Introduction
Identification evidence has long been recognised as one of the most unreliable and dangerous forms of evidence in the criminal justice system. Despite often appearing compelling, history demonstrates that even confident, sincere, and well-intentioned witnesses can be mistaken. Some of the most serious miscarriages of justice in English law have arisen from wrongful identification. It is for this reason that identification evidence attracts particular judicial caution, crystallised in the seminal decision of R v Turnbull.¹
Alongside this substantive caution sits an important procedural safeguard: the submission of no case to answer, also referred to as a half-time submission. Where, at the close of the prosecution case, identification evidence is insufficiently reliable to support a conviction, the defence may invite the court to stop the case altogether. In the Magistrates’ Court, this is governed by Criminal Procedure Rules 24.3(3)(d),² and in the Crown Court by R v Galbraith.³
This article examines the intersection between Turnbull directions and no case submissions, exploring when identification evidence is so weak that it should not be left to the tribunal of fact at all.
The Inherent Risks of Identification Evidence
Identification evidence is uniquely vulnerable to error. Factors such as fear, stress, poor lighting, distance, brief observation, and assumption can distort perception and memory. Psychological research has consistently shown that memory is not a recording device but a reconstructive process, influenced by later information and expectation.⁴
Crucially, confidence is not a reliable indicator of accuracy. A witness may be entirely honest and convincing, yet wholly mistaken. These dangers underpin the law’s insistence that identification evidence must be treated with particular caution.
PACE Code D and Identification Procedures
Identification evidence is governed at the investigative stage by PACE Code D, which regulates the methods by which witnesses may be invited to identify suspects. Code D establishes a hierarchy of identification procedures, including video identification (such as VIPER), identification parades, and group identification. Confrontation is expressly identified as a method of last resort.⁵
The purpose of Code D is fairness. While non-compliance does not automatically render identification evidence inadmissible, breaches may significantly undermine its probative value. A flawed identification procedure can exacerbate the inherent weaknesses of an already fragile initial observation.
The Turnbull Framework: Judicial Directions and the Quality of Identification
In R v Turnbull, the Court of Appeal laid down mandatory guidance for cases depending wholly or substantially on identification evidence. The court held that the tribunal of fact must be warned of the special need for caution before convicting.⁶
Turnbull requires the court to:
- Identify the quality of the identification evidence;
- Direct attention to factors such as lighting, distance, duration, stress, and prior familiarity;
- Remind the tribunal of the possibility of honest mistake;
- Consider whether there is supporting evidence capable of confirming correctness.
Importantly, Turnbull applies equally to recognition evidence. Although recognition may be more reliable than identification of a stranger, the Court of Appeal made clear that recognition can still be mistaken and must be approached with caution.⁷
Where the quality of identification evidence is poor, the judge should withdraw the case unless there is supporting evidence capable of making it safe.
From Caution to Termination: No Case to Answer Submissions
Turnbull’s most powerful effect is often felt at the halfway stage of the trial.
The Magistrates’ Court
Under CrimPR 24.3(3)(d), the court may acquit if the prosecution evidence is insufficient for any reasonable court properly to convict.⁸
The Crown Court
In R v Galbraith, the Court of Appeal established that:
- Where there is no evidence on which a jury could convict, the case must be stopped;
- Where the evidence is so weak that no reasonable jury, properly directed, could convict, the case should also be withdrawn.⁹
Identification cases frequently fall within the second limb of Galbraith. Where identification evidence is of poor quality and unsupported, the law does not require the matter to be left to “jury common sense”.
When Identification Evidence Is Too Weak to Proceed
A submission of no case to answer may properly succeed where:
- The prosecution case depends wholly or substantially on identification;
- The identification amounts to a fleeting glimpse or compromised observation;
- The witness was under stress or shock;
- The identification procedure was flawed or suggestive;
- There is no independent supporting evidence, such as CCTV, forensic material, admissions, or corroborative eyewitnesses.
At this stage, the court is not assessing credibility but legal sufficiency. The question is whether a conviction could safely be reached, not whether the witness might be believed.
Preparing Identification Cases for Trial
Effective preparation in identification cases requires early forensic analysis. Defence practitioners must scrutinise:
- The initial observation conditions;
- Compliance with PACE Code D;
- Discrepancies in descriptions;
- The evolution of witness confidence;
- Whether alleged supporting evidence genuinely supports identification or merely demonstrates motive or opportunity.
These considerations inform cross-examination, legal submissions, and the strategic decision to advance a no case submission.
Conclusion
Identification evidence sits at the intersection of human fallibility and legal consequence. R v Turnbull remains a vital safeguard, ensuring tribunals of fact are properly warned of the dangers of mistaken identity. Where identification evidence is so weak that it cannot safely support a conviction, the law does not require a defendant to face the risk of an unsafe verdict.
The submission of no case to answer provides a crucial procedural backstop. Together, Turnbull, PACE Code D, and Galbraith form a coherent framework designed not merely to regulate trials, but to prevent injustice before it occurs. When identification evidence is insufficient, the correct course is clear: the case should stop.
R v Turnbull [1977] QB 224 (CA).
Criminal Procedure Rules 2020, r 24.3(3)(d).
R v Galbraith [1981] 1 WLR 1039 (CA).
Cheryl Thomas, ‘Understanding the Reliability of Jury Verdicts’ (2010) MoJ Research Series 1/10.
Police and Criminal Evidence Act 1984, Code D.
R v Turnbull [1977] QB 224, 228–231.
R v Bentley (1991) 92 Cr App R 96.
Criminal Procedure Rules 2020, r 24.3(3)(d).
R v Galbraith [1981] 1 WLR 1039, 1042.
Written by Georgina Davies, who works at POCA Solicitors. She holds an LLB degree and is a Bar student at the University of Law, Birmingham.