English

Rules and word lists for English text. The lists at the end are read by scripts/lint.py; keep their format (one entry per line, regex entries in backticks).

Typography and grammar

Headings and titles: sentence case ("How survey weights work"), unless the publication requires title case.
Quotes: straight or curly as the target format uses; be consistent.
Compound modifiers: hyphen before the noun ("a high-quality survey"), none after it ("the survey is high quality").
Ranges in data may use an en dash (1990–2000). Do not use dashes to join clauses; see slop.md C4.
Serial comma: follow the profile; if the profile is silent, follow the existing text.
Prefer is, are, has over "serves as", "stands as", "boasts".

Academic and technical register

Keep field terms exact: headcount ratio, poverty gap, PPP, survey weights, imputation. Do not swap them for plain synonyms.
Hedges that report real uncertainty stay ("suggests", "is associated with", "under these assumptions"). Cut only stacked hedges.
The passive voice is fine in methods text when the actor is obvious or irrelevant.

Lint: error

These are near certain AI tells. Remove them.

delve
delves
delving
tapestry
a testament to
stands as a testament
in today's fast-paced
in today's world
in today's digital age
ever-evolving landscape
it's worth noting
it is worth noting
it's important to note
it is important to note
let's dive in
let's dive into
dive deep into
let's break this down
let's explore
here's the thing
here's what you need to know
without further ado
at its core
let that sink in
make no mistake
read that again
great question
i hope this helps
hope this helps
feel free to reach out
as of my last update
as an ai
plays a pivotal role
play a pivotal role
a pivotal moment
marking a pivotal
plays a crucial role
sets the stage for
paving the way for
the future looks bright
exciting times ahead
nestled
breathtaking
game-changer
game changer
embark on a journey
in the realm of
underscores the importance
serves as a reminder
the real question is
the uncomfortable truth is
in a world where
what really matters

Lint: warn

Common in AI text but also in careful human text. Fix them when several gather or when they add nothing.

crucial
pivotal
robust
seamless
seamlessly
leverage
leveraging
utilize
additionally
moreover
furthermore
notably
landscape
realm
intricate
intricacies
meticulous
meticulously
showcase
showcasing
underscore
underscores
highlighting
fostering
vibrant
enduring
garner
bolster
bolstered
multifaceted
holistic
transformative
groundbreaking
cutting-edge
really
truly
genuinely
literally
simply
actually
deeply
fundamentally
inherently
honestly
interestingly
importantly
crucially
serves as
stands as
boasts
a wide range of
navigate the
empower
elevate
resonate
align with
shed light on
key role
deep dive
at the end of the day
when it comes to
to be clear
in conclusion
in summary

Lint: regex warn

\bnot (?:just|only|merely|simply)\b[^.!?\n]{1,80}?\bbut\b
\b(?:it|this|that)(?:'s| is) not (?:just |only |merely )?[^.!?\n]{1,60}?[,;] (?:it|this|that)(?:'s| is)\b
\b(?:doesn't|does not) mean [^.!?\n]{1,60}?\. (?:it|this) means\b
\b(?:could|might|may) (?:potentially|possibly|arguably)\b
\bthe (?:data|evidence|numbers|market|decision|conversation|culture) (?:tells|tell|rewards|reward|emerges|shifts|speaks)\b
\bhere's (?:what|why|how|the)\b

Measure: hedges

Words that soften a claim (Hyland). measure.py counts them per 1,000 words. A count is a hint, not a judgment: "may" can also mean permission.

may
might
could
perhaps
possibly
probably
likely
unlikely
suggest
suggests
appear to
appears to
seem to
seems to
apparently
roughly
approximately
somewhat
to some extent
in part
tend to
tends to
generally
arguably

Measure: boosters

Words that strengthen a claim (Hyland).

clearly
obviously
certainly
definitely
undoubtedly
indeed
of course
in fact
surely
evidently
demonstrate
demonstrates
prove
proves

Measure: self mention

Measure: engagement

Words that address the reader directly.

you
your
let us
let's
note that
consider
imagine

Measure: connectors

One line per connector: the relation, a colon, the phrase.

contrast: however
contrast: but
contrast: yet
contrast: whereas
contrast: on the other hand
contrast: in contrast
concession: although
concession: even though
concession: despite
concession: nevertheless
concession: still
cause: because
cause: therefore
cause: thus
cause: hence
cause: as a result
cause: consequently
elaboration: for example
elaboration: for instance
elaboration: in particular
elaboration: that is
elaboration: in other words
elaboration: specifically
addition: also
addition: moreover
addition: furthermore
addition: in addition