In 1924, Lord Justice Atkin offered a clear explanation for why good fact-finding is key to the success of a case:
“An ounce of intrinsic merit or demerit in the evidence, that is to say, the value of the comparison of evidence with known facts, is worth pounds of demeanour.”
Société d’Avances Commerciales (Société Anonyme Egyptienne) v Merchants' Marine Insurance Co
The point here is that a case should turn on whether evidence aligns with known facts, not assessed on how persuasively that case is delivered. A century later, Atkin LJ’s justification remains the keystone of modern litigation.
Courts today are naturally cautious about using demeanour as a measure of credibility. The Judicial College, in Dame Hazel Genn’s ‘Assessing Credibility’, calls demeanour “a poor lie detector” and notes that judges are “no better than lay observers at detecting lies”. Cultural background, stress, and courtroom context can all shape how a witness presents to the judge, but confidence and accuracy rarely align.
This caution has developed over the last forty years. A brief history follows. In Armagas Ltd v Mundogas SA (The “Ocean Frost”) [1986] AC 717, Goff LJ said that veracity should be tested “by reference to the objective facts proved independently of their testimony, in particular by reference to the documents in the case”.
Leggatt J developed that reasoning in Gestmin SGPS SA v Credit Suisse (UK) Ltd [2013] EWHC 3560 (Comm), describing memory as “a mental record which is fixed at the time of experience of an event and then fades (more or less slowly) over time”. In Blue v Ashley [2017] EWHC 1928 (Comm), Leggatt J went further, advising judges to “place little if any reliance on witnesses’ recollections…and to base factual findings on inferences drawn from the documentary evidence and known or probable facts”.
Warby J subsequently applied the same principle in R (Dutta) v General Medical Council [2020] EWHC 1974 (Admin) and Khan v General Medical Council [2021] EWHC 374 (Admin), calling reliance on demeanour “a discredited method of judicial decision making”. He reiterated that credibility should first be tested against the documentary record–effectively, the facts.
From Atkin LJ in 1924 to the Commercial Court today, the message is consistent: facts are to be verified by comparison with what is known, not by performance on the stand.
Documents are the foundation of fact-finding
As a result, the ‘contemporaneous document’ has become the cornerstone of modern fact-finding. It captures events before litigation distorts perception of the parties. It reflects what was actually known, done, or believed at the time, rather than what the parties might want to recall to the benefit of their case.
Cross-examination turns less on catching a witness in contradiction with themselves and more on confronting them with their own correspondence, emails, or messages. Skeleton arguments cite Gestmin almost as routinely as Henderson v Henderson. Trial preparation has become a forensic exercise in aligning testimony with the documentary record.
But the documentary record itself has grown unmanageable. What used to be a few folders of letters is now terabytes of email, chat logs, WhatsApps, and attachments. The “intrinsic merit” Atkin described is now buried deep within the evidence. Courts expect decisions grounded in objective evidence, but the quantity of that evidence often now exceeds what humans alone can interpret.
From documents to facts
eDiscovery tools were the first step for processing, de-duplicating, and tagging millions of documents. They made disclosure manageable but are focussed at a document level, rather than telling the parties what those documents mean against the objective facts.
Meanwhile, the first wave of legal AI tools focussed on productivity across the firm. Platforms such as Harvey, CoCounsel, and Legora deliver drafting and summarisation. They improve general efficiency but were not built for the realities of contested evidence.
Fact intelligence (Wexler’s domain) takes the Gestmin principle to its logical conclusion. Instead of processing documents in bulk, it analyses what those documents actually say about who did what, when, and how. Wexler builds chronologies that link directly back to the underlying documents, maps relationships between people and events, and surfaces contradictions across the record, even live during depositions.
Wexler enables lawyers to test evidence rather than merely sort it, and to focus on how the objective facts fit with this evidence. It makes it possible to apply Atkin LJ’s principle of “comparison with known facts” at the scale that modern litigation requires.
Document discipline
Of course, documents aren’t infallible–a contemporary record can be wrong, biased, or incomplete. Mills & Reeve recently called this the “top trump” fallacy: essentially that is the assumption that documents always win. Although documents should be the first test, and the tools for testing them should be rigorous, documents do not always win.
Atkin LJ’s emphasis on “comparison with known facts” remains the most reliable framework for evaluating evidence. However, if Atkin LJ was sitting in a modern trial, he would find his principle stretched to breaking point. The challenge is no longer a shortage of facts but a mountain of them.
Wexler makes those facts manageable. It allows litigators to compare testimony with known facts continuously, to verify every connection, and to build a record grounded in objective truth.
An ounce of intrinsic merit is still worth pounds of demeanour. Now, for the first time, we have the tools to find that ounce in real time.