A reviewer "with knowledge" is not someone who has memorized a lot of rules. It is someone who, the instant they see a sentence, recalls without pause the several issues it might trip over — legal ones, disease ones, statistical ones. What matters is not the amount memorized but how densely the mind links one issue to another. This time we look at how to spot that density, using one scene: reviewing a comparison material that reports a trial showing a drug is "not inferior."
Knowledge Is a Mental Web of Associations
The internal document this is based on defines knowledge this way: "a connection web that instantly recalls the multiple regulations a single expression touches at once. What matters is not the amount of facts but the density of the web." Stiff wording, but the idea is simple. Knowledge is not "the number of facts you have stored" but "how densely those facts are linked."
A state of holding the systems of regulation, medicine, and statistics, including the connections among them.
The key phrase is "including the connections among them." Think of cooking. You can know the names of a hundred seasonings, but if your head lacks the links — "this flavor goes with this ingredient" — you cannot make a good dish. Conversely, even with few seasonings, if you feel the combinations in your bones, several ideas spring to mind from one taste. Reviewing is the same: you may have memorized a thousand provisions, but if seeing a statement does not make the related provisions link up in your head, it is useless at the desk. The real work of review is decided by speed of recall and breadth of connection. The strength of knowledge is set not by the number of dots but by the density of the lines tying dot to dot.
The Four Threads That Weave the Web
This web of connections is made of four kinds of "thread." One thick thread alone does not make a net. Only when all four are present does the scooping power appear.
Knowledge of the rules
The rules you must follow. Industry guides, advertising rules, the Pharmaceutical Affairs Law (the law on drugs). For example, the ban on exaggerated promotion (Art. 66), the ban on promoting unapproved drugs (Art. 68), and how far each reaches.
Medical and disease knowledge
How a disease works and how it is treated. For the disease the statement addresses, what matters to the patient and what invites misunderstanding.
Reading trials and statistics
How a trial is built and how to read "is there a difference." "Better than the rival" and "not worse than the rival" are different stories. You also read the meaning of confidence intervals and the margin (allowed range), which come up later.
Past cases and penalties
Past violations and their cleanup. How a similar statement was handled before — the "precedent."
Take a comparison material that shows a drug is "not worse than" the rival. First, with ③, you read that "this only shows 'not losing,' not 'winning.'" Next, with ①, you connect it to the rules on "exaggerated promotion" and "overstating." With ② you add what it actually means in that disease, and with ④ you recall a similar old case. Only when the four threads cross at one point does the judgment "this statement is risky" rise with real, solid weight. With one thread alone, it stays a mere hunch.
The Four Types of "Breadth" and "Depth"
The source re-measures knowledge with two rulers. One is scope (breadth): how far you can connect across the separate fields of regulation, medicine, and statistics. The other is abstraction (depth): do you only see the surface wording of a provision, or do you grasp "why this rule exists" — its aim? Lay these two as vertical and horizontal, and four types appear.
School math makes it clear. Just memorizing a formula is shallow; understanding why the formula holds is deep. Solving only one topic is narrow; solving problems across all units is broad. The ideal is the upper-right (L4), with both depth and breadth; the start is the lower-left (L1), weak in both. The problem is the two types that grew on only one side — these become traps.
| Type | Breadth × Depth | What state | Main path or trap |
|---|---|---|---|
| Local application (L1) | Shallow × narrow | Applies each rule by surface wording, to that one case only | Start of growth |
| Systemic reconstruction (L4) | Deep × broad | Rebuilds the whole from its aim and sets standards across fields | Goal of growth |
| Principle-heavy | Deep × narrow | Speaks deeply on one field's aim, but the range is narrow and cannot be connected | One-sided trap |
| Case-exhaustive | Shallow × broad | Memorizes a lot but never reaches the "why," so cannot apply to a first-time case | One-sided trap |
These two traps are easy to miss. The principle-heavy person speaks deeply and looks "capable," but the range they can handle is narrow. The case-exhaustive person knows a lot and seems dependable, but freezes when a statement with no precedent arrives. Both are unfinished, grown on one side, and neither is L4. So when developing people, the aim is to stretch depth and breadth at the same time — in the chart, to follow the diagonal line from lower-left to upper-right.
The "Signs" Visible from Outside
The density of connections cannot be told from self-report. "I have knowledge" measures nothing. Like a health check, you measure not by how the person feels but by numbers and behavior visible from outside. The source's list of "signs of a strong person" are all concrete actions you can watch with your own eyes.
Recalls at once
The moment they see a statement, they name several related provisions without pause. Not speed of "looking up," but speed of "recall."
Says the connection out loud
"This statement overlaps both the exaggeration issue and the overstating issue" — they voice the connection aloud.
States it without the materials
They state the required items and figures accurately without opening the materials at hand.
Locates new information themselves
When a new notice comes out, they can place it: "this part of my knowledge changes."
What the evaluator should watch is not "do they know the right answer." It is "do the words tying issue to issue come out of their mouth." Especially ② — saying the connection out loud — is the moment the density of the mental web surfaces most plainly.
Measuring with a Comparison Material — L1 to L4
The source builds the ruler from how four levels handle the same single scene. The scene is fixed: "show a comparison material carrying a 'not worse' trial result." Looking at the same material, what differs from person to person? It is the same method as a sports coach showing the same game footage to a beginner and an expert and measuring the gap in their comments.
| What to look at | L1 Local application | L2 Linking | L3 Integration | L4 Systematization |
|---|---|---|---|---|
| Position (breadth×depth) | Shallow × narrow | Medium × medium | Deep × mid–broad | Deep × broad |
| Shape of the web | Dots only | Dots become lines | Lines become a plane | The plane becomes a solid |
| How the mind works | Applies each provision by surface wording, case by case | Bundles the main rules at the requirement level, applies them one by one | Understands from the "why," surveys the connections among provisions | Rebuilds the whole rule set from its aim and turns it into a cross-field standard |
| Behavior on the spot | Opens the guide, then notices "the trial design probably needs to be stated" | Without the materials, points out "say whether it 'won' or is 'not worse,' and state the result accurately" | Connects it: "this overlaps exaggeration and overstating — to prevent the reader's misunderstanding" | Writes up an in-house rule for "not worse" trials in general and turns it into training material |
L1 notices only after opening the guide. Knowledge sits as scattered dots, used only for that one case. L2 can bundle the requirements and point them out without the materials — dots become lines, with the main rules grasped at requirement level. But it is still applying them one by one.
At L3 the quality changes. Seeing the same "not worse" statement, the reviewer surveys the connections among provisions — "this overlaps exaggeration and overstating" — and traces back to the aim: "to prevent the reader's misunderstanding." From this stage, where lines become a plane, density worth calling a "connection web" first appears. L4 turns that plane into a solid: not stopping at one judgment, it writes up an in-house rule for "not worse" trials in general and leaves it to the organization as training material. One person's mental web becomes the whole organization's web.
One caution: these four levels show the order, but they are not an equal-spaced ruler. L1 to L2 is just a difference in how requirements are bundled. But L2 to L3 is a jump in quality, from "applying one by one" to "surveying connections," and the step height differs. L3 to L4 is yet another kind of jump — a step that turns one person's knowledge into an organizational standard. On a staircase, each step has a different height, and the higher you go, the taller the step. So not treating the L3-to-L4 gap as "just one more step" is the crux when designing how to develop people.
Competency Framework ── Map of all 10 episodes
- Vol. 1: The Essential Question ── Who Detects, Pushes Back, and Embeds the Gap ── Series opener in plain language: the essential question, the three roles and eight-part map, and how to read each part (essence, four-box view, scale).
- Vol. 2: Seeing People on Two Axes ── Quadrant, and Scale/Level/Divergence ── A map that sees ability in two directions — breadth and depth — and the basics of keeping the ruler, the reading, and the gap apart.
- Vol. 3 (this episode): Knowledge ── Not Volume but the Density of the Connective Web ── Reframes the reviewer "with knowledge" as not someone who has memorized many facts but someone whose mental web of associations is dense — where one expression instantly links regulatory, medical, and statistical issues. Using a drug-comparison material, it shows the gap from beginner (L1) to organization-wide standard-setting (L4).
- Vol. 4: Intelligence ── Seeing Through Form to Judge by Substance ── Not fitting knowledge to a case but stretching it past the edge: taking the label off and reasoning from basic principles to read what is really going on. Measured on a four-step L1-L4 scale and four types built from scope and depth.
- Vol. 5: Risk Detection ── Reading What Is Not Written ── Catching what was left out and what is merely hinted at, by picturing what goes through the reader's head. The hardest part of perception, and the "all-theory-no-practice" trap.
- Vol. 6: Intuition ── The Alarm That Precedes Words ── A mental alarm that rings — "something feels off" — before you can put the reason into words. Its value is speed and the first flag that says "check this one carefully." But raising the flag is never the verdict; you always confirm it afterward.
- Vol. 7: Communication ── A Correctness That Doesn't Land Doesn't Exist ── Communication means re-saying your judgment in words the listener can grasp — a kind of translation. From handing over the bare fact (L1) to building a shared wording anyone understands (L4), read it by two things: how wide the audience is, and how much you reword.
- Vol. 8: Behavior-Change Inducement ── Intrinsic, Even When Unwatched ── The power to get someone to produce it right next time on their own — even when nobody is watching. We trace it from the "fixes it only because told" stage (L1) to the stage where "not doing it" becomes the team's air (L4), with one axis: is the person motivated from inside?
- Vol. 9: Relationship Building ── Neither Enemy Nor Ally, a Trusted Third Party ── The seventh skill: keeping your distance (staying independent) yet still being trusted. Getting too cozy fails, and so does being combative. The right route is the diagonal toward "strict but fair."
- Vol. 10 (final): Trust Density ── The Medium That Makes It Work, and the Whole ── The same point passes or bounces depending on who raises it. The gap is not brainpower but the thickness of trust a person has built up over time. Three things — never wavering, being readable, never bending to power — harden over years into an asset you cannot buy today. This finale folds all eight skills into one picture and hands off to the next series.
Measure knowledge by "amount memorized," and only heavy memorizers score high while the case-exhaustive trap (broad but never reaching the "why") stays invisible. Measure it by "density of connection," and the axis changes: from one "not worse" statement, how many issues rise, and are they tied back to the aim? At L3 lines become a plane; at L4 one person's web is left to the organization as its standard. Show the same fixed scene and read where the person sits on the diagonal — this is the starting point of developing people.
The next installment takes up the second "seeing" skill, Intelligence: the power to derive from principle the cases the rules do not directly state, peel off the surface form, and see the substance. Not "applying" knowledge but "stretching" it onto an unfamiliar case.
- Knowledge strength is "density of links," not "amount memorized." As cooking depends more on a sense of combination than on the number of seasonings, measure by how densely the mind ties issue to issue from one expression.
- The two types grown on one side are traps. Neither principle-heavy (deep but narrow) nor case-exhaustive (broad but never reaching the "why") is L4. Only growth along the diagonal, stretching depth and breadth together, is the main path.
- L1–L4 is an ordinal scale with uneven steps. Distinguish the L2→L3 jump ("from applying one by one to surveying connections") and the L3→L4 jump ("from individual knowledge to organizational standard") by showing the same comparison material.
- McClelland, D. C. Testing for Competence Rather Than for Intelligence. American Psychologist, 1973. (Origin of the idea of measuring ability by actual behavior, not by amount of knowledge or aptitude tests)
- Boyatzis, R. E. The Competent Manager. Wiley, 1982. (Defines ability as "behavior you can see with your eyes")
- Spencer, L. M. & Spencer, S. M. Competence at Work. Wiley, 1993. (Separates knowledge and skill from the deeper motives behind them and offers a behavior-based ruler)
- Dreyfus, S. E. & Dreyfus, H. L. A Five-Stage Model of the Mental Activities Involved in Directed Skill Acquisition. 1980. (Staged improvement from novice to expert — the theoretical backdrop for L1–L4)
- MHLW. Guideline on Sales Information Provision Activities for Prescription Drugs. 2018. (Regulatory primary source bearing on judgments of "not worse" and comparative expression)