Narrow Section 26 hyphen rule to drop only predicate-position cases

[Longman006]

Narrow Section 26 hyphen rule to drop only predicate-position cases

The problem

Section 26 currently tells writers to strip hyphens from compound modifiers like well-known, high-quality, real-time, data-driven, client-facing uniformly. The Before/After example demonstrates exactly that — every hyphenated compound modifier loses its hyphen, regardless of position.

That produces ungrammatical English in attributive position:

• a high-quality report is correct (standard attributive hyphenation)
• a high quality report is wrong

A user (or an automated tool) applying the rule literally to text that uses standard attributive hyphenation will end up with grammatically incorrect prose.

What the actual AI tell is

The pattern that signals LLM output isn't any hyphenation of compound modifiers — it's uniform hyphenation that doesn't distinguish position:

• AI: the report is high-quality and a high-quality report (hyphen in both)
• Human: the report is high quality but a high-quality report (hyphen only in attributive)

Humans hyphenate inconsistently — typically only when the compound is attributive (sits before the noun it modifies) and often drop the hyphen in predicate position (after a linking verb).

Proposed change

Reword the Problem statement to spell out the attributive-vs-predicate distinction:

Problem: AI hyphenates these uniformly, including in predicate position (the report is high-quality). Humans hyphenate inconsistently — typically only when the compound is attributive (a high-quality report) and often dropping the hyphen otherwise (the report is high quality). Keep attributive-position hyphens; drop them when the compound follows the noun.

Rework the Before/After so the same compound modifiers appear in both positions, and only the predicate-position hyphens get dropped:

Before:

The cross-functional team delivered a high-quality, data-driven report. The team is cross-functional, the report is high-quality, and the methodology is data-driven.

After:

The cross-functional team delivered a high-quality, data-driven report. The team is cross functional, the report is high quality, and the methodology is data driven.

This makes the contrast actually illustrate the human-vs-LLM pattern: the attributive uses keep their hyphens (correct English); only the predicate uses lose them (where the AI tell lives).

Context

This was caught during an external code review on a private repo where this skill is vendored as a dependency. The reviewer (independent of either the vendor or the maintainer) flagged the over-broad rule because it would corrupt prose if humanizer is applied as instructed to writing that uses standard attributive hyphenation.

Happy to adjust the wording or examples if you'd prefer a different framing — the main concern is that the rule as written can produce ungrammatical output.