Anatomy of Rightness── Whose rightness, and what kind, is in a reviewer's words? Between assertion and advice, where does conscience stand?

This installment borrows tools from the philosophy of language and the philosophy of law to take apart two words we use every day in materials review: "right" and "wrong." Right for whom? Is it an assertion or advice? And how many layers is rightness made of? I look for the conditions under which a reviewer's words earn trust, starting from concrete scenes in the review room.

01The Day I Wrote "This Is Not Right"

I was reviewing a brochure for a new product. The adverse reaction information sat two page-turns away from the page describing efficacy. In the comment field I wrote, "This layout is not right," and sent it back. The person in charge fixed it without a word. The comment went through, the material was corrected, and as a piece of work that should have been the end of it.

A few days later, that person came to my desk and asked, "Was that a rule violation? Or was it your opinion?" I could not answer on the spot. No notification (= the practical instruction documents issued by regulators) says "place adverse reactions next to efficacy." So was it my opinion? But to say I made someone redo their work on mere personal taste — that didn't feel right either. The ground under those few words, "not right," started to wobble the moment someone asked about it.

Come to think of it, a reviewer's words don't get through because they are always correct. They rest largely on the trust that makes the other person think, "If this person says so, I'll listen." Philosophy calls this epistemic authority (= authority about knowledge; the power to move people through trust in what you know, not through rank). Did that person quietly fix the brochure because they were convinced — or because I was the reviewer? If the latter, I moved someone by authority alone, without showing my grounds.

So in this essay I want to work through three questions in order.

• Whose rightness was my "not right" for? The patient, the physician, the company — or myself?
• Was it an assertion to be obeyed, or advice that could be declined? Which did I mean, and which did the other person hear?
• What, in review work, does "rightness" even consist of? What must be satisfied for something to be right?

The short answer: none of these questions has a single, solid answer. But there is a difference in weight between writing "not right" while staying vague, and writing the same words knowing exactly where the vagueness lies. Let me start the dissection with the first question: right for whom?

02Change Chairs, and Rightness Changes

Say there is a graph on my desk — results of a clinical trial (= a study that tests efficacy and safety in actual patients) comparing a new drug with an existing one. Sometimes I leave the graph where it is and, in my head, change the chair I'm sitting in. The same sheet of paper looks different from each chair.

The strange thing is that none of these "rightnesses" is a lie. Consideration for patients, honesty toward physicians, the company's business, the regulator's order — each makes sense on its own terms. And yet the four cannot coexist in one piece of text. Write out every limit of efficacy and the patient may grow anxious; put reassurance first and the physician finds it thin.

The economist Amartya Sen (= an Indian-born Nobel laureate in economics, known for his work on poverty and justice) called this positional objectivity (= a "positioned" objectivity: anyone standing in that position would see it that way). From the ground, the sun and the moon look the same size. That is not an illusion — as long as you stand on the ground, it looks that way to everyone. It is a fact tied to a position. Sen's point is this: objectivity is not a god's-eye view seated in no chair at all. It is how things look once you have declared your position — "this is how it looks from this chair." Rightness seen from the patient's chair and rightness seen from the physician's chair can both be objective, as long as the position is declared.

Up to here it is a pleasant thought experiment. But to be honest, my review room holds one more chair, a hidden one: the chair of self-protection. From that chair, the "right" wording is whatever keeps me from being blamed if trouble comes later. Make people revise on the strict side, and the reviewer stays safe. Unlike the other four, this chair is one you don't want to admit you're sitting in — even to yourself. Was that fifth chair mixed into my "not right" that day? If I follow Sen, this may be the chair I should declare first. Which chair was I speaking from? A "not right" without that declaration becomes a voice with no owner — words that make people obey without saying where they were spoken from.

03The Same Sentence Arrived as an Order

To a second-year employee I once said, "This data isn't the primary endpoint (= the yardstick chosen in advance as the trial's main measure of success), so please delete it." The next week, when the material came back, I was lost for words. Not just the one spot I had flagged — similar wording had been scrubbed from every page. Graph footnotes, the supplementary data column. Even entries I had thought fine to keep were gone, root and branch.

He said, pale-faced, "I thought there might be the same problem elsewhere." I had meant to hand him one piece of input for his judgment. To his ears, it had arrived as an order. Same wording, entirely different landing. For a while afterward I kept turning over what that gap really was.

The clue came from the British philosopher J.L. Austin's speech act theory (= a philosophy of language holding that to say something is itself to do something — to promise, to command). Austin argued that words are not merely tools for stating facts. Say "I promise" and a promise comes into being. Say "I order you" and an order exists. Saying is doing. His intellectual heir John Searle organized this further, classifying utterances into assertions, directives, promises, and so on (= a taxonomy of speech acts, sorted by what the speaker takes responsibility for). With an assertion, the speaker is responsible for its being true. With an order, responsibility extends to the consequences of moving the other person. Advice sits between: you are responsible for the quality of the input, but the decision stays with the other person.

Which was my remark that day? In my own mind, it sat between assertion and advice: the factual point that "this isn't the primary endpoint," plus my read that "so deletion is probably appropriate." But inside the relationship of reviewer and reviewed, the same sentence functioned as a directive. If it doesn't pass review, the material never reaches the world. Because of that difference in power, my "this isn't X" sounded, in his ears, like "delete X." The character of an utterance is not fixed by its wording alone. Who says it, and from what position, changes what the words do.

I had not handed over input for a judgment. Without knowing it, I had issued an order. And if it landed as an order, then part of the responsibility for that excessive response — a page-by-page purge — belongs to me, the one who gave it. Only when I saw it that way did I feel I finally understood his pale face.

04Whether to Take It Is Your Call

On another day, a similar scene came around again. This time I changed how I put it: "Deletion is one option. If you keep it, the question becomes how to add a note on the limitations of the source paper (= the parts that one study alone cannot settle)." A few days later, the person in charge brought a third option: take the data out of the body text and instead explain the trial design one notch more carefully, so readers could weigh the data for themselves. Honestly, it was better than my idea.

What was different? Last time I handed over an answer; this time I handed over options and the point at issue. Research on advice taking (= the field of psychology that studies how people accept others' advice) offers a clue. A wide-ranging review of past experiments by the researchers Bonaccio and Dalal reports a tendency: advice is more readily accepted when offered with room to choose rather than as compulsion. A tendency, not a law. Still, it fits what I see in practice. People use their own heads on judgments they chose themselves. On a judgment forced on them, there is no reason to.

A reviewer not monopolizing the answers is not an abdication of responsibility. I have come to think it is the opposite: an investment in the other person's judgment. Epistemic authority (= the other person's recognition that "what this person says is worth hearing") does not grow with the number of orders given. It grows out of accumulated experience — the input you handed over kept turning out to be sound. In a word, trust. Someone you moved by orders will wait for orders next time too. Someone who moved on input will go looking for input on their own next time.

• What I hand over as advice: judgments about wording, structure, whether to keep or cut. The decision stays with the other person.
• What I hold as assertion: lines that touch safety. Adverse reaction information, the handling of contraindications (= conditions under which the drug must not be used). Here I never say "one option would be."

Since imposing this distinction on myself, I spend a little longer choosing the words of a comment. Is this advice, or an assertion? Is it something the other person may decide, or something they must not? One breath before speaking. That single breath is my own remedy for never again producing that pale face.

05The Strength to Say "I Could Be Wrong"

Some years ago, a phrase I had passed as "no problem" was reclassified as inappropriate when a notification was revised. A junior colleague, going back through past materials against the revised standard, found my review record and brought it to me. In the comment field, in my own handwriting from back then: "This wording is correct. No revision needed." No citation of grounds, no reservation. A flat verdict. Failing to foresee the revision was nobody's fault. What embarrassed me was the way I had written it. A "correct" that read as if time stood still — and I had written it without a second thought.

Since then I have attached one qualifier to my comments: "Judged acceptable in light of the current notifications and the evidence at hand." I have been told it's long-winded. I have been asked whether I lack confidence. But this is not about confidence. Philosophy has a position called fallibilism (= the stance that any judgment can be mistaken). The nineteenth-century philosopher Peirce advanced it, and later Popper approached the same territory from a different angle with "falsifiability" (= the scientific requirement that a claim be stated in a form that could be shown wrong). Popper's point is simple: a statement that claims to be "absolutely never wrong" is, precisely because it cannot be tested, actually the weakest kind. Only statements that admit they could be wrong get tested, corrected, and survive.

Translated into review comments, it goes like this. "This wording is correct" does not say correct as of when, or against what — so the moment the notification changes, it collapses whole. And after the collapse, no one can reconstruct why the judgment was made. By contrast, a comment that says "acceptable in light of Section 3 of current Notification A and the package insert (= the official document of precautions for each drug)" tells you automatically, if Notification A is revised, where to look again. The comment carries its own expiry date and its own entry point for re-inspection.

Softening a verdict is not weakening a judgment. I would say the opposite. A judgment offered with the window for refutation left open (= an opening through which someone can later point out "this part was off") stays useful as time passes. A verdict with the window shut is impressive only while it happens to be right, and becomes a mere relic the moment it isn't. The review record my colleague brought me was, for me, the specimen of that. These days I regard writing in a correctable form not as a reviewer's weakness but as part of the craft. Write so that some future person can correct your comment. That is my answer to the embarrassment of that day.

06The Rightness of the Conclusion, and of How It Was Reached

At a review meeting, we once decided to withdraw a piece of material. I still believe the decision itself was sound. The citation of evidence (= scientific supporting data) was strained, and the content could not pass against the notifications either. The problem was how we decided. We were pressed for time, reached the conclusion before hearing the responsible department out, and the minutes recorded only the result. Later the person in charge told me, "I accept the conclusion. But I won't forget that I wasn't allowed to say a single word in that room." The conclusion was right, and only distrust remained. That one has stayed lodged in me for a long time.

I have the opposite experience too. On a case with a conclusion even more unfavorable to the other side — a full replacement of the material — I first showed the relevant passage of the notification, heard the department's counterarguments, and recorded, with reasons, which of those counterarguments we adopted and which we did not. The conclusion was heavy, yet the person in charge said, "I'm not convinced, but I have no complaint about how it was decided." Setting the two cases side by side, I finally saw it: rightness has two layers.

The philosophy of law calls this the distinction between substantive fairness and procedural fairness (= a basic concept: the rightness of what was decided and the rightness of how it was decided are separate things). The fact that a sense of fairness in the process shapes trust in the outcome has itself been confirmed again and again in social psychology experiments.

The troublesome part is that when the second is missing, it drags the first down with it. Someone who distrusts how you decide will look at your next conclusion through the lens of "decided one-sidedly again, no doubt." The rightness of review accumulates — or leaks away — as credit, not in single conclusions but in a relationship that repeats. Sen, mentioned earlier, argued in The Idea of Justice for removing the plain injustices in front of us rather than designing a perfectly just system, and for exposing decisions to public reasoning (= the parties exchanging reasons and examining them in an open forum). Our review meeting is a small forum, but I think the core of what it does is the same. The perfect review standard will never arrive. Even so, the plain unfairness of "decided without a chance to object" can be removed from today's meeting.

So I have decided to treat three unglamorous things — review records, stated reasons, a route for objection — not as paperwork but as part of rightness itself. Writing dissenting views into the minutes. Attaching reasons to conclusions. Telling people, "If you object to this judgment, here is the procedure for requesting a rehearing." None of these changes a single character of the conclusion. And yet a review that has them and a review that lacks them give the same conclusion an entirely different weight. How you decide is not a container for the conclusion. It is, in itself, the other layer of rightness.

07Which Layer of Rightness, and for Whom

Back to the opening scene. In the meeting room, a young colleague asked me, "Is that a rule, or is it your opinion, Suwa-san?" At the time I could not answer. If I were asked the same question now, I think I could do a little better. First, I would restate it: "Good question. Let me answer it in parts."

As this series has shown, the "rightness" a reviewer speaks of is not one solid block. It has at least four layers. First, accuracy: does it match the evidence as a matter of fact? Second, the soundness of the process: was the path to the conclusion a reasoned one? Third, honesty toward the reader: is it offered with nothing hidden? Fourth, endurance over time: is it a judgment you could reread in five years without shame? These are four separate yardsticks, and a perfect score on one with zero on the others earns no trust. Cite the numbers accurately, and you will still be doubted if the decision was made behind closed doors. Decide with care, and you are still dishonest if you fold the inconvenient warnings into small print.

And how to answer the question turns on one more axis: is it an assertion, or advice? "This wording violates the notification. Please revise it" is an assertion, and I bear the responsibility of stating the regulatory layer of rightness flatly, grounds attached. "I would write it this way. The reason: it reduces the reader's chance of misunderstanding" is advice, and the other person is free to decline. What Sen urged, again, was that people should declare where they see and speak from. Borrowing that, what a reviewer needs is not the face of an all-knowing judge but a declaration of position: "Up to here is the rule; from here on is my opinion." Only when that line is disclosed do a reviewer's words become checkable — and because they are checkable, they can be trusted. As Austin's speech act theory showed in section 03, a review comment is not a book report; it is the act itself of stopping a material or making it change. Being an act, the person who spoke must be able to explain which layer of rightness was exercised, and for whom.

Then what, in the end, underwrites that explanation? Not, I think, a chart of virtues. Integrity, fairness, humility — slogans on the wall do not work at five o'clock on a Friday before a deadline. What works is a much more concrete procedure. After finishing a comment, picture one face — just one is enough — of a patient beyond the physician who will read this material. Then reread the sentence you just wrote, and if needed, choose again. That much, and only that, is what I want to call conscience. Conscience is not a virtue perched somewhere high. It is the habit of pausing once, for a few seconds, just before you hit send. "Is that a rule, or your opinion?" As long as I keep answering that question honestly, a reviewer's rightness stays what it should be: not for the reviewer, but for the person who, at the end of it all, holds the medicine in their hand.