Travis Gilly (Real Safety AI Foundation) has posted Children as Canaries: Disaggregating the Sentinel Claim into Three Mechanisms of Early Harm on SSRN. Here is the abstract:
The claim that children function as canaries for harms later reaching adults is intuitive, rhetorically powerful, and used loosely. It is applied indiscriminately to cases sharing only the fact that children were hurt. This paper does not argue that every harm announces itself in children first; harms attaching to adult roles have no pediatric precursor and none should be sought. It argues that a real and recurring pattern in the documented record is being obscured by the imprecision, and that the sentinel claim collapses three mechanisms with different causal structures, different evidence, and different remedies. Mechanism one, sentinel proper, describes harm registering first in children through developmental susceptibility and becoming legible in adults only later: leaded gasoline, where 170 million Americans were exposed as children and roughly 824 million cumulative IQ points and 151 million excess mental disorders arrived decades after the pediatric signal; the Great Smog, where prize cattle sickened before the human toll was counted; radioiodine concentrated through milk at Nevada and Hanford; and swill milk in New York. Mechanism two, targeted selection, describes harm in which children were chosen because they were children: the devshirme, which took roughly 200,000 boys; the Mamluk system, which excluded its own elite’s sons and so required perpetual purchase; Canada’s residential schools, through which more than 150,000 children passed and at least 4,200 died; Willowbrook; Fernald; and the North Carolina Eugenics Board, which sterilized more than 7,600 people, some as young as ten. Fernald establishes the decisive point: a federal advisory committee found the radiation doses unlikely to have caused physical injury and concluded the studies wronged the boys regardless. Mechanism two is not a dose story. Mechanism three, captive absorption, describes children neither selected nor more susceptible who could not leave: asbestos in roughly 34,800 schools, lead in public housing, Flint, and 4 million lead service lines still in the ground. Only the first is a sentinel effect. The second and third are childism. The paper closes on a recourse asymmetry running from North Carolina, where fewer than 220 of more than 7,600 sterilized people were ever compensated, to the contemporary platform cases, where the barriers defeating child plaintiffs are general rather than child-specific, and therefore will not yield to child-specific advocacy.
To receive new posts from Legal Theory Blog by email, get a free subscription to Legal Theory Stack.
Lawrence Solum