Reasoning (A-33): The operational criterion for classifying a nutrient as essential is: when it (and only it) is removed from an otherwise-complete diet, a specific deficiency sign develops that is corrected by restoring that nutrient. The design isolates choline as the sole manipulated variable - a semisynthetic diet identical between arms except for choline, with adequate methionine and folate so that the methyl-group and de-novo-phosphatidylcholine (PEMT) pathways are not artificially starved. Under that control, the deficient arm alone shows (1) ~30% depletion of plasma choline and phosphatidylcholine and (2) a rising ALT indicating incipient hepatocellular dysfunction, while every other liver/renal test and the control arm stay flat; both abnormalities reverse when choline is returned in week 5. The counterfactual is supplied by the concurrent control group, so the changes cannot be attributed to hospitalization or the semisynthetic diet per se. The crucial subtlety the argument turns on: humans do synthesize choline de novo via PEMT, which is why choline had not been classified essential; but this experiment shows that under normal (non-excess) methionine/folate that endogenous route is insufficient to prevent deficiency, so intake is still required. Removal-plus-reversal with a matched control is a strong (near-deductive within the essentiality framework) warrant for the essentiality conclusion; the main limitation is the small, short-term, male-only sample, which bounds generalization of thresholds but not the qualitative essentiality claim.
Validity verdict (step 6)
status: approved, checked. Reconstruction: the step is subsumption of the experiment under an operational criterion — “removed as the sole variable from an otherwise-complete diet → a specific deficiency sign appears → it reverses on repletion, with a concurrent control ruling out non-specific causes ⇒ essential.” Load-bearing implicit premises surfaced: (i) the ALT rise (0.42→0.62 µkat/L, deficient arm only, reversing on repletion) counts as a genuine deficiency sign of hepatocellular dysfunction rather than a bare biomarker wobble, and (ii) adequate (non-excess) methionine/folate means the PEMT/de-novo route was not artificially starved, so its insufficiency is real, not an artifact of methyl-donor deprivation. Conditional on the premises, the subsumption is near-deductive within the essentiality framework and traces cleanly. Undercutting-defeater probe: candidate defeater is “hospitalization/semisynthetic-diet per se caused the change” — blocked by the concurrent choline-continued control staying flat; a second candidate is “one reversible enzyme elevation is too thin to be a deficiency sign” — this attacks premise (i) (truth), not the inference, so it is priced later as a likelihood, not a validity failure here. No weaker conclusion is forced; qualitative essentiality (scoped to non-excess methyl donors, matching H-39) holds as stated. checked: the criterion and its application are elementary, traced without leaning on source authority.