A dictionary presents itself as a record of what words mean. It is a record of who was asked. Every entry is the settled residue of a decision about which uses were citable, which speakers were evidence, and which mouths produced data rather than error — and the finished page shows none of that. What it shows is a word and its meaning, in a typeface that has never argued with anyone.
That is the department's interest here. Not whether a definition is right, but that the form of a definition cannot express doubt. A sentence can hedge. A dictionary entry has no grammar for we were not sure, this was contested, the citations came from four men in one city. It has headword, part of speech, sense, and the confidence is structural — you cannot write an uncertain one without breaking the form.
Ask what a definition would look like if it recorded its own disputes, and you have described an argument. Which is why it does not.
The authority is circular and the circle is the product. A lexicographer records usage: the dictionary is descriptive, it follows the language, it claims no power to legislate. Then the entry is printed, and it is consulted — by a teacher marking an essay, a court reading a statute, an editor refusing a word. Those consultations are themselves usage. The next edition records them.
So a description becomes the thing it describes, by being believed. The dictionary does not need to claim authority; it only needs to be looked at, and the looking supplies what the claim would have. This is the cleanest instance on the site of a myth in the sense the Myth Index uses — a productive fiction acting as infrastructure, generating the reality it reports.
FOUR THINGS AN ENTRY DOES WHILE APPEARING TO DO NOTHING
It ends a sentence that was still running. Meaning in use is negotiated across a conversation — corrected, narrowed, taken back. An entry takes one cross-section and stops time on it, then that cross-section is quoted against the people still negotiating.
It converts a live disagreement into an error. Before the entry there are two communities using a word differently. After it, one of them is wrong, and can be told so with a citation rather than an argument. Nothing about the two uses changed.
It orders the senses, and the order is read as rank. Sense 1 reads as what the word really means; later senses read as extension, slang, drift. Frequency, or age, or the corpus that happened to be at hand — presented as depth.
It decides what was never a candidate. The most consequential judgement is the one with no entry at all: a use that was not collected is not marked disputed or rejected. It is absent, and absence has no page a reader can turn to.
The clearest evidence is that definitions are fought over at all. If a dictionary merely reported, the fights would be pedantry. They are not: what counts as a marriage, a refugee, a recession, a genocide, a terrorist — each has been contested by people who understood exactly what the entry does. It is not that the word changes the world. It is that the entry becomes the place the argument has to be had, and whoever holds the entry gets to make everyone else the challenger.
The technical version of this is the department's account of vocabulary as a border: a term that functions as a lock rather than a key, producing insiders and outsiders as its primary output. A definition is the lock's specification, published, so that being outside can be demonstrated rather than merely felt.
The instruction the department gives is not to distrust dictionaries. It is narrower and harder: when a definition is produced in an argument, ask what the entry was for. Somebody assembled the citations, somebody ordered the senses, somebody decided the matter was settled enough to print. The entry is the end of that work and shows none of it. Meaning is produced — and here is a whole industry for producing it, which describes itself as a record.