A man crossing a river is struck by another boat. If someone is in it, he rages. If it is empty, he pushes it aside and rows on. Same collision, same damage, no anger. On the passenger we install in every bump.
One note on the source: this story sits in the outer chapters, generally considered a later layer than the Cook Ding material in essay 3. Different hands, different periods, same compilation. It does not weaken the story, but I would rather not present a multi-authored text as one voice.
The experiment is elegantly controlled. Same river. Same impact. Same damage to the hull. One variable changes and the entire experience changes with it.
The empty boat hits you: mild annoyance, a push, back to rowing. The occupied boat hits you identically: fury, the shouting, the story about the other person’s carelessness, the grievance carried for a week.
Which locates the suffering somewhere unflattering. Not in the collision. In the occupancy.
Psychology named the installation. The tendency to explain other people’s behavior by their character while explaining our own by circumstance has been studied for half a century under the heading of attribution error.3
I should note that the strong universal version of this has taken real criticism — the effect varies across cultures and situations more than the textbook presentation suggests, and it is better described as a common bias than a law of human cognition. What is robust enough to build on is the narrower observation: we generate character explanations quickly, circumstantial ones slowly, and the quick one arrives first and sets the emotional tone.
There is a related and more clinically studied version, hostile attribution bias, where people prone to aggression are more likely to read ambiguous actions as deliberately hostile.4 That is a well-supported finding about a specific population, and it illustrates the mechanism cleanly: the same event, read for intent, produces a different person’s afternoon.
The teaching’s move is to reverse the default. Generate the circumstantial reading first, as a discipline, because the character reading will show up on its own.
The online version of the river is collision by design. The quote-post, the reply, the pile-on — every disagreement arrives with a face attached and every criticism reads as aimed.
There is research on how moral and emotional language spreads through networks faster than neutral language, which is at least consistent with what the architecture rewards.8
What makes it hard is that online, the boats frequently are occupied. Someone did write that deliberately. The teaching works best on ambiguous collisions, and a medium engineered to produce unambiguous ones is a poor place to practice it.
This is the teaching in the series most likely to be used to talk someone out of a justified reaction.
Sometimes it was deliberate. Sometimes the person really is treating you badly, and the generous circumstantial reading is a way of not dealing with it. Applied indiscriminately, the empty boat becomes a machine for excusing other people’s behavior and swallowing your own response — and it lands hardest on people already inclined to assume they are the problem.
It is also wrong for patterns. A single ambiguous collision deserves the charitable reading. The fourth one from the same boat is data, and treating it as another accident is not equanimity.
The version I would defend: use it on first collisions and ambiguous ones, where it is usually right and always cheaper. Do not use it to file down a pattern. And notice that the question is “what if the boat is empty,” not “the boat is empty” — it is a question precisely because sometimes the answer is no.
Note — the fourth collision is not an accident
Applied to a pattern, it becomes a way of excusing behavior that should be named, and it does the most damage to people already inclined to assume they are at fault.
The order matters: emptying theirs is a trainable reflex, and emptying your own takes considerably longer.
This list mixes primary texts, translations, scholarship and modern trade books, and the entries vary in how precisely they can be located. Quotations from the discourses follow Bhikkhu Sujato’s translations, freely available at SuttaCentral, cited by the standard sutta reference. Where a phrase is traditional and circulates in several renderings I say so rather than inventing a citation, and anything marked as a paraphrase is my own wording.
The collision is identical either way. What differs is whether you installed someone in the other boat, and the installing happens in about a second.
Ask the question — it is a question, because sometimes the answer is that the boat has a person in it, and the fourth collision from the same boat is not an accident.