Identification evidence: proving who committed the offence

Overview

A court may be sure that an offence occurred while remaining uncertain who committed it. Proof of identity connects the accused to the relevant conduct. That connection can come from a witness, a recording, physical traces or several circumstances considered together. The legal question concerns the strength of the connection, rather than the number of people willing to name a suspect.

Identification evidence (preuve d'identification) requires particular care because an honest account can be mistaken. The witness may have had a poor opportunity to observe, may have received suggestive information later, or may now remember a photograph more clearly than the event. The apparent confidence of courtroom testimony does not remove those possibilities.

This lesson concerns Canadian criminal trials, including those in Quebec. It starts with the observation, follows the identification process and then considers independent recordings and forensic evidence. The admissibility lesson explains which information the court can receive. The reasonable-doubt lesson explains the Crown's ultimate burden.

Learning Objectives

• Assess identification through observation conditions and the complete history of the witness's account.
• Distinguish a supported identification from a general resemblance, repeated selection or forensic association.
• Apply case-specific caution and the Crown's burden to the whole evidentiary record.

Reconstruct the original observation

Separate honesty from accuracy

An eyewitness (témoin oculaire) may sincerely identify the wrong person. The court must therefore examine what the witness could perceive, as well as whether the account is honestly given. Length of observation, distance, lighting, obstructions, movement and the visible part of the face matter. The opportunity to observe should be established through evidence rather than assumed from a witness's proximity to the event.

Attention matters too. A person may have concentrated on a weapon, an escape route or an injured companion. Stress does not have one mechanically predictable effect in every case. The assessment concerns the particular observation and the features actually noticed. Describing an entire encounter as lasting several minutes does not establish that the witness saw the perpetrator's unobstructed face for that period.

In R. v. Hibbert, 2002 SCC 39, paras. 44–53, the Supreme Court explains the danger of apparently persuasive but unreliable identification, including the effects of later exposure to the accused's image. The court's task is not completed by describing the witness as sincere. Confidence and accuracy must be considered in light of the circumstances in which the identification developed.

Record the description before later influences

The first description provides a reference point for evaluating later accounts. Preserve what the witness actually said, including uncertain language, omissions and estimates. A description recorded before exposure to a suspect's photograph can reveal what the witness initially remembered. It is not automatically accurate, but it allows the court to compare later changes against a less influenced account.

General characteristics have limits. A dark coat, approximate height and broad age range can narrow a search without identifying one individual. A distinctive feature may carry more significance, depending on observation quality and how common it is. Both matches and discrepancies deserve attention. The absence of an obvious feature in the first account may matter, although it should not be treated automatically as proof that the witness saw someone else.

An identification history should include failures to select, tentative choices, selections of other people, changed descriptions and the timing of each. The final confident statement cannot replace that history. In R. v. Hay, 2013 SCC 61, paras. 40–45, the Court examines why a particular weak identification could not support conviction by itself despite the witness's later explanation of what the earlier selection meant.

Examine the basis for recognition

Recognition (reconnaissance) depends on actual familiarity. A neighbour seen daily under ordinary conditions presents a different starting point from someone glimpsed once at a crowded event. Establish the frequency, duration, recency and nature of earlier contact. A bare assertion that the witness knows the person supplies little information about whether recognition was realistically possible on the occasion in question.

Previous familiarity does not remove poor viewing conditions or later influence. The witness may recognize a gait, voice or other characteristic while never seeing a face clearly; that basis must be specified and tested. Canada Evidence Act s. 6.1 confirms that identification may use vision or another sensory mode. The label recognition does not itself determine the weight of the evidence or create a lower criminal standard of proof.

Follow the identification process from beginning to end

Understand what a lineup can test

A photo lineup presents a further opportunity to test the witness's memory. Its usefulness depends on how it was prepared and administered. If the suspect stands out because of appearance, image quality or some other feature, the witness may select that photograph for reasons unrelated to the original observation. The record should preserve the photographs, instructions, sequence and complete response.

The federal-provincial-territorial report Innocence at Stake, chapter 3, recommends safeguards such as an administrator who, where possible, is independent of the investigation and unaware of the suspect's identity; an instruction that the perpetrator may be absent; photographs that do not make the suspect stand out; and a record of all responses and initial confidence. It also recommends sequential presentation and steps to prevent feedback or contact between witnesses afterward.

These are prevention recommendations, not a statutory checklist that automatically decides admissibility. An irregular procedure can weaken the inference without necessarily excluding the evidence. Conversely, departure from a recommendation does not make every legal objection a matter of weight only. Applicable exclusionary rules, constitutional issues and trial fairness require their own analysis. Proper procedure improves the opportunity to assess reliability; it does not guarantee correctness.

A one-person presentation carries a pronounced risk of suggestion because the expected selection may be obvious. The report restricts its recommended use to rare circumstances, such as an apprehension near the scene soon after the event. The reasons for using that procedure and its actual conditions should be documented. Its urgency does not erase the potential influence on a later lineup.

Identify later exposure and feedback

Contamination can arise from conversations with other witnesses, news coverage, social-media searches or police comments. It need not involve dishonesty or intentional coaching. A person may merge the event with later information while believing that the resulting memory comes entirely from the event. Repetition can then strengthen confidence without adding an independent observation.

The circumstances of an online selection therefore belong in the record. Identify who initiated the search, what names or accusations were supplied, which images were viewed and whether other people suggested a choice. A witness who had already seen a tagged suspect photograph did not approach a later police lineup with the same information as a witness seeing it for the first time.

Feedback after selection can also matter. Telling the witness that the chosen person is the police suspect may affect later certainty. The initial confidence statement and the later courtroom confidence should be considered in their respective contexts. The 2018 report discusses research on initial confidence under carefully controlled conditions; it does not justify treating every confident witness as accurate or treating confidence as wholly irrelevant in every setting.

Two witnesses are not necessarily two independent sources of support if they discussed identity before making their selections. Nor does one witness's repeated identification create three independent identifications. The court must trace where the information originated and whether the apparent confirmation depends on the same potentially mistaken source.

Treat courtroom identification with particular caution

A dock identification often occurs in a setting that plainly identifies the person accused. Its dramatic effect can exceed its evidentiary value, especially where a stranger was never reliably identified beforehand. The witness may now recognize the accused from earlier hearings or images. Asking the witness to deny that influence does not establish that the influence has been eliminated.

Hibbert explains why the warning must fit the actual identification history. A weak courtroom identification does not become stronger simply because the witness points confidently in front of the jury. The law does not impose a universal ban on courtroom identification, but the court must assess what the act of pointing actually contributes and the danger of overvaluing it.

Evaluate recordings on what they actually show

Separate the recording from an observer's conclusion

A properly admitted recording permits the trier of fact to inspect features that may be unavailable through a witness's memory. In R. v. Nikolovski, [1996] 3 SCR 1197, paras. 22–23 and 28–32, the Supreme Court recognizes that sufficiently clear video can itself establish identity. A human witness who names the person in the image is not invariably required. The quality and clarity of the recording, the view provided and the opportunity to examine it remain central.

The recording's authenticity and its identifying force are separate questions. Establishing that a clip comes from the relevant camera does not prove that a blurred face is the accused's. A poor-quality image can still contribute information about movement, clothing or timing, even where it cannot independently establish facial identity. Repeated viewing cannot create detail that the recording does not contain.

Nikolovski also requires care in the decision process. A jury making its own identification must receive directions to examine whether the recording's clarity, quality and duration permit identification beyond a reasonable doubt. If the recording is the only identification evidence, the jury must be reminded of that limit and the burden. A judge sitting alone must apply the same cautions. The trier of fact must be able to review the recording during deliberations using the same or similar viewing equipment as at trial. The Court also recommends that a judge sitting alone indicate when the images appear capable of supporting identification, so counsel can address image quality or changes in the accused's appearance through evidence and submissions.

Playback should preserve a fair opportunity for both parties to examine the evidence. The complete relevant sequence may explain a misleading still frame. Cropping, slowed playback or an enhanced image should be distinguished from the original recording, with an adequate explanation of what was done. The physical and documentary evidence lesson addresses authenticity, integrity and the proper use of digital exhibits.

If someone else offers a recognition opinion about a person in the recording, the court must separately examine that proposed testimony. In R. v. Berhe, 2012 ONCA 716, paras. 13–26, the Ontario Court of Appeal applies the requirements of prior familiarity and a better position than the trier of fact to assist identification. The judge should determine that foundation in a voir dire, a hearing on the proposed opinion's admissibility; its ultimate reliability remains separate. There is no additional threshold requirement to identify a unique physical idiosyncrasy. Merely repeating an investigator's conclusion cannot supply the required foundation. An expert claim requires the distinct screening explained in witnesses and testimony.

Worked example: a photograph shared after a brief observation

Factual Scenario

In this hypothetical, a witness sees a person leave a tool depot at night for approximately four seconds, through rain on a window. The witness initially describes a medium-height person wearing a bright cap and cannot describe the face. The next day, a friend sends a social-media photograph naming Morgan as the suspected thief. The witness later selects Morgan in a police lineup and expresses certainty at trial.

An authenticated camera clip shows a person in a bright cap leaving the depot, but no usable facial detail. Assume the Crown has no other evidence linking Morgan to the event, no recognition based on previous familiarity, and no dispute that property was unlawfully taken. The defence disputes identity without identifying another perpetrator.

Connect forensic findings to the alleged conduct

A match answers only the question it addresses

Source attribution concerns the origin of a trace. Depending on the method and result, a fingerprint or DNA analysis may strongly associate a person with material recovered from an object. The further inference concerns how, when and why that material came to be there. An association with a movable object does not automatically establish participation in the offence in which the object later appeared.

The court needs evidence about the sampling location, handling, opportunity for innocent contact and the expert opinion's actual limits. The statistical expression accompanying a DNA result must not be recast as the probability that the accused is guilty. Scientific strength on one proposition cannot silently establish a different proposition about activity or timing.

In R. v. Metzger, 2023 SCC 5, paras. 1–9, the majority found the convictions unreasonable on the particular combined record. DNA on a cigarette butt in a stolen vehicle established at most a connection with the vehicle before recovery; the surrounding evidence did not adequately establish the necessary timing and participation. The other identifying evidence had serious weaknesses. The decision demonstrates the need to examine the combined reasoning, rather than treating a match as a complete answer.

The absence of a trace likewise has no fixed meaning in every case. Its significance depends on whether the alleged activity would be expected to leave a recoverable trace, what was sampled and what the methods could detect. A non-match may undermine a proposed source attribution, but does not necessarily exclude involvement if the trace could have come from someone else. The actual scientific evidence must support the proposed inference.

Worked example: a print on a reusable delivery container

Factual Scenario

In this hypothetical, an expert reliably attributes a fingerprint on the outside of a reusable delivery container to Riley. The container is found inside a warehouse after a theft. Admissible employment records establish that Riley handled that same container during an authorized delivery the previous week. No witness or recording identifies Riley at the theft, and the expert cannot date the print. Assume there is no other incriminating evidence.

Reach the verdict from the whole record

Confirmatory evidence must support the disputed connection. A second source proving that a crime occurred may add little to the identification of its author. Several genuinely independent circumstances can nevertheless establish identity even though none does so alone. Circumstantial evidence must be considered together, with reasonable alternative inferences grounded in the evidence or its absence. R. v. Villaroman, 2016 SCC 33, paras. 35–38, distinguishes those reasonable alternatives from speculation and rejects requiring the accused to prove facts supporting every alternative.

There is no general requirement of a DNA result, fingerprint or second eyewitness before conviction. Hay recognizes that a single reliable eyewitness can suffice, while also explaining that a particular identification may be too weak to support conviction alone. A standard caution cannot supply missing identifying evidence. Directions must connect the recognized risks to the weaknesses and supporting features of the actual case. Hay, para. 41, adds a procedural consequence: on a motion for a directed verdict, the trial judge must direct an acquittal where the Crown's entire case consists of eyewitness testimony that, even if believed, would necessarily leave reasonable doubt in a reasonable juror's mind. That threshold concerns whether the evidence can legally support conviction, rather than the judge simply preferring a different assessment of a witness's credibility.

An alibi does not shift the burden of proving identity from the Crown to the accused. Hibbert, paras. 58–67, distinguishes rejecting an alibi from finding deliberate fabrication involving the accused. Mere disbelief does not strengthen the Crown's case. There must be other evidence supporting deliberate fabrication and the accused's involvement in the attempt to mislead. Where that foundation exists, the trier of fact may, but need not, infer consciousness of guilt from the attempted deception. That inference is not conclusive proof of guilt and, depending on the case, may support presence without establishing participation in the offence.

The accused's silence cannot repair an otherwise insufficient identification case. Metzger rejects that reasoning on its weak evidentiary record. The final assessment must identify what the admitted evidence establishes, what remains uncertain, and whether the Crown has proved the accused's participation beyond a reasonable doubt. Where that doubt remains, the verdict must reflect it.

Practice Checklist

• Record observation duration, visibility, attention and sensory limitations.
• Preserve the earliest description and every identification attempt.
• Identify earlier familiarity and all later exposure to suspect information.
• Examine lineup composition, instructions, administration and feedback.
• Separate initial confidence from confidence after repeated exposure.
• Test whether supposed confirmation is truly independent and identity-specific.
• Distinguish authentic footage from sufficient identifying detail.
• Separate forensic source, timing and activity inferences.
• Apply the Crown's burden without demanding proof of an alternative perpetrator.