The same sentence on every item conveys no information

What it is

Ten recommendations sit under “because you watched similar content,” “because you might like this.” The syntax is locked. Even if the slot words differ slightly, people quickly learn: this line does not move with the item. Boilerplate recommendation reasons are one sentence pattern laid on every row, so between-item information is flattened by syntax. The reason rail can no longer aid open/skip, and it cannot be a handle for rebuttal.

Semantic vacuum is “this sentence is compatible with any item.” Boilerplate is “even the grammatical shell is copied.” Vacuum can swap synonyms; boilerplate is the form itself saying “do not bother reading.”

Why it happens

People read information from difference. Once syntax freezes, attention marks the reason rail as decoration and skips stop going through it. Even if a slot word on one row is a faithful feature, it is buried in a template already ignored, and never reaches the decision. Rebuttal fails too: denying “similar content” denies nothing, because every row says it.

Templates usually come from scale: one copy frame, fill a feature name. Scale’s cost is that the frame eats discriminability. If you want to keep the syntax tidy, slot words must contrast sharply across rows; when the slots are homogeneous too, tidiness is zero information.

Studying it

Three presentations: no reason, a distinct sentence per row, one sentence for the whole list with only slot words changing. Measure: gaze on the reason rail, recognition of between-item difference, whether open/skip uses the reason, whether a rebuttal can be aimed at one row. Independent variables: whether syntax is locked, discriminability of slot words, list length. Dependent variables: gaze time, recognition, rate of judging the rail “not worth reading.”

Split boilerplate from vacuum: one condition says empty things in different syntax; another says specific features in one syntax. If the latter still goes unread, the failure is formal repetition, not semantics.

Where it stops holding

On a rail of two or three items, one syntax is not yet marked as decoration. In accessibility, stable syntax is sometimes a virtue (a screen reader can predict it); then discrimination must live entirely in the slot words, and those words must sound different when read aloud. A section heading may need one lead-in sentence; that is not an item reason. This entry is about copied syntax. It does not fold in post-hoc hallucination or tracking grain.

Applying it

• Ban one sentence skeleton for a whole rail. Switch skeletons by kind of basis: “because you finished X” and “now trending” must not share “you might like this.”
• Slot words must contrast on neighbouring rows: two filled with “similar content” count as boilerplate and must go back to a concrete title or class.
• Check: hide covers, keep reasons, ask people to pair which reasons could swap as whole sentences. High swap rate is boilerplate. Add replay or gaze: if the reason rail is no longer looked at after the third row, the form has already taught people to ignore it.

Related

• Same group: L6.07.1 A reason helps the user decide whether an item is worth opening, not to persuade them to open it · L6.07.2 Reasons must come from the actual ranking basis; a post-hoc story is another kind of hallucination · L6.07.3 Reasons expose data sources and can make users feel tracked · L6.07.4 Reasons make recommendations rebuttable, so users can correct a wrong profile
• Nearby: L6.01 Recommendation Reasons · L5.05 The Moderation Principle of Transparency · L1.03 Visualizing Uncertainty
• Search terms: boilerplate recommendation reasons · template why · identical explanation