism recognise the systemic pressures that make defence statements do too much too early. They are drafted under time constraints and often before meaningful disclosure is complete. Perry does not deny this reality; it simply refuses to graft confession-law protections onto a different genus of document (at [29]). The ethical and professional response is ours to own. Say only what you can presently justify; give the reasons you actually have; amend promptly when the picture changes. These are not mere defensive habits. They re-align the CPIA process with its proper aim – clarifying live issues – while reducing the scope for ‘gotcha’ credibility contests that pivot on wording choices made in a hurry. Doctrinally, Perry is conservative in the best sense. It places ‘What did the defendant mean?’ in the category where it logically belongs – questions of fact – even when the vehicle is a document. That sits comfortably alongside criminal law's everyday inferential method for intention, knowledge and belief. And by reiterating the Supreme Court's institutional role – ‘very rare’ interference with concurrent findings because its ‘prime function’ is to decide points of law (at [33]) – the judgment stabilises expectations about second appeals. For defence advocates, that stability is both a constraint and a gift: a constraint because you cannot count on a legal relabel to reopen factual evaluations; a gift because it tells you exactly where to invest your energy – front-load precision, build your evidential record, and treat meaning as the fact question it is.
Perry is best understood as a constitutional restatement about where epistemic authority sits in criminal adjudication. By insisting that disputes over a defence statement's meaning fall to the fact-finder while disputes about its legal effect remain for the judge, the Court re-anchors interpretation within the ordinary architecture of the adversarial trial (at [24]–[25]). That allocation matters because it resists the juridification of credibility: not every argumentative difficulty raised by a defendant's words may be laundered into a legal question. Instead, the case affirms a norm of epistemic humility – that factual understanding arises from the total evidential matrix, not from doctrinal fiat. This posture also clarifies the kind of inference the criminal courts are prepared to countenance. The trial judge's reading of adjacent sentences in para 4(n) exemplifies a permissible contextual inference rather than a formalist exercise in textualism. The Supreme Court's point – that one does not need an explicit connective (‘therefore’) to read the sentences as linked, particularly where no rival explanation was advanced – exposes the deep structure of fact-finding: tribunals draw sensible, defeasible inferences from ordinary language and context (at [31]). The defence complaint, stripped to theory, was that such inferencing risks over-production – for example, generating more meaning than the author intended. The answer is to make that risk a matter for evidence, not a reason to relocate the dispute into law. If that is the epistemology, the procedural corollary is a renewed ethic of reasons-giving in defence case construction. The animating idea behind my shorthand ‘put the because on the page’ is not a practice tip but a commitment to inferential transparency. In a system that treats meaning as fact, the defendant's narrative must supply the warrants that make the conclusion plausible within the record. That ethic sits comfortably with CPIA's telos: defence statements are designed to sharpen the real issues, not to perform cryptic gestures. Perry thus encourages what we might call a discipline of explication: if ‘currency had long since dissipated’, the justificatory because – time elapsed in relation to specified events; loss of intelligibility; absence of onward operational use – ought to be articulated ex ante rather than reconstructed ex post.
The Court's rejection of confession-law analogies reinforces this frame. By stating, crisply, ‘[T]his is not … analogous to a confession’ (at [29]) and reminding that the criminal standard attaches to guilt, not to each evidential step along the way (at [29]), Perry refuses to exceptionalise defence statements. The deeper theoretical move is to keep the law of proof and the law of admissibility in their proper lanes. Defence statements, once in play, are simply part of the trial's evidential ecology; their persuasive force turns on coherence and corroboration, not on threshold doctrines designed for qualitatively different risks. The protection against over-reach is not a heightened gateway but the tribunal's duty to reason from the whole.
Appellate restraint supplies the institutional backstop for this settlement. By restating that primary fact-findings ‘can rarely be disturbed’ and that disturbing concurrent findings is a ‘very rare’ step reserved for exceptional cases (at [26] and [33]), the Court casts the Supreme Court's role as one of normative calibration, not renewed fact-weighing. Theoretically, this preserves the vertical separation of functions: trial courts resolve meaning; appellate courts police law. Normatively, it incentivises parties to litigate meaning where it belongs – through evidence and reasons at first instance – rather than to relabel factual disputes as ‘construction’ on a second appeal. There is, nevertheless, an important fairness undertone. Defence statements are often produced under disclosure scarcity and managerial time pressure. Perry does not deny those structural conditions; it simply declines to let them recalibrate doctrine. The theoretical response is not to immunise defence statements from ordinary inferential use but to press for an ethic of candour and amendability: say what can be justified now; explain why; and amend when, and if, the picture moves. That approach takes seriously the Court's insistence that meaning is found, not declared; it also aligns with an adversarialism that, at its best, treats clarity as a condition of fairness rather than as its enemy. Seen this way, Perry is not a narrow ruling about a paragraph in a defence statement but a reminder about the grammar of criminal proof. Meaning is a factual phenomenon; inferences arise from context; and appellate law is not a solvent for disappointments in fact-finding. The defence response, theoretically grounded, is to practise an expressivist style of pleading and proof: articulate the reasons that underwrite your client's conclusions, integrate them with corroborative materials, and situate both within the statutory purposes of the CPIA. If the trial remains the privileged site of truth-seeking, then the defence's central intellectual task is to render its account legible to that site. To borrow the Court's own idiom, the division of labour between effect and meaning (at [24]–[25]) is not an inconvenience; it is the condition under which our arguments can be heard – and fairly judged.