Written Down Is Not Proven

What private recovery writings, inventories and personal disclosures can and cannot tell a Canadian court, a child protection worker, or a victim-survivor

A document can be intensely private yet still contain information that matters to someone else’s safety. Those two things are not in tension. Privacy, truth, admissibility and probative value are four distinct questions, and answering one does not answer the others.

This article is written for people, most often victim-survivors, who have encountered detailed personal writings by another person: a journal, a workbook, a written inventory, an unsent letter, or a file on a shared computer. It explains what Canadian law and credible research can say about such records and, just as importantly, what they cannot.

One conclusion is stated at the outset because it governs everything that follows. The available evidence does not establish why any particular individual wrote what they wrote. Research can describe the functions that written self-examination may serve for people in general. It cannot reach into a single person’s motive. Anyone who tells you otherwise is guessing.

Why recovery contexts can generate detailed personal records

Twelve-step programs can involve extensive personal self-examination, acknowledgment, disclosure and amends. The wording of the Steps themselves is public and stable. Step Four asks members to have “made a searching and fearless moral inventory of ourselves.” Step Five involves admitting “to God, to ourselves, and to another human being the exact nature of our wrongs.” Step Eight asks for a list of people harmed. Step Ten calls for members to have “continued to take personal inventory and when we were wrong promptly admitted it.”

Several points follow, and each matters.

The Steps do not require writing. They speak of inventory, admission and listing. They do not prescribe a notebook. In practice, many groups, sponsors and published workbooks encourage written inventories, and written formats are widely used. But the practices of individual sponsors and groups are not the same as the official program, and it is wrong to assume that every participant creates a written record, or that any record found belongs to a program at all.

Twelve-step frameworks extend well beyond alcohol. Programs addressing compulsive sexual behaviour, including Sex Addicts Anonymous, state that their basic principles are drawn from the Twelve Steps and Twelve Traditions of Alcoholics Anonymous. The structure - inventory, disclosure, amends, continuing inventory - carries across.

Participation proves nothing about conduct. Twelve-step attendance is not a diagnosis, not a confession and not evidence of dangerousness. A major 2020 Cochrane systematic review by Kelly, Humphreys and Ferri evaluates these programs as interventions for alcohol use disorder. It makes no claims about participants’ character or criminality, and no one else should draw such conclusions from it.

It is also worth naming a common category error. “Sex addiction” is not a diagnosis in the DSM. A proposed diagnosis of hypersexual disorder was developed and field-tested during the DSM-5 process but was not included. The World Health Organization’s ICD-11 does include compulsive sexual behaviour disorder (code 6C72), classified among impulse-control disorders, and it expressly cautions that distress arising entirely from moral judgment or disapproval of one’s own sexual impulses does not meet the diagnostic threshold. Recovery-program participation is not a diagnosis, and a diagnosis is not proof of harm to anyone.

What can research tell us about written disclosure?

Research on written disclosure has examined outcomes associated with writing about difficult experiences under experimental conditions. A major meta-analysis in this area, Frattaroli’s 2006 synthesis of 146 randomized studies in Psychological Bulletin (DOI 10.1037/0033-2909.132.6.823), found that experimental disclosure produced a positive and statistically significant but small average effect (r = .075), with substantial variation across contexts and populations.

That is the honest state of the general evidence. It concerns outcomes and moderators of disclosure under experimental conditions. It does not establish why any particular person creates a private record.

Recovery frameworks themselves expressly emphasize personal inventory, acknowledgment or admission of wrongs, identification of people harmed, amends and continuing inventory. But those are the stated elements of a program, not a finding about an individual. The line that must not be crossed:

• Research and program documents describe this as a possible function of such writing. Supportable.
• This is why this person wrote it. Not supportable, absent evidence about that person.

A record may have been produced as recovery work, as a private diary, as fiction, as a rehearsal for a conversation, as fantasy, as a legal or therapeutic exercise or for reasons no one will ever recover. The document’s genre is itself a fact requiring proof.

Private does not mean evidentially irrelevant

The starting point in Canadian law is broad: everyone owes a general duty to give evidence relevant to a matter before a court, so that the truth may be ascertained. The Supreme Court of Canada restated that principle in M. (A.) v. Ryan, [1997] 1 S.C.R. 157. Exceptions to it must be justified. The fact that an author intended a document to remain private does not, by itself, establish an evidentiary privilege.

At the same time, nothing about a document’s usefulness suspends privacy law. Both propositions are true simultaneously, and confusing them is a common error in this area.

Six words that are not synonyms

• Relevance - does this make a fact in issue more or less likely?
• Admissibility - may the trier of fact receive it at all, given the exclusionary rules?
• Authenticity - is the document what it is claimed to be, created by whom it is claimed to have been created by?
• Reliability - can the content be trusted for the purpose it is offered for?
• Corroboration - does genuinely independent evidence confirm parts of the account?
• Weight - having been admitted, how much does it actually prove?

A record can clear the first four and still carry little weight. It can be highly relevant and still be excluded. It can be admitted and still fail to establish the event it describes.

Admissions, and their limits

Out-of-court statements offered to prove the truth of their contents are presumptively inadmissible as hearsay. A long-established evidentiary rule permits admissions of a party to be tendered against that party. The Supreme Court explained the rationale in R. v. Evans, [1993] 3 S.C.R. 653: the rule does not rest on any circumstantial guarantee of trustworthiness, but on the fact that a party cannot complain about an inability to cross-examine themselves. Appellate courts have confirmed that it operates in civil as well as criminal proceedings, though it is invoked less often there (see R. v. Foreman (2002), 62 O.R. (3d) 204 (C.A.)).

Three cautions belong with that:

1. The rule is about parties. It applies where the author is a party to the proceeding in which the record is tendered. It is not a general principle that anything anyone wrote down comes into evidence.
2. Admissibility is not proof. Getting a document before a judge is the beginning of the analysis, not the end. Weight remains fully in play.
3. Not every “confession” fits a doctrine. A “declaration against interest” is a distinct hearsay exception that concerns statements by people who are not parties to the proceeding and is generally invoked where the declarant is unavailable and the other requirements of that exception are met. The Supreme Court said in R. v. Schneider, 2022 SCC 34, that the party admission exception should not be confused with it, because the two rest on different foundations. Assuming a written passage automatically qualifies under one doctrine or another is exactly the sort of error that collapses a case.

There is also a category of use that sidesteps hearsay: a statement offered simply to prove that it was made, rather than that its contents are true, is not hearsay for that purpose - though other admissibility rules may still apply, and “not hearsay” does not mean “automatically admitted.” That distinction can matter enormously and is easy for non-lawyers to lose.

Authentication, provenance and corroboration

Before its contents can be considered, the document must be authenticated as what it purports to be. Electronic records are also subject to specific evidentiary rules.

The applicable statute depends on the proceeding. In a family proceeding, s. 23(1) of the Divorce Act directs the court to the laws of evidence of the province in which the proceeding is taken, subject to the Divorce Act and other Acts of Parliament. In Alberta, that means ss. 41.3 to 41.5 of the Alberta Evidence Act, R.S.A. 2000, c. A-18: s. 41.3 places the burden of proving authenticity on the party introducing the record; s. 41.4 addresses application of the best evidence rule to electronic records; and s. 41.5 establishes specified presumptions concerning the integrity of electronic records systems. Those are specified presumptions, not a general assumption that any digital file or system is inherently reliable. The federal Canada Evidence Act, R.S.C. 1985, c. C-5, contains analogous electronic-document provisions at ss. 31.1 to 31.3. It should not be treated as the default evidentiary statute for an ordinary provincial family proceeding under the Divorce Act.

Practically, a record’s significance often turns on less dramatic questions:

• Who created it, and how can that be shown?
• When was it created, and is that verifiable?
• Is it contemporaneous with the events described, or reconstructed years later?
• Has it been altered, and can that be detected?
• Do any dates, places, names or events in it match independently verifiable facts?
• Does similar information appear across records that do not depend on each other?

Independent corroboration can materially strengthen the evidentiary significance of a particular assertion. A record naming a date that can be checked against a receipt, a travel record or a third party’s independent account does something a bare narrative cannot.

Corroboration must be genuinely independent. Two documents do not provide two sources if one was copied from, summarized from or generated out of the other. A notebook entry, a message quoting that entry and a later account written from memory of both may look like three sources and function as one. Before treating records as mutually corroborative, you must examine their provenance and their relationship to each other.

Finally, a firm boundary: this article does not describe, and no one should attempt, methods of accessing another person’s devices, accounts, passwords or protected records without authorization. How material was obtained is itself a legal question with consequences, and it is a question for a lawyer before it is a question for anyone else.

Privacy, confidentiality and privilege are three different things

Privacy is a protected interest whose legal consequences depend on the context. Canadian courts recognize that a reasonable expectation of privacy can persist in personal communications even after they leave the author’s own space.

Confidentiality describes an expectation or a promise between people or within an organization. It is real and socially important. It is not, by itself, a legal shield.

Privilege is the legal doctrine that keeps relevant evidence out. Canadian law recognizes very few class privileges. In R. v. Gruenke, [1991] 3 S.C.R. 263, the Supreme Court declined to create a class privilege for religious communications, and instead confirmed that privilege may be recognized case by case, guided by the four Wigmore criteria: the communication originated in a confidence that it would not be disclosed; that confidentiality is essential to the relationship; the relationship is one the community should sedulously foster; and the injury from disclosure outweighs the benefit to the correct disposal of the litigation. M. (A.) v. Ryan confirmed that approach and held that a privilege, where recognized, may be partial rather than absolute.

Applied to peer-support settings, this yields an answer that is genuinely unsettled and context-dependent, and should be described that way.

Within AA, anonymity and confidentiality are related but distinct norms. Tradition Twelve describes anonymity as the spiritual foundation of all the Traditions, placing principles before personalities, and AA’s own materials explain anonymity as protecting members from identification as alcoholics at the personal level and as restraining the exploitation of AA affiliation at the level of press and media. Separately, AA states that members understand personal disclosures made in meetings to be treated as confidential. Both operate as important norms within the fellowship. Neither is a statutory confidentiality regime, and neither is, in itself, Canadian evidentiary privilege. Equally, recovery-related communications are not always obtainable; a case-by-case claim can be advanced, and its outcome will turn on the specific facts.

A further distinction is worth drawing. A purely private record created and retained solely by its author presents a different privilege question from communication made within a confidential relationship - the Wigmore framework is directed at the latter. That said, privilege analysis is fact-specific, and a private document can later be incorporated into confidential communication or prepared for another person’s review.

A note on health records, which are outside this analysis

Psychotherapy notes, clinical charts and other professional health-care records raise a separate set of questions: health-information legislation, professional confidentiality obligations, third-party production procedures and, potentially, case-by-case privilege - the very issue litigated in M. (A.) v. Ryan. They should not be treated as equivalent to a person’s own journal, a recovery workbook, or correspondence they voluntarily sent to someone else. This article deliberately leaves them aside, precisely to avoid creating the false impression that every record touching on personal behaviour should be treated as a protected medical record.

Why patterns matter

Canadian family law expressly requires courts to consider patterns of coercive and controlling behaviour, while also assessing individual acts of family violence in their context. Section 2(1) of the Divorce Act defines family violence as conduct, whether or not it constitutes a criminal offence, by a family member towards another family member that is violent or threatening, or that constitutes a pattern of coercive and controlling behaviour, or that causes the other family member to fear for their safety or another’s - and, in the case of a child, direct or indirect exposure to such conduct. The definition expressly includes psychological abuse, financial abuse, harassment including stalking, sexual abuse and threats.

Section 16(3)(j) requires courts to consider family violence in determining a child’s best interests, and s. 16(4) directs attention to its nature, seriousness and frequency; whether there is a pattern of coercive and controlling behaviour; whether a child is directly or indirectly exposed; harm or risk of harm; compromise to safety; and any steps taken by the person engaging in the violence. Section 16(2) makes the child’s physical, emotional and psychological safety, security and well-being the primary consideration. Section 16(5) is the necessary limit: past conduct is considered only where it is relevant to exercising parenting time, decision-making responsibility or contact.

A 2023 research report by Peter G. Jaffe, Nicholas Bala, Archana Medhekar, Katreena L. Scott and Casey Oliver, Making Appropriate Parenting Arrangements in Family Violence Cases, published by the Department of Justice Canada, makes the practical point plainly: recognizing patterns of coercive control often requires consideration of a series of events over time by someone able to take a more objective view, rather than focus on the most recent or most serious incident. The report’s authors also state that denial and minimization of intimate partner violence by genuine abusers are more common than false or exaggerated claims by alleged victims.

Four distinctions have to be kept in view when applying this to a written record:

• Repetition is not corroboration. One person saying the same thing five times in one notebook is one source, not five.
• Consistency is not proof. Internally coherent accounts can still be false; inconsistent accounts can still be true.
• An allegation is not an established fact, whoever made it and however detailed it is.
• A behavioural pattern is not a psychiatric diagnosis. Describing conduct is not diagnosing a person, and no reader of a document is qualified to do the latter.

Multiple genuinely independent records that converge - a written passage, a dated message, a third party’s contemporaneous account, a document generated by an unrelated system - have a different evidentiary character from a single ambiguous sentence. That is a statement about corroboration, not about certainty.

When child protection genuinely becomes relevant

Child-protection systems assess risk and circumstances. The connection has to be specific and evidence-based. Information in a private record may become relevant to a child-safety assessment where it bears on matters such as: conduct directed at or involving a child; a pattern of coercive control operating within the household; a child’s direct or indirect exposure to family violence; parenting capacity; or identifiable risk factors that a qualified professional should assess.

The concern is not hypothetical. The Divorce Act itself treats a child’s direct or indirect exposure to family violence as within the definition, and the 2023 Justice Canada-published report notes that child exposure to domestic violence is among the most common forms of child maltreatment substantiated by Canadian child protection services. Child-protection legislation is provincial, and intervention thresholds vary by jurisdiction.

The distinction to hold onto: conduct that affects a child’s safety is the trigger. Not sexuality. Not an addiction label. Not recovery participation. Not the existence of a notebook.

Different systems, different questions

Three Canadian processes may engage the same document to different ends.

• Family proceedings decide parenting arrangements and support. The standard of proof is the civil one. In F.H. v. McDougall, 2008 SCC 53, the Supreme Court held that there is only one civil standard - the balance of probabilities - and that it does not rise with the seriousness of the allegation, though the evidence must always be scrutinized with care and be sufficiently clear, convincing and cogent to satisfy it.
• Child-protection proceedings are forward-looking and risk-focused, governed by provincial statutes with their own thresholds and procedures.
• Criminal proceedings operate under different evidentiary and constitutional rules. At trial, the Crown must prove guilt beyond a reasonable doubt. Police investigations are governed by different legal thresholds depending on the investigative step, and state evidence-gathering is subject to the Charter.

Relevance in one forum does not guarantee admissibility, or sufficiency, in another. And in every forum, one distinction matters more than any other for a person holding a document: an investigative lead is not proof of a fact. A record may point investigators toward a date, a location, a witness or another record. That is valuable. It is not the same as establishing that an event occurred.

What this means if you are the one holding the document

A record containing a serious allegation or admission warrants careful assessment. It is not proof that conclusively establishes the underlying event. Holding both of those thoughts at once is difficult and necessary.

What such a record may legitimately do:

• give context to experiences that previously seemed isolated or inexplicable;
• suggest a pattern that a qualified professional should assess;
• identify dates, places or events capable of independent verification;
• point toward other evidence that does not depend on the writer;
• confirm or contradict something already known;
• inform of a decision to seek legal advice or a professional safety assessment.

What to do:

Preserve the original state of the material as far as reasonably possible. Do not alter, annotate, reorganize or overwrite it. For digital material, the original file, its file path, its metadata and its storage context can all matter to proof of authenticity later. Obtain legal advice about appropriate preservation and storage where evidentiary significance may be in play.

Get legal advice before acting on it. Questions of lawful access, disclosure obligations, privacy and admissibility are fact-specific, and getting them wrong can harm the very case you are trying to make.

Where family violence, coercive control or other safety concerns are present, do not raise the material with the person who created it until you have safety or legal advice. This is not a point about politeness. The 2023 Justice Canada-published report documents that the post-separation period can involve heightened risk of continued or escalating violence.

If the information raises a credible concern about a child’s safety, seek professional input rather than assessing it yourself. Depending on the circumstances, that may mean police, a child-protection authority, a specialized family-violence service, a lawyer or a qualified risk assessor. Immediate danger requires emergency assistance.

Where the evidence stops

When you discover someone else’s words, there is a strong pull to believe you have finally found the truth about that person. Resist that conclusion, not because the writing is unimportant, but because it goes beyond what documents, messages, or other records can establish on their own.

A record shows that words were written. Whether they are true, when they were written, why they were written and what they mean for anyone’s safety are separate questions, each requiring its own evidence. Diagnosis is beyond any reader’s reach. Motive is beyond the reach of the research.

Private writings should neither be treated as unquestionable truth nor dismissed merely because they originated in a personal or recovery context. When they contain information relevant to violence, coercion or child safety, the responsible approach is the same in every case: preserve the evidence, establish its provenance, seek corroboration, respect applicable privacy and privilege rules, and allow qualified professionals and justice institutions to assess its significance.

If you or someone you know is in immediate danger, call 9-1-1. This article is general information about Canadian law and research. It is not legal advice, and it does not create a lawyer-client relationship. Legal thresholds, child-protection legislation and evidentiary procedure vary by province and territory.

Glossary

Admissibility - whether a judge may receive a piece of evidence at all, given the exclusionary rules. Separate from whether it is true and from how much it proves.

Admission of a party - a statement made by someone who is a party to the proceeding, tendered against that party. A long-established rule permits this. It applies because the party cannot complain of being unable to cross-examine themselves, not because such statements are assumed reliable.

Authentication - proving that a document is what it is claimed to be, and who created it. For electronic records, this is a threshold step, and the burden falls on the party introducing the record.

Balance of probabilities - the civil standard of proof: more likely than not. There is only one civil standard, and it does not rise because an allegation is serious, though the evidence must always be sufficiently clear, convincing and cogent to satisfy it.

Best evidence rule - a rule concerned with proving the contents and integrity of documentary evidence. For electronic records, applicable evidence legislation contains specific rules for satisfying the best evidence requirement, including provisions concerning the integrity of electronic records systems.

Beyond a reasonable doubt - the criminal standard the Crown must meet at trial to prove guilt. It is a trial standard, not the threshold governing what police may investigate.

Class privilege - a privilege attaching automatically to an entire category of relationship. Canadian law recognizes very few. Where no class privilege exists, a claim may still be advanced on a case-by-case basis.

Coercive and controlling behaviour - a pattern of conduct, addressed expressly in the Divorce Act definition of family violence. Because it is pattern-based, recognizing it generally requires looking at a series of events over time rather than at a single incident.

Confidentiality - an expectation or promise between people, or within an organization, that information will not be passed on. Real and socially important, but not by itself a legal barrier to evidence.

Corroboration - independent evidence confirming part of an account. Independence is what makes it corroboration; records that derive from one another are one source, not several.

Family violence (Divorce Act) - conduct by a family member toward another that is violent or threatening, or that constitutes a pattern of coercive and controlling behaviour, or that causes fear for safety, and in the case of a child, direct or indirect exposure to such conduct. It applies whether or not the conduct is a criminal offence.

Hearsay - an out-of-court statement offered to prove that its contents are true. Presumptively inadmissible, subject to exceptions. A statement offered only to prove that it was made is not hearsay for that purpose, though other admissibility rules may still apply.

Indirect exposure - a child’s exposure to family violence without directly witnessing an incident. The Divorce Act definition captures both direct and indirect exposure.

Investigative lead - information that points toward a date, a place, a witness or another record. Valuable, and not the same as evidence capable of proving a fact.

Privilege - the legal doctrine that keeps relevant evidence out of proceedings. Distinct from privacy and from confidentiality. Where recognized case by case, it may be partial rather than absolute.

Probative value - how much a piece of evidence actually contributes to establishing a fact in issue.

Provenance - where a record came from, who created it, when and what has happened to it since. Often important in assessing authenticity, reliability and evidentiary weight.

Reasonable expectation of privacy - a protected interest that can survive even after material leaves its author’s own device. Under s. 8 of the Charter, it constrains state action, and that framework does not translate directly into disputes between private individuals.

Relevance - whether a piece of information makes a fact in issue more or less likely.

Reliability - whether the content can be trusted for the purpose it is offered. A record can be authentic and still unreliable.

Statement against interest - a distinct hearsay exception concerning declarations by people who are not parties to the proceeding, generally invoked where the declarant is unavailable and that exception’s own requirements are met. Not interchangeable with an admission of a party.

Weight - once evidence has been admitted, how much it proves. Admission is the beginning of the analysis, not the end.

Wigmore criteria - the four-part test guiding case-by-case privilege claims: communication originated in confidence; confidentiality is essential to the relationship; the relationship is one the community should sedulously foster; and the injury from disclosure outweighs the benefit to correctly disposing of the litigation.

Sources

Legislation - Divorce Act, R.S.C. 1985, c. 3 (2nd Supp.), ss. 2(1), 16(2)-(5), 23(1). Justice Laws Website, consolidated text. Readers should check the consolidation date shown on that page, as it changes. - Alberta Evidence Act, R.S.A. 2000, c. A-18, ss. 41.3-41.5. - Canada Evidence Act, R.S.C. 1985, c. C-5, ss. 31.1-31.3.

Jurisprudence - R. v. Evans, 1993 CanLII 86 (SCC), [1993] 3 S.C.R. 653. - R. v. Schneider, 2022 SCC 34, [2022] 2 S.C.R. 619. - R. v. Foreman (2002), 2002 CanLII 6305 (ON CA), 62 O.R. (3d) 204. - R. v. Gruenke, 1991 CanLII 40 (SCC), [1991] 3 S.C.R. 263. - M. (A.) v. Ryan, 1997 CanLII 403 (SCC), [1997] 1 S.C.R. 157. - R. v. Marakah, 2017 SCC 59, [2017] 2 S.C.R. 608. - F.H. v. McDougall, 2008 SCC 53, [2008] 3 S.C.R. 41.

Government research - Jaffe, Peter G.; Bala, Nicholas; Medhekar, Archana; Scott, Katreena L.; with Casey Oliver, Research Associate. Making Appropriate Parenting Arrangements in Family Violence Cases: Applying the Literature to Identify Promising Practices. Ottawa: Department of Justice Canada, 2023. The report states that the views expressed are those of the authors and do not necessarily represent the views of the Department of Justice Canada. - Department of Justice Canada, The Divorce Act Changes Explained; Fact Sheet - Divorce and Family Violence.

Peer-reviewed and systematic literature - Frattaroli, J. (2006). Experimental disclosure and its moderators: A meta-analysis. Psychological Bulletin, 132(6), 823-865. DOI 10.1037/0033-2909.132.6.823. - Kelly, J.F., Humphreys, K., & Ferri, M. (2020). Alcoholics Anonymous and other 12-step programs for alcohol use disorder. Cochrane Database of Systematic Reviews, CD012880. DOI 10.1002/14651858.CD012880.pub2. - Potenza, M.N., Gola, M., Voon, V., Kor, A., & Kraus, S.W. (2017). Is excessive sexual behaviour an addictive disorder? The Lancet Psychiatry, 4(9), 663-664. DOI 10.1016/S2215-0366(17)30316-4.

Clinical classification - World Health Organization, International Classification of Diseases, 11th Revision (ICD-11), code 6C72, Compulsive sexual behaviour disorder (impulse-control disorders).

Recovery organizations, describing their own programs - Alcoholics Anonymous World Services, Twelve Steps and Twelve Traditions (Steps Four, Five, Eight and Ten; Tradition Twelve). - Alcoholics Anonymous, Understanding Anonymity (pamphlet) and AA General Service Office FAQ, on the two levels of anonymity and on the understanding that personal disclosures made in meetings are treated as confidential. - Sex Addicts Anonymous (saa-recovery.org), statement that its basic principles are found in the Twelve Steps and Twelve Traditions of Alcoholics Anonymous.

Read Further

The full research and structural analysis are available in: