A Floor on Two Axes ── Not Letting Desktop Detection Pass

The morning at the materials review office, Yui saw the desk of a man named Minami for the first time. A chart of problem types, plastered with sticky notes. \"Deviation comparison,\" \"implied superiority,\" \"out-of-bounds extrapolation\"—the technical terms lined up in neat order. Minami can name every one of the problems. And yet Mio will not make him an independent reviewer. The reason lay in a single piece of material and one thin line.

He Could Name Every Type

That day, Minami recited the problem types of review material in front of Yui. Using the explanatory material for a cardiovascular treatment drug as his subject, he counted on his fingers: "There are six deviations that tend to happen with this kind of material." How data is cropped, how the control group is hidden, how an effect is reworded. The textbook headings came straight out as his own voice. Yui was honestly impressed. "Minami, you've got it all memorized."

Mio silently placed a different sheet in front of Minami. An actual proof copy. "Look at this as if you were doing an independent review. If there's a problem, catch it."

Minami read it for ten minutes and looked up. "No major problems here. The wording is reasonable too."—That material contained one trick: partway up, the vertical axis of a graph changed its scale. The very "deviation comparison" he had named first on his chart was happening right in front of him.

Why a Floor Can't Be Measured by 'Height' Alone

Yui asked Mio about it later. "Minami's knowledge is around L3 (the top of the detection skill that lets you review independently), isn't it? Why won't you pass him?"

Mio drew a horizontal axis and a vertical axis on the edge of the paper. "If you only look at the single yardstick called L, Minami really is high. But detection skill can't be measured by height alone. You read it as coordinates."

"Minami is high in abstraction and narrow in field of view. We call this 'desktop detection.' On the desk—that is, inside his head—he can see everything. But on the actual item, he can't catch it."

Think of Airport Security Screening

Mio offered an analogy. "Think of training for a security screener. There's someone who has memorized every type of dangerous item. Knives, liquids, batteries—if you ask him to explain, it's perfect. But sit him in front of an actual X-ray monitor and he lets the blade on the screen pass right through. Can you put this person alone in charge of an inspection station?"

Yui shook her head. "No way. Screening is meaningless if you can't catch it on the real item."

"Right. Materials review is the same. A dangerous piece of material goes out into the world on one person's judgment. A medical professional reads it and decides on a prescription. So what we ask of a reviewer isn't 'being able to talk about types' but 'being able to catch the danger in the single sheet in front of you.' No matter how fluent the talk, it doesn't become a floor."

Yui remembered Higuchi from before. Higuchi was outstanding at explanation and persuasion, but weak at finding danger. "So Higuchi's weak spot is in a different place." "Different. Higuchi is strong at persuasion and weak at detection. Minami is strong at knowledge and weak at detection—more precisely, strong at abstract detection but weak at detection on the real item. So the prescription is different too."

Pass Only What's Above the Diagonal

Mio drew a single line across the coordinates, from lower left to upper right. "This is the real body of the qualifying line. The floor isn't a single point called the L value—you impose it along this diagonal."

"If you only look at the L display, Minami looks like L3, because his talk is full marks. But his field of view S-hat doesn't reach 2. He can't catch it on the real item. So no matter what the L number is, he doesn't pass as independent. Height that strays off the diagonal isn't accepted as a floor."

Yui felt she understood the meaning of the line. "If you're fooled by the L number and pass someone who's only high in abstraction but narrow in field of view—" "You end up putting someone who overlooks danger on the real item alone in charge of the inspection station. That's the scariest thing."

How to Handle Minami

So is Minami a fail? Mio shook her head. "Minami isn't like Higuchi. He has the foundation of knowledge. What he lacks is only the field of view. So it's a conditional pass—we won't let him work independently, but with supervision he can join the review."

Taking notes, Yui asked, "Supervision—concretely, what does that mean?" "Someone with a wide field of view checks the material Minami reviewed once more on the real item. We make use of Minami's power of abstraction and just plug the overlooks on the real item with another set of eyes. It's because we read detection skill on two axes that we can write a prescription like this."

Mio said at the end, "If we measured people on the single point of L, Minami would have been 'L3, so independent OK.' Because we read it as coordinates, it became 'knowledge is high but field of view is narrow, so supervised.' Defining the floor on two axes—that's what this means." In Yui's notebook were a diagonal line and one point sitting just a little below it. That was Minami's position.

Minami could talk about problem types better than anyone. Even so, he did not become an independent reviewer. The power to talk (abstraction A-hat) and the power to catch it in the single sheet in front of you (field of view S-hat) are separate axes, and the qualifying line is drawn only on the diagonal where both axes are present at once.

The positions of Higuchi and Minami are now fixed. What remains is Wada. That plain, unremarkable man who can still catch danger on the real item—where will he land on the diagonal? The day the three verdicts line up side by side is near.