In 2012 the ENCODE consortium announced that at least eighty percent of the human genome is 'functional,' a figure that appeared to overturn the long-standing view that the great bulk of our DNA is inert - so-called junk. The claim provoked immediate and unusually sharp rebuttal from evolutionary biologists, and the controversy, though often cast as a dispute about data, is better understood as a collision between two incompatible definitions of a single word. ENCODE's operational criterion for function was biochemical: a stretch of DNA counted as functional if it was transcribed into RNA, bound by a protein, or marked by a particular chemical modification. By that standard the eighty-percent figure is unremarkable, even conservative, for the genome is a biochemically busy place, and transcription in particular is promiscuous. Critics, however, deployed a different and older criterion, rooted in natural selection: a sequence is functional only if it does something the organism needs - something whose disruption selection would tend to oppose, and whose presence selection therefore explains. On this 'selected-effect' definition, mere biochemical activity is no evidence of function at all, since much transcription is incidental noise that selection neither maintains nor removes. The distinction is not pedantic. A transcribed sequence under no selective constraint can drift freely, accumulating mutations without consequence; a sequence whose disruption is opposed by selection cannot. The two definitions therefore make divergent predictions about conservation across species, and here the evidence favors the critics: only about ten percent of the human genome shows the signatures of purifying selection, a figure sharply at odds with eighty percent and strikingly consistent with the older estimates ENCODE claimed to have overturned. To announce that most of the genome is functional because most of it is biochemically active is, the critics charged, to redefine the term so broadly that the interesting question - how much of our DNA is maintained because it matters - is quietly assumed away. It would be a mistake, however, to treat the episode as a simple case of one side's error. Each definition answers to a legitimate research program. The biochemist mapping regulatory elements needs a catalogue of what the genome does, molecule by molecule, and for that purpose selective history is beside the point. The evolutionary biologist asking why a genome has the contents it has needs a criterion that tracks selection, and for that purpose biochemical bustle is a distraction. The error lay not in either definition but in the rhetorical slide between them: in presenting a biochemical inventory as though it settled an evolutionary question it was never designed to address. The word 'functional,' doing double duty, concealed the equivocation. What the controversy ultimately illustrates is a hazard peculiar to sciences that inherit their vocabulary from earlier and less precise eras. A term like function carries the authority of long use while harboring more than one meaning, and a measurement impeccable under one meaning can be gravely misleading when reported under another. The remedy is not to ban the word but to specify, on each occasion, which sense is intended.