The name arrives before the sentence
A widely shared quotation carries a familiar name beneath it, and our posture changes. If an ordinary person said it, we might hesitate. If Zhuangzi said it, the sentence seems to have passed a preliminary examination already.
But are those words in the book? If they are, does a character say them, does the narrator say them, or has a later explanation been printed alongside the text? Before that, what happened to the book on its way to us?
This reading series now changes direction. So far, we have mostly asked how a story opens a question. We will also begin asking whether the book always speaks in one voice, and whether an explanation appended to a story narrows what the story made possible.
That is not an invitation to a trial at which only specialists may speak. It requires me to explain what a reader can check directly, and what remains uncertain even after careful examination.
Su Shi did pick up the scissors
In 1078, the writer Su Shi composed a memorial essay for a shrine to Zhuangzi. His understanding of the book differed from treating it as straightforwardly hostile to Confucius. Some passages that appeared to attack Confucius, he thought, were actually helping him indirectly.
That view made several chapters troublesome. Su Shi suspected Robber Zhi and The Fisherman because they seemed genuinely to denigrate Confucius. He described Giving Away a Throne and Discussing Swords as shallow and failing to enter the Way.
Those are distinct reasons. The judgment about shallowness directly concerns the latter pair; it should not be made into his sole explanation for rejecting all four.
He also proposed a structural connection. Remove the four intervening chapters, and the end of Borrowed Words—the story of Yang Ziju losing his preferential treatment at an inn—seems to meet the opening of Lie Yukou, where Liezi is disturbed by preferential service on his travels. Su Shi describes realizing the connection and laughing: these belong to one chapter.
His suspicion is therefore not simply a vague feeling that the prose sounds wrong. He is trying to recover a continuity interrupted by the received arrangement.
His premise is on the desk too
The argument has force, but Su Shi does not approach it without a prior understanding. His conviction that Zhuangzi assists Confucius helps determine which passages appear out of place.
That does not automatically invalidate the reading. Without some sense of a whole, it is difficult to notice an anomaly. Yet the question must also run backward. Could the overall interpretation need revision? Could a collection contain voices that do not agree?
Two passages reading smoothly together is a clue worth pursuing. It is not the same as seeing them together in an original manuscript. An editor can discover a connection and can also create one. The distance between coherence and historical reconstruction still requires evidence.
What Su Shi leaves us is consequently more valuable than four chapter titles crossed out. He leaves his interpretation, reasons, and proposed rearrangement available for inspection. Later readers can examine them and disagree.
A judgment can be sharp without having to become the last judgment anyone is permitted to make.
We have more than “It sounds like him”
Saying that the book is ancient and therefore offers no external evidence would erase evidence that has actually survived. The bibliographic treatise in the History of the Han records a Zhuangzi in fifty-two chapters. Lu Deming's account of earlier editions lists commentaries with different chapter totals and describes selections among the Outer and Miscellaneous Chapters. It also preserves remarks attributed to Guo Xiang about material he regarded as intrusive or miscellaneous.
The familiar thirty-three-chapter text is closely tied to the transmission of Guo Xiang's commentary. These records do not restore all the lost wording, but they are not nothing. Catalogues, old commentaries, quotations in other works, and differences among transmitted versions can constrain a reconstruction.
Nor can the arithmetic between fifty-two and thirty-three simply be laid out as nineteen complete manuscripts whose removal we understand. Chapter divisions, combinations, and selection have their own history. The surviving count does not reveal every editorial action.
The responsible position is that the evidence is incomplete and many conclusions remain difficult to establish. It is not that evidence is absent, leaving the reader's ear as the only available instrument.
A smaller unit brings possibilities and risks
The coming essays will often stop at a passage rather than judge an entire chapter. A story leaves room for discovery; a following summary seems to close that room. A scene invites the reader to notice something; the next voice insists on how it must be understood. Such seams deserve attention.
But a seam is not an author's signature embedded in the paper. It may reflect different writers, the same writer at different times, a changed purpose, editorial selection, or our own misunderstanding of the preceding passage.
At least three questions therefore need separating. Is there a tension between these passages? Does it support a hypothesis of distinct textual layers? Can we assign a passage to, or exclude it from, the historical Zhuang Zhou?
Close reading may establish the first question quite persuasively. The second needs wider comparison. The third is usually harder still. An elegant answer at the first stage does not authorize automatic passage through the other two.
The aim is to preserve the attentiveness that comes from working at a smaller scale. Smaller scissors do not, by themselves, make every cut more certain.
A form of address is a clue, not a seal
Who calls whom what is worth noticing. A questioner suddenly becomes an instructor; a respectful address differs from the pattern elsewhere. These changes can give us a reason to investigate.
But ancient forms of address require attention to period, occasion, and tone. We cannot first define fuzi as a word that only a junior may use for his own teacher, then declare every use outside that rule impossible.
Courtesy, respectful address, irony, and the literary staging of a conversation also matter. If versions differ, those differences cannot be skipped on the way from one unexpected word to a second author.
An inconsistency may justify “Look at this more closely.” To justify “This cannot be the same hand” requires additional reasons. That restraint does not weaken close reading. It tells the reader exactly which claim the observation supports, and where further material is needed.
A work can be questioned by its own practice
An essay recommends independent thought while forbidding disagreement. An arrangement advertises trust while determining every action in advance. Noticing those discrepancies can be illuminating.
What they first show is whether a form of expression or a practice fulfills its own claim. They do not immediately establish a false attribution. Nor can we insist that anyone who genuinely understood the subject would be incapable of writing inconsistently.
People contradict themselves, fail, and change their manner for different audiences. Detailed explanation is not necessarily indoctrination; a suggestive silence does not necessarily respect the reader. Giving reasons can make disagreement easier by giving it something precise to engage.
I am explaining how to read while writing this essay. I cannot consistently make explicit explanation itself a mark of inferiority. The better question is whether these explanations help you return to the text and see for yourself, or merely make you accept my conclusions faster.
The writer cannot settle that question by declaring the essay successful. Readers must actually encounter the work.
Confucius need not have one permanent role
Characters also move. We have met a Confucius who teaches, one who acknowledges a limitation, and one who asks to follow Yan Hui after the student's account of sitting in forgetfulness. If we assign every character a single permanent function, many of the book's stories will be rejected before they begin.
A change in role may be evidence of another voice. It may also be the point of the story. We need to see how it works locally, not merely compare it with a character profile we prepared in advance.
Forms of address, the relationship between manner and claim, and the placing of characters will all remain useful observations. They are ways to continue reading, not three keys any one of which closes the case.
“I cannot decide that yet” may indicate not a failure to read carefully, but arrival at the edge of what the evidence will bear.
Quality and authorship need separate answers
A later addition may be excellent. A passage securely associated with a great writer may still be mistaken, incomplete, or unsuitable as guidance in a present situation.
A name does not guarantee the content. Admirable content does not authenticate a name. Liking a sentence cannot prove that a great thinker wrote it; disliking one cannot automatically remove it from that thinker's work.
For modern quotations, we can often make the distinction practical. Look for the work, edition, and surrounding passage. Where an original publication or recording is available, do not stop at the impression that the wording sounds characteristic. Style can be imitated; genuine statements can surprise our expectations.
If the source has not been found, leave the uncertainty visible. A sentence need not borrow an unverified name in order to deserve thought. It may be worth discussing even while its attribution remains unsettled.
No certificates of authenticity
The following essays will select passages worth reading and consider how they work, how they meet the text around them, and why I prefer to follow a particular line of thought. Where a judgment concerns layers or attribution, its limits belong beside its reasons.
You may agree with an interpretation of a story while rejecting a hypothesis about its composition. You may find continuity exactly where I see a seam. These need not be bundled into a single vote for or against the whole reading.
Su Shi moved four chapters aside and listened for another continuity. We can learn from the seriousness of that rereading while retaining the right to put the chapters back and ask again.
Before deciding which pages should be torn out, turn the next one. It still deserves to be read.