A lie detector rejected in 1923 still decides which experts a jury hears.
A Washington murder defendant asked a court to let a blood pressure machine vouch for him. The judges said no, in under a thousand words, and wrote a sentence that still keeps novel science out of courtrooms in California, New York, Illinois, Pennsylvania, Washington, and Minnesota.
By Helena Harper, Editorial Director
A confession, then a retraction
On November 28, 1920, Dr. Robert W. Brown was shot and killed at his home in Washington, D.C. The case went cold. Roughly a year later a young Army veteran named James Alphonso Frye was picked up on an unrelated matter and told police he had killed the doctor. Then, once he had lawyers, he took it back. The confession was false, he said, part of an arrangement to collect and split the reward money. A jury in 1922 did not believe the recantation. Frye was convicted of second degree murder and sentenced to life. He was paroled in 1939.
Between the confession and the verdict, his defense tried something no American court had approved. It offered a psychologist, William Moulton Marston, who had developed a systolic blood pressure deception test. The theory was that lying under questioning is work, and that the effort of concealment drives blood pressure up in a pattern a trained observer can read. Marston had tested Frye in jail and concluded he was truthful. The defense wanted the jury to hear that, and offered to repeat the test in open court. The judge refused both.
Two pages that reorganized evidence law
The appeal produced one of the shortest landmark opinions in American law. Decided on December 3, 1923 by the Court of Appeals of the District of Columbia, predecessor of today's D.C. Circuit, it runs under a thousand words, barely two pages in the Federal Reporter. Justice Josiah Van Orsdel wrote it. It affirms the conviction by way of a general observation about how courts should treat science they have not seen before.
The hard part, the opinion says, is that a scientific principle passes only gradually from the experimental to the demonstrable, and it is difficult to say exactly when the line is crossed. Somewhere in that "twilight zone," as the court put it, the principle must be recognized. So the court supplied a proxy. The thing from which the deduction is made must be sufficiently established, in the court's words, to have gained "general acceptance in the particular field in which it belongs." Marston's test had not. It was excluded, and Frye's conviction stood.
A gate, not a scale
Frye asks one threshold question before a jury hears anything: has the technique gained general acceptance in the field it belongs to?
The field decides, not the judge
The court counts heads in the relevant scientific community rather than grading the science itself. Consensus is the proxy for reliability.
Federal courts left in 1993
Daubert held that the Federal Rules of Evidence, not Frye, govern in federal trials, and made general acceptance one factor among several.
Many state courts did not
The 1923 test still controls novel scientific evidence in several of the largest state court systems, so the forum can change the answer.
There is a footnote that has kept the case in circulation outside law schools. Marston kept working on deception, and in 1941 a character he helped create first appeared: Wonder Woman, whose signature weapon is a lasso that compels the truth.
What the test actually asks
Read closely, the standard does something modest and something radical at once. It does not ask a judge whether a technique works. It asks whether the people who would know already think it does. The court delegates reliability to the relevant scientific community, then does something close to counting: is this accepted in its field, or still contested?
The thing from which the deduction is made must be sufficiently established to have gained general acceptance in the particular field in which it belongs.
Two consequences follow, and both still shape motion practice. The inquiry aims at method rather than conclusion, so an opponent challenges the technique, not the number the expert reached. And because it is a threshold question about novelty, courts reserve it for genuinely new or contested science. Nobody demands a hearing before a physician testifies about a broken arm. The fights are over a comparison discipline, a statistical model, a diagnostic method that has outrun the literature behind it.
The federal courts left in 1993
Seventy years after Frye, the Supreme Court held that the Federal Rules of Evidence had displaced it. In Daubert v. Merrell Dow Pharmaceuticals, decided June 28, 1993, Justice Blackmun wrote for the Court that Rule 702 says nothing about general acceptance and that a rigid acceptance requirement is at odds with the Rules' permissive approach to evidence. In its place came a flexible reliability inquiry with the trial judge as gatekeeper, guided by whether the theory can be tested, whether it has been peer reviewed and published, its known or potential error rate, the existence of controlling standards, and general acceptance, now one factor among several.
The federal framework kept growing. Kumho Tire Co. v. Carmichael (1999) extended the gatekeeping duty beyond strictly scientific testimony to technical and other specialized knowledge, while stressing that the factor list is not a checklist. Amendments to Rule 702 effective December 1, 2023 then made two points explicit that courts had been getting wrong: the party offering the expert must establish the rule's requirements as more likely than not, and the opinion must reflect a reliable application of the method to the facts.
Where the 1923 rule still governs
State courts were free to follow Daubert or not, and many did not. California still applies the general acceptance test through People v. Kelly (1976), layered since 2012 with the separate gatekeeping duty described in Sargon Enterprises v. University of Southern California. New York adheres to it through People v. Wesley (1994). Illinois retained it in Donaldson v. Central Illinois Public Service Co. (2002), and its Rule 702 says so expressly. Pennsylvania reaffirmed it in Grady v. Frito-Lay (2003), holding that it reaches an expert's methods rather than the conclusions drawn from them. Washington applies it under State v. Copeland (1996). Minnesota uses a two part version known as Frye-Mack, reaffirmed over Daubert in Goeb v. Tharaldson (2000).
The map is not static, and it has moved in one direction. New Jersey incorporated the Daubert factors in In re Accutane Litigation on August 1, 2018. The Supreme Court of Florida adopted the Daubert amendments on May 23, 2019, receding from its earlier decisions to keep Frye. Maryland replaced its longstanding Frye-Reed rule in Rochkind v. Stevenson on August 28, 2020, by a four to three vote. Published fifty state charts disagree at the margins, which is itself a reason to check current authority in the forum rather than rely on a table.
Why any of this reaches your file
The practical effect is that the same expert, with the same report, can be admissible in one courthouse and excluded across a state line. That turns an evidentiary rule into a strategic one: where a case can be filed, and whether removal is available, carries a consequence for proof that has nothing to do with the merits. Counsel who practice in both systems keep two challenge templates, because the arguments do not transfer. A Frye motion is built out of literature, professional bodies, and prevailing practice in a defined field. A Daubert motion is built out of testing, error rates, and the fit between method and conclusion.
It also changes what an expert must be ready to say. In a general acceptance jurisdiction, the decisive testimony is often not about the expert's own work. It is about the field: who uses this method, where it is published, whether the professional community treats it as settled. A hundred and three years on, the question the D.C. judges asked about a blood pressure cuff is the question that decides whether a jury hears the science at all.
A man asked a machine to speak for him. The court said the machine had not yet earned the right, and then wrote down how anything else would.
Frye lost his appeal in two pages. The rule those pages created still sits at the courthouse door in some of the busiest trial systems in the country.
Archivar
The Archivar Editorial Desk
Researched and fact-checked against the primary opinions and multiple independent sources. General information for attorneys, not legal advice.
Sources: Frye v. United States, 293 F. 1013 (D.C. Cir. 1923), the opinion of the Court of Appeals of the District of Columbia (Van Orsdel, J.), via Wikisource and Justia; Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993), and Kumho Tire Co. v. Carmichael, 526 U.S. 137 (1999), via the Legal Information Institute at Cornell Law School; the 2023 amendments to Federal Rule of Evidence 702, as reported by the American Bar Association and by law firm analyses of the Advisory Committee's notes; People v. Kelly, 17 Cal.3d 24 (1976) and Sargon Enterprises, Inc. v. University of Southern California, 55 Cal.4th 747 (2012); People v. Wesley, 83 N.Y.2d 417 (1994); Donaldson v. Central Illinois Public Service Co., 199 Ill.2d 63 (2002) and Illinois Rule of Evidence 702; Grady v. Frito-Lay, Inc., 576 Pa. 546 (2003); State v. Copeland, 130 Wn.2d 244 (1996); Goeb v. Tharaldson, 615 N.W.2d 800 (Minn. 2000); In re Accutane Litigation (N.J. 2018); the Supreme Court of Florida's May 23, 2019 order adopting the Daubert amendments, as reported by The Florida Bar; and Rochkind v. Stevenson (Md. 2020). Case background from "Frye's Backstory," Journal of the American Academy of Psychiatry and the Law (2014), and from Encyclopaedia Britannica on William Moulton Marston.
