The AI Journal

On 25 April 1938 the Supreme Court of the United States reinstated a criminal indictment against an Illinois cannery. Carolene Products Company had shipped tins of Milnut, a compound of condensed skimmed milk and coconut oil sold as a cheap substitute for evaporated milk, across a state line. A federal statute of 1923, the Filled Milk Act, made that shipment a crime. A district judge had thrown the indictment out. Justice Harlan Fiske Stone, writing for the Court, put it back. The opinion is now read for a footnote. The footnote is not what the Court did.

The Filled Milk Act prohibited the interstate shipment of skimmed milk compounded with any fat other than milk fat so as to resemble milk or cream. Congress had declared the product an adulterated food, injurious to health and a fraud on the public. Stone accepted those declarations as legislative facts to which a court should defer. Regulatory legislation affecting ordinary commercial transactions, he wrote, is not to be pronounced unconstitutional unless the statute precludes the assumption that it rests upon some rational basis within the knowledge and experience of the legislators. Milnut was an ordinary commercial transaction. The cannery lost.

That sentence, not the footnote appended to it, is the holding. The footnote — the most famous in American constitutional law — reserved the possibility that the presumption of constitutionality might operate more narrowly when a statute appeared to violate a specific prohibition in the Bill of Rights, or restricted the political processes that could be expected to repeal bad laws, or was directed at religious, national or racial minorities, “discrete and insular” groups against whom prejudice might disable the ordinary political remedy. None of those reservations described Carolene Products Company. The company was a manufacturer. Its product was a food. Its complaint was due process. Stone classified that complaint as the kind of claim courts would no longer hear.

The classification was the point of the case, and it had a history. Filled milk was the successor to margarine in a fifty-year campaign by dairy producers against cheaper fats. Ruth Dupré has shown that federal margarine policy from the Oleomargarine Act of 1886 through the 1950s is intelligible as interest-group legislation: a tax on a competing spread, raised in 1902 to ten cents a pound on the yellow product and a quarter-cent on the uncoloured one, designed not to raise revenue but to keep margarine from looking like butter. Geoffrey Miller, in the fullest study of the filled-milk litigation, places the 1923 Act in the same sequence. Dairy farmers understood, correctly, that coconut oil in filled milk undercut the domestic butterfat market. The Act did not require better labels. It closed the interstate market. Miller’s conclusion is blunt: the effect of Carolene Products was to deprive working and poor people of a healthful, low-cost food. One may resist the nutritional claim — butterfat does carry vitamin A that coconut oil lacks, and in 1923 that was not a trivial point for infant feeding — without resisting the industrial one. A labelling rule would have addressed fraud. A ban addressed competition.

The Court was not unaware of the competitive structure. It saw no constitutional difficulty in the fact that the ban left oleomargarine and other butter substitutes untouched. Congress could strike at one substitute and leave another. In 1919, in Hebe Co. v. Shaw, the Court had already upheld an Ohio statute that kept a similar coconut-oil compound out of the condensed-milk trade. Justice Day, joined by Van Devanter and Brandeis, dissented on the ground that Hebe was a wholesome compound, distinctly labelled as such, and not sold as condensed milk. By 1938 the federal government had simply nationalised the Ohio result. The “ordinary commercial transaction” in the sentence to which Footnote Four is appended was a transaction the dairy industry had spent a generation trying to criminalise.

Law students do not meet the case this way. They meet Footnote Four, isolated, as the origin of the two-tiered structure of modern constitutional review, and as the seed of John Hart Ely’s representation-reinforcing theory of judicial review. Ely built Democracy and Distrust on the footnote’s second and third paragraphs: courts should police the political process and protect minorities who cannot win inside it, and should otherwise leave policy to legislatures. That is a coherent theory of the judicial role. It is not an account of what the 1938 Court did to the litigant who had invoked its protection. Carolene Products Company was a small firm selling a cheap food against a better-organised agricultural lobby. If the footnote’s third paragraph describes a process failure, the company was a candidate for the paragraph. The Court placed it in the category the footnote was written to exclude.

The strongest objection to this reading is that it confuses a holding with a theory, and then blames the theory for the holding. Louis Lusky, Stone’s clerk that term and the footnote’s drafter, later described the note as a starting point for debate in the spirit of inquiry, not as a settled theorem. It did not purport to decide anything. Chief Justice Hughes asked Stone to add the first paragraph, on specific constitutional prohibitions; Stone kept Lusky’s process paragraphs as well. On this view the filled-milk facts are pedagogically awkward and legally irrelevant. The Court was liquidating the old economic due process of the Lochner era, as it had to do after 1937, and it was simultaneously sketching, in a footnote, the grounds on which it might still intervene. Later cases — the white primary, Brown, the reapportionment decisions — gave the footnote its meaning. The dairy statute is the vehicle. The footnote is the cargo. To read the case as a dairy decision with a famous reservation is to miss what made it a landmark.

The objection would be stronger if the Court had treated the 1938 result as a one-off compelled by the vitamin findings of 1923. It did not. In 1944, in a second case also styled Carolene Products Co. v. United States, Justice Stanley Reed wrote for the Court that the Filled Milk Act survived even if one assumed the product was wholesome and properly labelled. Vitamin deficiency had been an efficient cause of the 1923 Act, Reed said, but not the sole one. The compounds lent themselves to substitution and confusion with milk products, and Congress could exclude them for that reason alone. Black and Douglas, no friends of economic due process, concurred in the result. The 1944 decision is the untaught sequel. It tells us what the 1938 Court thought the category of ordinary commercial transactions contained: not a provisional deference pending better nutrition science, but a lasting permission to suppress a competitor of milk, labels and vitamins notwithstanding.

Footnote Four became the charter of a rights jurisprudence the 1938 Court did not apply, and did not need, in order to decide the only question before it. The question was whether a federal ban on a cheap milk substitute, enacted at the instance of the industry it protected, required more than a rational basis. The Court said no, twice. The footnote made that answer safe to write. It did not make it a theory of minority protection. The discrete and insular minority in the case was the cannery. The political process had worked exactly as the dairy states had designed it to work. What the textbooks call the birth of modern constitutional law is the moment the Court announced it would no longer notice.