But they seemed to argue that the cohort born after 1889 were exposed to it as children and that's what made them weaker against the 1918 outbreak, which seems to run counter to received wisdom about immunity.

Maybe I misinterpreted and as you're saying, they were arguing that the cohort born after 1889 weren't exposed and were thus had less immune defence against the 1918 virus. It's just, that's also confusing because I thought the whole thing they were trying to explain was why people who otherwise mightn't be as strongly affected by a novel influenza virus, generally (young healthy adults), seemed in this case to be affected more so than in comparable outbreaks. If people born before 1889 had prior exposure to a similar virus to the then novel, 1918 influenza virus and were thus, a little bit better protected against it, that would manifest in an unusually lower number of older people being as badly stricken not a higher than usual number of young people. Also if the logic follows, it would imply that for young healthy people not to be as affected as they had seemed to be in 1918, or in any case of a novel influenza virus outbreak, they would normally have to have been partially exposed to a similar virus, but that would mean either that every novel influenza virus affected a cohort not usually as badly affected meaning paradoxically that in fact that cohort would seem to usually be that strongly affected by novel viruses, or that somehow coincidentally every such novel influenza virus outbreak before had had a similar preceding outbreak timed to give that cohort time to form antibodies.