|Pierce Butler (Library of Congress)|
On July 4, 1927, Joe Knudson, a 55-year-old Norwegian rigger, was injured while splicing a cable on a boat owned by one Charles Benson, who had loaned it to another person. The boat was in dry dock near the Mobile River in Alabama when Knudson was injured. Knudson lost 65 percent of the use of his right leg as a result of the injury and filed a claim against Benson under the federal Longshoremen’s and Harbor Worker’s Compensation Act of 1927. Letus Crowell, a deputy commissioner of the agency charged with enforcing the act, upheld Knudson’s claim, but Benson got a federal District Judge to enjoin Crowell’s order. The judge decided after a trial de novo that Knudson was not Benson’s employee at the time of the injury. A Circuit Court of Appeals affirmed the District Court, and Crowell appealed. (For a fuller statement of the facts and a penetrating analysis of the case, see Mark Tushnet’s “The Story of Crowell: Grounding the Administrative State,” in Federal Courts Stories, ed. Vicki C. Jackson and Judith Resnik (Foundation Press, 2010).)
By a 5-3 vote the Supreme Court affirmed, in an opinion by the Chief Justice. Hughes distinguished between two different categories of facts when it came to the judicial review of agencies’ fact finding. The first category included “the circumstances, nature, extent, and consequences of the employee’s injuries and the amount of compensation that should be awarded.” For such run-of-the-mill facts, the findings of deputy commissioners, if supported by substantial evidence, were final. De novo review of these facts would destroy the “prompt, continuous, expert and inexpensive method” of dealing with workplace injuries that the statute contemplated, Hughes explained.
But “a different question is presented,” he continued,
where the determinations of fact are fundamental or “jurisdictional,” in the sense that their existence is a condition precedent to the operation of the statutory scheme. These fundamental requirements are  that the injury occurs upon the navigable waters of the United States, and  that the relation of master and servant exists. These conditions are indispensable to the application of the statute, . . . because the power of the Congress to enact the legislation turns on the existence of these conditions.If the injury did not occur on a navigable waterway, Hughes explained, it was beyond the reach of the federal government and was a matter for the states. If the injury did not occur within the course of employment, forcing a business owner like Benson to “compensate” an injured person was a violation of substantive due process, a taking of the property of A and giving it to B. Federal judges had to be free to find such facts de novo, that is, on a record created in their own courts. To require federal judges to defer to administrators on such matters would be to “sap the judicial power as it exists under the Federal Constitution, and to establish a government of a bureaucratic character alien to our system.”
Brandeis, joined by Roberts and Stone, dissented. He trained considerable fire on “the inherent uncertainty” of Hughes’s category of jurisdictional facts. Yet he identified “the primary question for consideration” not as the existence of some category facts that courts could determine upon their own reading of the weight of the evidence. Rather, it was "upon what record shall the District Court’s review of the order be the deputy commissioner be based?” The District and Circuit Court judges decided that “all the evidence introduced before the deputy commissioner should go for naught; and that respondent should have the privilege of presenting new, and even entirely different, evidence in the District Court.” The Supreme Court had hamstrung the Interstate Commerce Commission by insisting upon de novo review of its fact-finding in Alabama Midland (1897). Congress had liberated the ICC with the Hepburn Act in 1906, but the dissenter feared that the category of jurisdictional facts might prove so expansive that agencies would be returned to judicial bondage.
From the report one might well conclude, as Alpheus T. Mason did in his biography of Stone, that “at conference [the justices] split 5 to 3, Roberts siding with Stone and Brandeis” (337). In fact, nine justices voted at conference. Mason evidently arrived at eight because by the time the decision was announced on February 23, 1932, Holmes had retired from the Court. But Holmes still sat when the Court heard argument on October 20 and 21, 1931, and both Roberts’s and Butler’s docket books have him voting at conference on October 31.
Mason also erred by stating that three justices dissented at conference; both docket books have only Brandeis dissenting. Holmes’s vote did not surprise me; he had followed the jurisdictional fact doctrine in Interstate Commerce Commission v. Northern Pacific Railway Co., 216 U.S. 538 (1910), even though he had acknowledged the “sometimes considerable” difficulty of distinguishing jurisdictional from other kinds of facts. (I learned from a citation in Butler’s docket book that Holmes had also followed the doctrine in Miller v. Horton, 152 Mass. 540 (1891).)
But the vote in conference raised an obvious question: what led Roberts and Stone to join Brandeis’s dissent? Butler recorded several issues that were easily dismissed. Did Congress interfere with the vesting of the judicial power in the Supreme Court by transferring cases involving injuries on navigable waterways from the jurisdiction of admiralty courts to a federal agency? Butler wrote, “No question whether admiralty court jurisdiction has been infringement [sic].” He also saw “No lack of due process.” The justices also considered some “question of authority under the act” and “whether any arbitrary action present[ed a] question of law.”
But “is there an interference with federal court jurisdiction,” Butler asked, “when we come to question of fact”? That is, did the statute profess to keep the federal courts from determining whether the commission exceeded its statutorily conferred jurisdiction and the constitutional limits Congress presumably respected in passing the organic act. At this point Butler wrote the word “jurisdictional" with the words "fundamental” and "essential" bracketed next to it–apparently a note on the definition of jurisdictional facts that would later appear in Hughes's opinion. Butler did not mention navigability, but he did write the words "employee" and “Here ques[tion] of m&s [master and servant] vel non.”
Butler’s entry in his docket book thus suggests that at conference Hughes articulated his new version of the jurisdictional fact doctrine and that an alarmed Brandeis dissented, fearful that the new formulation might expand weight-of-the-evidence review. But Butler did not record any discussion of the record upon which the federal courts were to determine jurisdictional facts. Of course, given the sketchiness of Butler’s notes, one draws the negative inference–that the issue was not discussed at conference–at one’s peril. Still, if Hughes did not argue that trial courts should hear new evidence relating to jurisdictional facts until his opinion circulated–no later than December 29, when Brandeis’s legal secretary Henry M. Hart flagged the issue for his justice–Roberts’s and Stone’s switch becomes understandable.
Apparently Brandeis circulated his opinion by January 10, when Roberts discussed it with him. The next day Roberts followed up with a letter, in which he ventured:
If a right to retry the question of jurisdiction on new evidence is accorded the employer, he may of course absent himself entirely from the proceeding before the deputy commissioner until an award is made, and then go to court and submit his evidence under a bill of equity. We would not permit a respondent in a Federal Trade Commission case to do any such thing, or in an Interstate Commerce Commission case, or in a number of other controversies in inferior tribunals to which you have referred throughout your opinion.Stone saw things the same way. On February 19 he wrote to Brandeis:
I think the point you make that even if the considerations which seem to weigh with the Chief Justice requires limiting the power of the Commission, it does not follow that you should destroy his power as a collector of evidence. It would certainly be easier to construe [the statute] as requiring the Commissioner to make up the record and requiring the Court to review that record even though it reviewed the evidence, than to construe it as the Chief Justice does. I am not sure but you might make even more of this point. . . . If that much of the act were saved, it could be made a workable act.I’m left wishing for some evidence that Holmes expressed his willingness to join Brandeis in dissent before January 11, the day Hughes made and Holmes promptly accepted the suggestion that he resign from the Court. Holmes would have seen Hughes’s draft opinion when it circulated in December, but perhaps Brandeis refrained from sending him his dissent or from otherwise sounding out his faltering friend. Still, I hope the forgoing is enough to encourage other historians to ask to consult the Butler and Roberts docket books when pursuing their own hunches about the Supreme Court's deliberations.