The deference of an appeal court to demeanour-based assessments of a witness’s credibility at trial is illustrated in Cox v R (Turks and Caicos Islands) [2023] UKPC 4, at [31], [34], [40].
There is always a kind of residual feeling that, even so, the credibility assessment may have been wrong. Deference avoids the need for appellate courts to deal with this. Plainly, the law has to be pragmatic about what risks are acceptable. Without challenging the correctness of the conclusion of this appeal, where circumstances supported the credibility assessment (see at [34(iv)], [46]-[47], [49]), we can still be left wondering about some general issues.
How reliable is an assessment of credibility based on demeanour? What does science tell us about this? Is an articulated warning about the need for caution in assessing credibility based on demeanour sufficient to counteract the risk of error, or might it be just a formulaic ritual?
Is it appropriate for an appellate court to defer to credibility assessments made at trial? Will technology, such as video recordings of trials, diminish the so-called advantage that the trial court has? Will the recently confirmed ability of appellate courts to reach what are really verdicts mean that appeals against conviction become replays of trials?
Should corroboration be required where the credibility of a witness is of central importance?
Another point brought to mind by this appeal is the inability of the courts to provide a remedy for unacceptable failure of timely disclosure if it does not affect the fairness of the trial (see [43]-[44]). Does unacceptable failure have no consequence? What does “unacceptable” mean?