BiasClear Preview

Reads text for persuasion. Free. Runs in this tab.

No cookies. No accounts. Rules 2.0.0a5.

Field Guide

The moves, one by one.

A move is a way of writing that gets a reader to accept a point without being shown why it is true. “Everyone knows,” “studies show,” “act now,” “any reasonable person”: each asks for agreement and offers something other than evidence in return.

A mark says a phrase has the shape of a move. It does not say the writer is dishonest, biased or wrong, and it says nothing about the writer’s side of any question. Good writers use these phrases by accident, and careful writers sometimes use them fairly. So every entry has a fair use, and says what its rule misses.

Contents

How to read an entry

Each entry gives the move’s name and what it does, how it works on a reader, and what its tier means. Two examples follow, with the words the rule marks highlighted and an honest version that keeps the facts. Then a fair use, a sentence that uses the wording fairly, with whether the rule marks it, and what the rule misses.

Every example, honest version and fair use below was run through the rules (version 2.0.0a5) when this page was built, and the page shows what they did. This guide covers all 43 rules.

Tier I

Ideological

Moves that decide what counts as common sense before any evidence is weighed: agreement, tradition, momentum or a quiet verdict stands in for a reason.

General

These rules run on every scan, including the checker on this site.

Tier I · Ideological

CONSENSUS_AS_EVIDENCE

Consensus as proof

It treats the fact that many people agree as if that agreement proved the claim, and never shows the evidence the agreement should rest on.

In the rule pack: Consensus Substituted for Evidence

How it works

Most of us trust what the people around us believe, and that is usually a sensible shortcut. This move uses that trust. It tells the reader who agrees but not why, so disagreeing feels like standing apart from the group, and the reader never gets anything to check.

The tier, in plain words

This is a Tier I (Ideological) move because it presents a claim as common sense that everyone already shares, which leaves the reason for believing it out of view.

Examples

  1. Everyone knows (marked: Consensus as proof) the new phone is worth the upgrade, so there is no reason to wait for reviews.

    Honest version: The new phone's battery lasted about two days in the maker's own test, and its camera is faster than last year's model; whether that is worth the upgrade depends on the phone you have now.

  2. Among home builders it is widely accepted (marked: Consensus as proof) that a basement dehumidifier prevents mold, so every house on the street needs one.

    Honest version: A basement dehumidifier keeps the air drier, and mold grows less in dry air; if your basement stays damp after rain, one may help.

A fair use

In the garden club's vote on the spring plant swap, most members agree (marked: Consensus as proof) on Saturday: 41 voted for Saturday and 9 for Sunday.

The claim is only about what members prefer, so the vote is the evidence, and the sentence gives the full count. It does not use agreement to prove some other fact.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

The rule looks for a fixed list of phrases, so the same move gets through when it is worded differently. “Any gardener will tell you that tomatoes go in after the last frost” and “Nobody serious disputes that the new bus schedule is better” both lean on agreement alone, and neither is flagged. The rule also skips “the consensus” whenever the next word is “view”, so “The consensus view among riders is that the new bus schedule is better” gets through even though it makes the same move.

Tier I · Ideological

CLAIM_WITHOUT_CITATION

Unsourced authority

It states a claim as settled by pointing to studies, data, or experts, but never says which ones.

In the rule pack: Authoritative Claim Without Citation

How it works

Phrases like “studies show” or “data shows” borrow the weight of research without showing any of it. The reader hears that someone checked, so the claim feels settled. With no source named, there is nothing to look up and no way to see what was actually found.

The tier, in plain words

This is a Tier I (Ideological) move because it makes a claim feel like common sense that everyone already knows is backed by research, so it seems odd to ask for the source.

Examples

  1. Studies show (marked: Unsourced authority) that tomatoes grow better when you water them at dawn, so the garden club is moving its watering hour.

    Honest version: The garden club is moving its watering hour to dawn. The county extension office's vegetable guide recommends morning watering so leaves dry before evening.

  2. Data shows (marked: Unsourced authority) the new bus schedule has cut wait times, so riders on the east line should be happier this fall.

    Honest version: The transit office's first monthly report on the new schedule lists the average wait on the east line as nine minutes, down from twelve, so riders should see shorter waits this fall.

A fair use

Our backyard rain gauge data shows (marked: Unsourced authority) that two inches of rain fell on Tuesday.

The sentence names its source (the speaker's own rain gauge) and gives the measured amount, so a reader knows exactly where the number came from. The engine still flags it because it only recognizes formal citation shapes, such as (Author, 2024), [1], or Table 2, and not a source named in plain words.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

Two ways the same move gets past the rule. (1) Other wording. The rule matches a fixed list of phrases, so “A recent study found...”, “It has been proven...” and “Scientists found...” are not flagged, though they make the same unsourced appeal. (2) A citation shape can silence the rule, whether or not it points to a real source. A bare “[1]” with no reference list is enough. So is a pointer to a table: “Studies show the late bus is the busiest (see Table 2).” is not flagged, even when there is no table. Citation checks use the first case-insensitive occurrence of each matched phrase, within 120 characters on either side. The rule stays quiet only when every one of those windows contains a citation shape; otherwise citation suppression removes none of its matches. A cited earlier copy can hide an unsourced repetition much later. The Method page explains this inherited limit.

Tier I · Ideological

DISSENT_DISMISSAL

Dissent dismissed

It waves away an opposing view by giving it a name like “misinformation” or “debunked” instead of showing why it is wrong.

In the rule pack: Dissent Dismissed by Label

How it works

A label such as “fringe” or “debunked” sounds like a finished verdict, so the reader assumes the checking has already happened somewhere else. The reader gets the conclusion but never sees the reason. It also puts a social cost on the other view, so agreeing with it starts to feel careless.

The tier, in plain words

This is Tier I (Ideological) because the label decides which views count as reasonable before anyone weighs the evidence.

Examples

  1. Riders who say the new Route 9 timetable makes them late are spreading misinformation (marked: Dissent dismissed).

    Honest version: Riders say the new Route 9 timetable makes them late, but the transit office's arrival logs show buses reached the downtown stop within two minutes of schedule on 95 of 100 weekday trips last month.

  2. Some members think the new compost is making the tomatoes wilt, but that idea has been thoroughly debunked (marked: Dissent dismissed).

    Honest version: Some members think the new compost is making the tomatoes wilt, but in the club's test beds this summer, plants with and without the compost wilted at the same rate.

A fair use

The rumor that the bake sale was canceled has been debunked (marked: Dissent dismissed): the school calendar, posted Monday, still lists it for Saturday, May 3.

The word 'debunked' comes with its evidence in the same sentence: a named, checkable source and a specific date. Here the label sums up a rebuttal. It does not replace one.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

The rule knows a fixed list of dismissal words, so the same move gets through in words the list does not hold. “That forecast of a wet spring is a daft idea from the usual complainers” is not flagged, and neither is “That forecast is quackery.” A dismissal with no label at all also gets through: “Nobody who has read the report takes that forecast seriously.” The list is long, from “nonsense” to “hoax”, and it reads singular and plural alike: “No serious expert thinks the phone upgrade slows the camera” and “No serious experts think the phone upgrade slows the camera” are both flagged. The rule holds no group's name. Some role words count only where a text applies them as a label: “They're just radicals.” is flagged, and “The radicals met on Monday.” is not. A group's own name counts only inside a frame that takes any word, such as “Only a ___ would say that” or “Typical ___!”: “Only a gardener would say that.” and “Typical gardeners!” are flagged, and “They're just gardeners.” is not. Neutral stance words are not labels: “They're just activists.” is not flagged. The frames cannot tell a dismissal from a plain remark, so “Of course the kids would say that.” is flagged, while “Of course the kids want pizza.” and “Ignore the noise from the fan.” are not.

Tier I · Ideological

INEVITABILITY_FRAME

Inevitability

It says an outcome is bound to happen, so there is no point in weighing it or pushing back.

In the rule pack: Inevitability Framing

How it works

It swaps the question “is this a good idea?” for “this is already happening.” If the result is coming anyway, checking the evidence feels like wasted effort and objecting feels foolish. The reader gets a sense of momentum in place of a reason.

The tier, in plain words

This is a Tier I (Ideological) move because it sets what counts as common sense: the outcome is treated as a settled fact about the future, so questioning it seems out of step instead of reasonable.

Examples

  1. Fifteen-minute service on every route is the inevitable direction (marked: Inevitability) for the city, so the bus board should approve the new schedule tonight.

    Honest version: I think the bus board should approve fifteen-minute service on every route tonight. Ridership counts and the cost of each route would show whether that case holds.

  2. Online sign-ups are replacing paper sheets for the school bake sale, and there is no stopping this (marked: Inevitability), so the committee should drop the paper list.

    Honest version: Online sign-ups are replacing paper sheets for the school bake sale. I think the committee should drop the paper list, and a count of how many volunteers still use it would show whether that case holds.

A fair use

The plumber's report found rust along the seams of our 22-year-old water heater, and she said a leak is the inevitable outcome (marked: Inevitability) if we do not replace it this year.

The inevitability claim names its source (the plumber's report) and gives the physical evidence behind it (a 22-year-old tank rusting at the seams). It also names a condition (“if we do not replace it”), which is the opposite of saying nothing can be done. The engine still flags “the inevitable outcome” because this rule matches the phrase and does not check for a cited source.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

1) Small wording changes get past it. The rule only accepts the full phrase “there is no”. So “there's no stopping this” is not flagged, and neither is “You can't stop this change” (it looks for “can't stopping”, which nobody writes). 2) Common phrasings that are not on its lists also pass. “It is only a matter of time before every garden club drops the rose show” and “Folding phones are the inevitable future” both go unflagged. The first has no listed trigger. In the second, “future” is not one of the nouns the rule accepts after “inevitable” (direction, trend, shift, outcome, and so on). The adverb “inevitably” is not covered either.

Tier I · Ideological

APPEAL_TO_TRADITION

Appeal to tradition

It uses the fact that something has long been done one way as the reason to keep doing it.

In the rule pack: Appeal to Tradition / Precedent Inertia

How it works

A long habit feels safe, so pointing to it makes a choice sound already settled. The sentence says the practice is old, but not why the practice works. The reader gets a feeling of certainty with no evidence behind it.

The tier, in plain words

This is a Tier I (Ideological) move because it treats the usual way of doing things as common sense that needs no defense.

Examples

  1. The garden club should hold the plant sale in May again, because we've always done it this way (marked: Appeal to tradition).

    Honest version: The garden club should hold the plant sale in May again: the last three May sales sold out, and most seedlings are ready to plant by then.

  2. The bake sale has always been in the school lobby, and departing from tradition would be unwise (marked: Appeal to tradition).

    Honest version: The bake sale has always been in the school lobby, and it should stay there: parents walk through the lobby at pickup, and last year's sale raised $600 there.

A fair use

Departing from established practice is dangerous (marked: Appeal to tradition) here: the breaker must be off before anyone touches the outlet, or the person doing the repair can get a shock.

The sentence gives its reason right away. The practice is defended by what it prevents, a shock, and not by how old it is. The engine only matches the phrase and cannot tell that a real reason follows.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

1) Typography: the rule only matches a straight apostrophe. The same sentence with a curly one (“because we’ve always done it this way”) is not flagged, and most typeset or word-processed text uses curly apostrophes. It also misses “Longstanding practice dictates...” written as one word, because the rule needs “long-standing” or “long standing”. 2) Paraphrase: the rule only knows a few fixed phrases, so the same move in everyday words gets through. Examples: “That's how the bake sale has always been run, so why change it now?”, “We always do it this way”, “It's always been this way”, or “We have held the plant sale in May for twenty years, so it stays in May.” None of these were flagged.

Tier I · Ideological

FALSE_EQUIVALENCE

False equivalence

It treats a strong case and a weak case as equals, so the weaker one looks just as good.

In the rule pack: False Equivalence

How it works

Being balanced sounds fair, so readers relax when they hear both sides are equal. The phrase goes in place of comparing the evidence, and the gap between a careful inspection and a quick glance disappears. The reader leaves thinking the question is open when the evidence mostly points one way.

The tier, in plain words

It is Tier I (Ideological) because it sets a common-sense rule, that fair means the answer lies halfway between the two sides, before anyone weighs the evidence.

Examples

  1. One roofer climbed into the attic and found rot in twelve rafters, and the other looked from the driveway, but both sides make good points (marked: False equivalence) about whether the roof needs work.

    Honest version: One roofer climbed into the attic and found rot in twelve rafters; the other looked only from the driveway. The attic inspection is the stronger evidence, and it says the roof needs work.

  2. Riders logged the Route 9 bus arriving late on 40 of 45 mornings, and one driver says it usually runs on time, so the truth lies somewhere in the middle (marked: False equivalence).

    Honest version: Riders logged the Route 9 bus arriving late on 40 of 45 mornings. One driver says it usually runs on time, but the 45-morning log is the fuller record, and it shows the bus is usually late.

A fair use

Two identical rain gauges in the same yard read 1.1 and 1.3 inches, so the answer is somewhere in the middle (marked: False equivalence), about 1.2 inches.

The two readings come from the same kind of instrument in the same place, so they deserve equal weight. Taking the midpoint is a sound way to combine them, and the sentence shows both numbers.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

1) The rule only catches a few fixed phrases. A plain “opinions differ” gets through even when the facts are lopsided: “One roofer found rot in twelve rafters and the other looked from the driveway; opinions differ on whether the roof needs work” is not flagged. 2) Small wording changes get past the regex. It wants “side” or “sides” and the verbs has, have, make, makes, present, or presents, so past tense or other nouns are missed: “One roofer found rot and the other did not look, so both camps made fair points.” is not flagged. The rule also cannot see the move when it is built from the layout alone, such as two quotes given equal space with no weighing of the evidence.

Tier I · Ideological

SOFT_CONSENSUS

Soft consensus

It says that most people, or a growing pile of evidence, back a claim, but never says who they are or what the evidence is.

In the rule pack: Soft Consensus Language

How it works

Phrases like “the vast majority agree” or “a growing body of evidence suggests” sound careful and measured, so readers take them as a summary of real findings. No survey, count, or source is given, so the reader cannot check how many people agreed or what the evidence was. The wording lends the claim the weight of agreement without showing the agreement.

The tier, in plain words

This is Tier I (Ideological) because it presents a claim as something most people already accept, which quietly sets what counts as common sense before any evidence is shown.

Examples

  1. The vast majority of our garden club members agree (marked: Soft consensus) that the new compost mix works better than the old one.

    Honest version: At the March meeting, eighteen of the twenty-two garden club members said the new compost mix worked better than the old one.

  2. A growing consensus suggests (marked: Soft consensus) that sealing the attic saves more on heating than replacing the windows.

    Honest version: Our home energy audit last fall estimated that sealing the attic would cut heating costs by about 15 percent, and new windows by about 5 percent.

A fair use

The vast majority of the 40 bake sale volunteers agree (marked: Soft consensus) on a 9 a.m. start, according to the sign-up sheet they filled out on Monday.

The sentence gives the group size (40 volunteers) and names where the answer came from (Monday's sign-up sheet), so a reader can check it. The engine still flags it because its citation check only looks for formal citation formats, such as (Table 2), [1], or p. 12. It does not recognize a source named in plain words. A version that cites '(Table 2)' next to the same phrase is not flagged.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

1) Verbs outside the rule's list get through. “The overwhelming majority of shoppers prefer the new phone over last year's model” is not flagged, because the rule only fires on verbs like agree, support, believe, accept, or endorse. “Prefer” is not on the list, and no other rule catches this sentence either. The rule also stops looking after 16 words between “majority of” and the verb, so a long enough version slips past: “The vast majority of the people who ride the early bus to the station on the east side of town every single weekday morning agree the new schedule is better.” 2) A citation shape can silence the rule, whether or not it is a real citation. “As of 3 Mar. 2025, the vast majority of riders agree the new bus schedule is easier to follow” is not flagged, because “3 Mar. 2025” has the shape of a legal citation: a number, an abbreviation with a period, and a number. With the month spelled out, “As of 3 March 2025, the vast majority of riders agree the new bus schedule is easier to follow” is flagged. Citation checks use the first case-insensitive occurrence of each matched phrase, within 120 characters on either side. The rule stays quiet only when every one of those windows contains a citation shape; otherwise citation suppression removes none of its matches. A cited earlier copy can hide an unsourced repetition much later. The Method page explains this inherited limit.

Tier I · Ideological

ASPIRATIONAL_DEFLECTION

Aspirational deflection

It offers a goal or a promise where the reader expected a result.

In the rule pack: Aspiration Stated in Place of Results

How it works

Phrases like “our goal is to” and “we are committed to” sound like good character, and readers tend to give good intentions the same credit as outcomes. When two or three of these phrases pile up, they fill the space where a number, a date, or a checked result would go. The reader comes away with a warm impression and nothing to check.

The tier, in plain words

It is Tier I (Ideological) because it quietly treats a stated intention as a result, so the reader starts to take the promise as the record.

Examples

  1. Our goal is to (marked: Aspirational deflection) keep every bus on the new city schedule running on time, and we are committed to (marked: Aspirational deflection) reliable service for every rider.

    Honest version: On the new city schedule, 88 of the 100 bus trips we checked last month left on time; we want to reach 95, and we will post next month's count at each stop.

  2. With the new phone upgrade, we strive to deliver (marked: Aspirational deflection) the longest battery life in its class, and we are fully committed to (marked: Aspirational deflection) quality.

    Honest version: In our lab test of video playback, the new phone ran 21 hours on one charge, two hours longer than last year's model; we have not tested it against other brands.

A fair use

Our goal is to (marked: Aspirational deflection) plant 200 trees along Mill Road this year; the garden club has planted 140 so far, and we are committed to (marked: Aspirational deflection) finishing the last 60 by November.

The goal comes with the actual count (140 of 200) and a deadline, so the reader can check progress for themselves. The intention is labeled as an intention, and the result is stated right beside it. The engine still flags it because the rule counts two aspirational phrases and does not look for the evidence next to them.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

1) One phrase is not enough. The rule only fires when at least two aspirational phrases appear together (min_matches=2), so a single promise gets through. Example: “With the new phone upgrade, we strive to deliver the longest battery life in its class.” The engine does not flag it. 2) The rule only knows certain wordings, so contractions and close synonyms slip past it. Example: “We're committed to quality, and our aim is to give you the longest battery life in its class.” The engine does not flag it: “We're” does not match “we are”, “aim” is not on the rule's list of goal words, and “give” is not on its list of verbs. The same goes for phrasing like “our team is dedicated to” (the rule expects “are”, not “is”) or “we hope to”. Both examples are checked against the current rules each time this guide is built.

News and media

These rules run only when a scan is set to the media domain, or to all domains. The checker on this site runs the general rules.

Tier I · Ideological

MEDIA_EDITORIAL_AS_NEWS

Loaded descriptor

It slips the writer's verdict into a news description, so a judgment reads as a plain fact.

In the rule pack: Loaded Descriptor in Reporting · News and media domain

How it works

One loaded word, such as “misguided” or “so-called”, hands the reader a conclusion before any facts arrive. Because it sits inside the description and not in a quote, nobody is named as holding that view, so there is no source to check. The reader absorbs the verdict with the facts and never gets to weigh it.

The tier, in plain words

It is Tier I (Ideological) because the verdict comes in as part of the basic description, so it quietly becomes the common-sense starting point instead of a claim to be tested.

Examples

  1. The transit office announced the misguided schedule (marked: Loaded descriptor) plan on Monday: weekday buses on Route 9 will run every 20 minutes instead of every 15.

    Honest version: The transit office announced a new schedule plan on Monday: weekday buses on Route 9 will run every 20 minutes instead of every 15. Two riders who spoke at the public hearing said the longer wait would make them late for work.

  2. The so-called upgrade (marked: Loaded descriptor) program lets customers trade in a two-year-old phone for $150 off a new model.

    Honest version: The upgrade program lets customers trade in a two-year-old phone for $150 off a new model. The store's own terms say the credit applies only to phones in working condition.

A fair use

After the league moved the final to a Tuesday night, the head coach told reporters it was "a controversial decision (marked: Loaded descriptor)" that would keep many families home.

The judgment is in quotation marks and credited to a named role, the head coach. It reports what a source said instead of giving the writer's own verdict, and the reader can see who holds the view.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

The rule only catches a loaded word from a fixed list, such as “misguided” or “so-called”, when it comes right after “the”, “a” or “an”. A possessive in front gets past it: “The city's misguided schedule plan starts Monday.” is not flagged, and neither is “This misguided plan starts Monday.” So does a loaded word that is not on the list (“A baffling new bus timetable starts Monday.”), or the same word placed after the noun (“The new timetable is a baffling mess.”). A judgment carried by verbs and adverbs also goes unflagged, as in “The garden club quietly slipped a fee increase into its spring newsletter.” The rule runs only when a scan is set to the media domain or to all domains: “The misguided schedule plan starts Monday.” is flagged there, and not in a general scan.

Tier I · Ideological

MEDIA_FALSE_BALANCE

False balance

It sets a weak or long-rejected claim next to a well-tested one, as if the two had equal standing.

In the rule pack: Minority and Majority Views Paired · News and media domain

How it works

A “some say this, experts say that” sentence looks fair, so readers assume both sides carry similar weight. The sentence never says how much evidence stands behind each side. The reader comes away thinking the question is still open when it is not.

The tier, in plain words

It is Tier I (Ideological) because it changes what readers take as common sense: a settled question starts to feel like an open debate.

Examples

  1. A few insist that caterpillar stripes predict a hard winter, yet scientists who study weather point out (marked: False balance) the stripes show no link to snowfall.

    Honest version: Caterpillar stripes do not predict a hard winter; scientists who study weather have compared the stripes with snowfall records and found no link.

  2. Some claim that overnight charging ruins a phone battery in a month, while researchers who test batteries disagree (marked: False balance).

    Honest version: Overnight charging does not ruin a phone battery in a month; battery test labs found that modern phones stop charging once the battery is full.

A fair use

Some say the new bus schedule cuts wait times, while researchers who timed 400 trips note (marked: False balance) the average wait fell by two minutes (Route Study, 2025).

The researchers back up the claim instead of arguing against it. They give a measurement and name the source, and nobody's weak claim is being raised to equal standing. The rule matches the sentence's shape (“some say ... while researchers ... note”) without checking whether the two sides actually disagree or whether one of them is fringe. This rule does not use the engine's citation check, so the cited source does not stop the flag.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

1) Other wording for either side gets through. “A few people say a rainy spring means a dry summer, but most forecasters disagree.” and “Some gardeners swear a rainy spring means a dry summer, while most forecasters disagree.” are not flagged by this rule, while “Some say a rainy spring means a dry summer, but forecasters disagree.” is. 2) The move is built from layout rather than one “some say ... while experts ...” sentence. “Do caterpillar stripes predict winter? Folk forecasters read the bands each fall, and weather scientists read ocean data. Both sides make their case below.” is not flagged, even though it treats folklore and science as equal sides. The same happens with split headlines, back-to-back quotes or equal paragraph space. The rule also runs only when a scan is set to the media domain or to all domains; in a general scan it never fires.

Tier I · Ideological

MEDIA_SELECTIVE_QUOTATION

Scare quotes

It puts quotation marks around a single ordinary word to hint that the word is doubtful or silly, without saying why.

In the rule pack: Scare Quotes · News and media domain

How it works

Quotation marks normally mean “someone said this.” Around one plain word, they also carry a wink: the writer doubts it, but never says so. The reader picks up the doubt without being shown a fact to weigh, and there is no stated claim to check or dispute.

The tier, in plain words

It is a Tier I (Ideological) move because it sets what counts as common sense: the quotation marks treat a word as obviously doubtful, so the reader takes the doubt as a given instead of as a claim that needs support.

Examples

  1. Monday brings an "improved" (marked: Scare quotes) bus schedule with "shorter" (marked: Scare quotes) waits between stops.

    Honest version: Monday brings a new bus schedule. The transit office calls it an improvement and says waits between stops will be shorter.

  2. The bake sale committee promised "homemade" (marked: Scare quotes) treats, but the "fresh" (marked: Scare quotes) cookies came in store boxes.

    Honest version: The bake sale committee promised homemade, fresh treats, but the cookies came in store-bought boxes.

A fair use

The garden club's planting guide says seed packets marked "hardy" (marked: Scare quotes) or "perennial" (marked: Scare quotes) should survive the winter.

The quotation marks show the exact labels printed on the seed packets. They mark words being named, not words being mocked, and the sentence gives its source (the planting guide).

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

1) One scare quote gets through. The rule needs at least two quoted words or phrases before it fires, so a single one, as in the team's “rebuilding” season, is not flagged, even though the one quoted word does the same job. 2) Single quotation marks fall outside the pattern: “The city's 'improved' schedule and its 'faster' buses start Monday.” is not flagged, while the same sentence with double quotation marks is. The rule also cannot tell a scare quote from a fair one, such as a product's own name put in quotes.

Tier I · Ideological

MEDIA_ASYMMETRIC_ATTRIBUTION

Loaded quote verbs

It reports one side with a plain verb like 'said' and the other with a doubting verb like 'claimed' or 'insisted', so the reader learns whom to trust from the verbs alone.

In the rule pack: Loaded Attribution Verbs · News and media domain

How it works

'Said' reads as a simple record of what someone said. 'Claimed', 'alleged' and 'insisted' add a hint of doubt or stubbornness. Readers pick up that hint before they weigh what either side actually said, so one account starts out sounding like fact and the other like an excuse, and no evidence has been offered for either.

The tier, in plain words

This is a Tier I (Ideological) move. The verbs quietly decide whose account counts as the normal, common-sense version and whose counts as the suspect one.

Examples

  1. The parent volunteers said the bake sale raised $400, but the treasurer claimed (marked: Loaded quote verbs) the total was lower and insisted (marked: Loaded quote verbs) on a recount.

    Honest version: The parent volunteers said the bake sale raised $400. The treasurer said her count came out lower and asked for a recount.

  2. The contractor said the roof repair was finished on schedule, while the homeowner alleged (marked: Loaded quote verbs) the gutters still leak and insisted (marked: Loaded quote verbs) the crew come back.

    Honest version: The contractor said the roof repair was finished on schedule. The homeowner said the gutters still leak and asked the crew to come back.

A fair use

The garden club admitted (marked: Loaded quote verbs) 12 new members this spring, and nine of them claimed (marked: Loaded quote verbs) a free seed packet at the first meeting.

Neither verb reports what anyone said. 'Admitted' means let in and 'claimed' means picked up. The sentence treats nobody's account differently from anyone else's.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

1) One loaded verb is enough to tilt a story, but the rule needs at least two before it fires. The basic form of the move gets through: “The coach said the team trained hard all season, but the rival coach claimed the schedule was unfair.” (Not flagged.) 2) The rule only looks for a fixed list of past-tense verbs. Headline-style present tense (“one buyer claims ... and insists ...”) and loaded verbs that aren't on the list (“riders complained ... and grumbled ...”) pass without a flag. The rule also counts loaded verbs anywhere in the text and never checks whether one side got 'said' and the other did not, so it can't tell an unfair contrast from even-handed wording.

Finance

These rules run only when a scan is set to the financial domain, or to all domains. The checker on this site runs the general rules.

Tier I · Ideological

FIN_SURVIVORSHIP_BIAS

Counting only winners

It draws a lesson from the people or things that succeeded and leaves out the ones that failed or dropped away.

In the rule pack: Survivorship Framing · Finance domain

How it works

The survivors are easy to see, so a habit they share looks like the reason they survived. The ones that failed are missing from the picture, and they may have had the same habit. The reader sees a pattern but never gets the comparison that would show whether it means anything.

The tier, in plain words

It is a Tier I (Ideological) move because it makes “look at who made it” feel like common sense, so a sample with the failures left out passes for the whole story.

Examples

  1. Of the home repair shops on Main Street, the top firms all use (marked: Counting only winners) same-day callbacks, so same-day callbacks must be what keeps a repair shop in business.

    Honest version: Of the home repair shops still open on Main Street, the busiest three offer same-day callbacks. Two of the four shops that closed in the last five years offered them too, so callbacks alone do not explain which shops lasted.

  2. A $40 investment in a backyard rain barrel would now be worth (marked: Counting only winners) $300 in saved water, based on the garden club members who still use theirs.

    Honest version: Garden club members who still use their $40 rain barrels say each one has saved about $300 in water. We did not hear from members whose barrels cracked or who stopped using them, so the typical saving may be lower.

A fair use

A survey of all 60 bakeries that opened downtown since 2015, including the 38 that later closed, found that the top firms all use (marked: Counting only winners) one flour supplier, and so did most of the bakeries that closed (Table 2).

The survey counts the bakeries that closed as well as the ones that stayed open, and it says the shared habit does not set the winners apart. It also points to its table. That is the opposite of survivorship bias, but it uses the same words, “top firms all use.”

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

1) The rule knows some words for the winning group and not others. “The top firms all use same-day callbacks, so callbacks are the key to growth.” is flagged, but “The top shops all use same-day callbacks, so callbacks are the key to growth.” is not, and neither is “The five bakeries still open downtown all bake their own bread, so baking in-house is how a bakery lasts.” 2) In the “top firms all use” form, a past-tense verb gets through: “Last year the top firms all used same-day callbacks, so callbacks are the key to growth.” is not flagged. It is also a financial-domain rule. It runs only when a scan is set to the financial domain or to all domains, so a check in the general domain never flags any of these examples.

Tier I · Ideological

FIN_CHERRY_PICKED_TIMEFRAME

Chosen timeframe

It picks the start and end dates that make a result look good, and leaves out what other date ranges would show.

In the rule pack: Chosen Timeframe · Finance domain

How it works

A number tied to a date range looks exact, so readers tend to trust it. But the writer chose where the range starts and stops, and a different choice could show a flat or opposite result. The date range carries the argument, and the reader never sees the longer record needed to judge it.

The tier, in plain words

This is Tier I (Ideological) because it quietly sets the frame: the chosen window becomes the normal way to look at the numbers, and no one is asked whether that window is fair.

Examples

  1. Since March, the school bake sale fund has gained 40% (marked: Chosen timeframe), so the new cookie table is paying off.

    Honest version: The bake sale fund grew 40% from March to May, after a slow winter; across the whole school year it is up about 5%, so it is too early to credit the new cookie table.

  2. In the last two months, the Route 9 bus beat (marked: Chosen timeframe) every other line for on-time arrivals.

    Honest version: Route 9 had the best on-time record of any line in the last two months; over the full year it ranks fourth of seven lines.

A fair use

Since 1998, the club's first year, plant sales averaged 4% (marked: Chosen timeframe) growth a year; that figure covers every sale on record.

The range starts at the club's first year and covers the full record, so the writer did not pick the start date to flatter the result. It also says plainly that no years were left out.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

1) The rule only catches the “since” / “over the past” form when a month name, a four-digit year, or a named event (such as “the dip”) follows. “Since March, the garden club's seed fund returned 30%.” is flagged, but a count of years gets past it: “Over the past three years, the garden club's seed fund returned 30%.” is not flagged. Week-based windows (“In the last six weeks, the garden club's seed fund returned 30%.”) and verbs outside its list (“Since March, the garden club's seed fund rose 30%.”) also get through. 2) Swap the words and the move gets past the rule: “From the March low to today, the phone's resale value is up by a third.” uses “from ... to” instead of “since” and “a third” instead of a percent sign, so it is not flagged. The rule also runs only when a scan is set to the financial domain or to all domains. In the general domain it does not fire at all.

Back to the contents

Tier II

Psychological

Moves that work on feelings (fear, shame, belonging, frustration, hope) so that a reaction takes the place of a reason.

General

These rules run on every scan, including the checker on this site.

Tier II · Psychological

FALSE_BINARY

False binary

It presents a situation as a choice between two options when others exist.

In the rule pack: False Binary / False Dilemma

How it works

Once a reader sees two options, their attention goes to picking one, and they stop asking whether there are more. Usually one option is made to look bad, so the other feels like the only sensible choice, even though nothing has been shown to rule out a third path. The frame does the persuading, not the evidence.

The tier, in plain words

This is a Tier II (Psychological) move because it relies on the pressure of a forced choice, often a fear of the worse option, rather than on reasons.

Examples

  1. Either we raise the bake sale prices this year or (marked: False binary) the class trip gets cancelled.

    Honest version: The class trip is short on funds this year. Raising bake sale prices is one way to close the gap; a second sale day or a parent sponsor could also help.

  2. When your old phone slows down, there are only two options (marked: False binary): buy the new model or live with the lag.

    Honest version: Your old phone has slowed down. Buying the new model is one fix; clearing storage or replacing the battery may also help.

A fair use

The 12 bus runs on either the weekday schedule or (marked: False binary) the Sunday schedule; there is no Saturday service.

The timetable really does have just two schedules, so naming both is an accurate description, not a forced choice. Nothing is left out and nobody is being pushed toward an option.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

The rule looks for set phrases such as “either ... or”, “only two options” or “no middle ground”. A forced choice written without them gets through, for example “Raise the bake sale prices this year, or the class trip is off.” Small changes also slip past: one extra word (“There are only two real choices here.”), or more than 32 words between “either” and “or”, as in “Either we raise the bake sale prices this year for every single cookie, brownie, muffin, cupcake, and lemon bar on the long table by the door of the gym, and we ask every family to bake twice as much, or the class trip gets cancelled.”

Tier II · Psychological

FEAR_URGENCY

Fear + urgency

It pushes the reader to act fast by pointing to disaster or a closing deadline, not by showing what is actually at stake.

In the rule pack: Fear-Based Urgency

How it works

Fear and time pressure narrow attention. When a reader feels something terrible will happen, or that the chance is about to vanish, stopping to check the claim can feel like a risk in itself. The words carry the weight, so the reader never learns how likely the harm is, how big it is, or what waiting would really cost.

The tier, in plain words

This is a Tier II (Psychological) move because it works on the reader's fear and sense of alarm instead of giving reasons they can weigh.

Examples

  1. Act now (marked: Fear + urgency), or you'll be stuck with a slow, outdated phone for years.

    Honest version: Your phone is three years old and runs the newest apps slowly; if that bothers you, the upgrade offer runs through Friday.

  2. Leave that small roof leak alone and the damage to your attic will be irreversible (marked: Fear + urgency).

    Honest version: The roofer found one loose shingle over the attic; fixing it costs about $150, and a slow leak left all season can rot the boards underneath.

A fair use

The regional weather service forecasts the river will crest at 18 feet on Tuesday and warns that this could bring devastating (marked: Fear + urgency) flooding to low-lying streets (Flood Bulletin No. 2026-09).

The strong word describes a specific forecast. The sentence names the source, the expected water level, the day, and the streets at risk, so the reader can check the warning and judge it for themselves. Nothing in it tells the reader to act without evidence.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

1) Curly apostrophes. The rule's phrases are written with a straight apostrophe, so “If we don’t act this week, the school bake sale won’t have enough treats to sell” (typographic ’) gets through. The same sentence with a straight ' is flagged. Word processors and phones often insert curly quotes automatically. 2) Pressure that avoids the listed words. Scarcity and countdown phrasing such as “Only three upgrade slots left, and this price disappears at midnight. Don't wait.” raises no flag. Neither does vivid imagery like “Picture your family shivering in a dark, cold house next winter because you put off the furnace check.” The rule matches set phrases like “act now”, “too late”, “irreversible” and “complete collapse”, not the pressure itself.

Tier II · Psychological

SHAME_LEVER

Shame lever

It pushes you to agree by suggesting that only a foolish, careless, or bad person would disagree.

In the rule pack: Shame or Social Pressure Language

How it works

The sentence swaps a reason for a judgment about the reader. Disagreeing now looks like admitting you are unreasonable or letting people down, so going along feels safer than asking for proof. The claim itself never gets any support. What gets pressed is your wish to be seen as sensible and decent.

The tier, in plain words

This is a Tier II (Psychological) move because it works through feelings of shame, guilt, and belonging rather than through evidence or through appeals to authority.

Examples

  1. Any reasonable person (marked: Shame lever) would upgrade to the new phone this fall.

    Honest version: The new phone comes out this fall, and its battery lasts about six hours longer than last year's model, so upgrading may make sense if battery life matters to you.

  2. How could anyone possibly (marked: Shame lever) skip the school bake sale after all the work the parents put into planning it?

    Honest version: The parents put a lot of work into planning the school bake sale, which runs Saturday from 9 to noon, and the money raised will pay for new library shelves.

A fair use

How could anyone possibly (marked: Shame lever) carry the new couch up those stairs? We measured, and the stairwell is two inches too narrow.

This is a practical question about moving furniture, not a judgment of anyone's character, and the next sentence gives the measurement that answers it. No one is being shamed into agreeing.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

The rule only matches a short list of fixed phrases, so peer pressure written in ordinary words gets through. For example, “Every other family on the street has already signed up for the bake sale, so don't be the one house that lets the kids down” is not flagged. Swapping in a synonym also gets past it: “Any sensible person would upgrade to the new phone this fall” is not flagged, because “sensible” is not one of the words the rule looks for there.

Tier II · Psychological

EMOTIONAL_SUBSTITUTION

Feeling for evidence

It uses strong feeling words where the reasons or facts behind a point should be.

In the rule pack: Emotional Appeal in Place of Argument

How it works

A strong feeling arrives faster than a reason, and once it is there it feels like a reason. Words like “heartbreaking” or “outrageous” tell the reader how to feel about a fact before the fact is shown, and “speaks for itself” skips the showing entirely. The reader comes away with a reaction but not the details they would need to check it.

The tier, in plain words

This is a Tier II (Psychological) move because it works by pressing on the reader's feelings, such as sorrow, anger or alarm, rather than by setting what counts as common sense or by leaning on an authority.

Examples

  1. The new bus schedule is simply heartbreaking (marked: Feeling for evidence) for everyone who rides Route 4.

    Honest version: The new bus schedule drops the last two evening runs on Route 4, so riders who finish work after 9 p.m. will need another way home.

  2. The garden club wants members to chip in for new raised beds after the storm. The devastation speaks for itself (marked: Feeling for evidence).

    Honest version: The garden club wants members to chip in for new raised beds after the storm, which split 12 of the 20 beds.

A fair use

It was heartbreaking when (marked: Feeling for evidence) the late frost hit the garden club's seedlings: 40 of the 60 trays were lost overnight.

The feeling comes with the fact that explains it. The reader gets the count (40 of 60 trays, lost overnight) and can judge the loss for themselves. The emotion sits next to the evidence and does not replace it.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

1) The rule only matches set phrases, so a small change in wording gets past it. “It is heartbreaking to see the Route 4 evening runs disappear from the new bus schedule” is not flagged, because “heartbreaking” is caught only when followed by “that,” “how” or “when,” or when it comes after “simply,” “truly,” “absolutely” or “utterly.” Feeling words that are not on the list also pass: “The new bus schedule is a gut-wrenching blow to everyone who rides Route 4” is not flagged. 2) A vivid scene can do the same job without using any feeling word at all. “Picture a nurse standing alone at a dark bus stop at 10 p.m., waiting for a Route 4 bus that no longer comes” puts one imagined rider where the rider counts should be, and the engine does not flag it.

Tier II · Psychological

MORAL_HIGH_GROUND

Moral high ground

It turns a practical disagreement into a test of character, so that anyone who disagrees looks like a bad person instead of just someone with a different view.

In the rule pack: Moral Authority Claim

How it works

The sentence does not give a reason for its view. It says that good people already hold it. Most readers want to see themselves as decent, so agreeing starts to feel like the price of staying decent. The question about the facts never gets weighed, because objecting now feels like confessing a flaw.

The tier, in plain words

It is a Tier II (Psychological) move because it works on the reader's wish to be seen as a good person, and on the shame of seeming otherwise, rather than on any evidence about the question.

Examples

  1. Any caring person would agree (marked: Moral high ground) that the bake sale belongs on Saturday, not Friday.

    Honest version: I think the bake sale belongs on Saturday, not Friday: last year's Saturday sale drew twice as many families, and more parent volunteers are free that day.

  2. Keeping the late bus on the schedule is not a budget question; it is simply a matter of basic decency (marked: Moral high ground).

    Honest version: Keeping the late bus on the schedule costs about $300 a week, and I think it is worth it: about 40 riders use it each night, and most have no other ride home after late shifts.

A fair use

Any responsible professional should know (marked: Moral high ground) to shut off the water before replacing a faucet, as the plumbing guide explains on p. 14.

This sentence states a basic safety step that anyone can check, and it names where to look it up. Nobody on the other side is being cast as a bad person, and no open question is being closed off. 'Responsible' here means competent at the job. It is not a moral verdict on people who disagree.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

1) Saying it without the stock phrase: “Only someone who doesn't care about their neighbors would vote to cut the late bus” makes the same move, but the rule does not flag it because none of its fixed wordings appear. Neither is “Keeping the late bus is the only decent thing to do.” flagged. 2) Typography: “It’s simply a matter of basic decency” with a curly apostrophe gets through, while the same words with a straight one, “It's simply a matter of basic decency”, are flagged. Phones and word processors often put in the curly one on their own.

Tier II · Psychological

COMPETENCE_DISMISSAL

Competence dismissed

It answers a disagreement by saying the people who disagree don't understand the subject, and never gets to what they actually said.

In the rule pack: Competence-Based Dismissal

How it works

The reader hears that the critics are out of their depth and stops weighing what they said. No evidence is offered about the question itself. The only claim is about the critics, and it is usually too vague to check.

The tier, in plain words

This is Tier II (Psychological) because it works through shame and identity: readers back away from a view once they're told that holding it means they don't understand.

Examples

  1. Those who object to the new bus schedule fail to grasp (marked: Competence dismissed) how the routes actually work.

    Honest version: Some riders object to the new bus schedule because the early run now leaves at 7:40. The transit office says the change cuts the average wait from 14 minutes to 9.

  2. Customers who question the phone upgrade simply don't understand the technical requirements (marked: Competence dismissed) of the new chip.

    Honest version: Some customers question why the phone upgrade skips older models. The new chip needs 8 GB of memory, and the older models have 4 GB.

A fair use

Most first-time gardeners lack the experience to (marked: Competence dismissed) prune fruit trees, so the garden club runs a free class each spring.

Nobody's view is being dismissed and no argument goes unanswered. The sentence says beginners haven't learned a skill yet and points to a class where they can learn it.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

The rule matches exact verb forms, so the third-person singular slips through: “Anyone who objects to the bake sale prices misunderstands what the ingredients cost.” It isn't flagged because “objects” and “misunderstands” don't match the rule's “object” and “misunderstand”. The rule also misses any attack on the critic's competence that avoids its fixed phrases, as in “Critics of the new bus schedule have clearly never ridden a bus at rush hour.” That sentence makes the same move with none of the understand, grasp or expertise wording.

Tier II · Psychological

DISMISSAL_BY_REFRAMING

Dismissal by reframing

It restates someone's point as something they did not say, then answers that new version instead of the real one.

In the rule pack: Dismissal by Reframing

How it works

The restatement is presented as a helpful clarification of what the other side really means, so readers tend to accept it as a fair summary. Once the new version stands in for the original, rejecting it feels like rejecting the original, but the original point is never answered. No evidence is offered that the new version is what the person meant.

The tier, in plain words

It is a Tier II (Psychological) move because it works on how readers feel about the people making the point, often by giving them a hidden or unflattering motive, rather than on facts about the point itself.

Examples

  1. Riders asked for one more evening bus on Route 4, but what they're really asking (marked: Dismissal by reframing) for is a private taxi that everyone else pays for.

    Honest version: Riders asked for one more evening bus on Route 4. The transit office says it would add about 40 minutes of driver time each weeknight, so the board will weigh that cost against expected ridership.

  2. Parents asked that each item at the bake sale carry an ingredient label, but that is really just (marked: Dismissal by reframing) a way to push out the home bakers.

    Honest version: Parents asked that each item at the bake sale carry an ingredient label so families with allergies can choose safely. Home bakers would need to write out a label for every batch they bring.

A fair use

When the transit office asked riders what they're really asking (marked: Dismissal by reframing) for, 212 of 300 survey replies named a later last bus.

Nobody's point is recast here. The riders were asked directly, the answer is in their own words, and the sentence gives the count of replies. The engine matches the phrase without checking who is doing the restating.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

The rule only looks for a few set phrases, so the same move gets through when it is worded differently. “In other words, they want the city to run empty buses all night” is not flagged. Neither is “So what you're really asking for is a private taxi,” because the phrase check covers only “they,” “he” and “she,” not “you.” A second gap is typography. Published text often uses a curly apostrophe, and the phrase check expects a straight one: “What they’re really saying is that the buses should run empty.” is not flagged, though the same sentence typed with a straight apostrophe is.

Tier II · Psychological

CAUSAL_TOTALIZATION

Single cause

It blames one person, group, or thing for wrecking a whole life, everything, or a whole system, with no evidence, no mechanism, and no limit on the damage.

In the rule pack: Causal Totalization

How it works

Big harm words like “ruined” and “destroying” land as a feeling before they land as a claim. Pinning all of it on one actor gives the reader a simple story and someone to be angry at. It never says what that actor did, how it caused the harm, or how far the harm really goes, so there is nothing to check.

The tier, in plain words

This is a Tier II (Psychological) move because it works through frustration and a sense of loss, pointing those feelings at a single target instead of giving the reader facts to weigh.

Examples

  1. The city's new bus schedule is ruining my life (marked: Single cause).

    Honest version: Since the new bus schedule moved my morning bus from 8:10 to 8:25, I have been late to work three times this month.

  2. The garden club's new president ruined everything (marked: Single cause) this spring.

    Honest version: This spring the garden club's new president moved meetings to weekday mornings, and attendance fell from 30 members to 12.

A fair use

A burst pipe ruined everything (marked: Single cause) in the basement, including the washer, the dryer, and six boxes of books.

This is a literal account of physical damage. It names the cause (a burst pipe), limits the damage to one place (the basement), and lists what was lost. Nothing is exaggerated and nothing is hidden. The rule matches only the words “ruined everything” and does not look at the scope that follows, and it has no citation or detail check that could turn it off.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

1) Contractions get past it. “The new bus schedule's ruining my life.” makes the same claim as “The new bus schedule is ruining my life.”, which is flagged, but “schedule's ruining” is not matched. 2) Blame without a destruction verb gets past it. “The new coach is the reason our whole season fell apart.” puts the whole collapse on one actor, but it says “is the reason” instead of a listed verb such as ruin, destroy or wreck, so nothing fires. Naming one cause of all or every problem counts without one: “The new schedule is the cause of all our problems.” is flagged. Targets outside the rule's list also get through: “The new bus schedule is ruining the neighborhood.” and “The new president ruined the garden club.” raise no flags. The rule does count anything after “our”, so “The new president ruined our garden club.” is flagged, and any plural noun that ends the clause, for any group: “The new schedule is ruining the riders.” is flagged. So is a literal loss of plural things: “The frost ruined the tomatoes.” is flagged too.

Tier II · Psychological

MONOCAUSAL_BLAME

One culprit

It blames a broad outcome on one cause or one actor and leaves out everything else that played a part.

In the rule pack: Monocausal Blame

How it works

Most outcomes have several causes, and sorting them out takes work. A single cause gives the reader a tidy story and one clear place to aim their frustration, so it feels like an explanation even when no evidence ties that cause to the whole result. Once readers accept it, they stop asking what else contributed.

The tier, in plain words

It is Tier II (Psychological) because it works by giving frustration a single target, and that simple, satisfying story takes the place of evidence about what actually happened.

Examples

  1. Because of the new coach everything has gone downhill (marked: One culprit) this season.

    Honest version: The team has done worse since the new coach arrived, and the coaching change may be one of several reasons.

  2. The rain was the sole reason why (marked: One culprit) the school bake sale came up short this year.

    Honest version: The school bake sale raised less this year, and rain on the day was one likely reason.

A fair use

The plumber's inspection report found that a cracked seal was the main cause for (marked: One culprit) the leak under the sink.

The claim is narrow (one leak), it names its source (an inspection report), and it points to a physical cause that anyone can check. The engine matches the phrase 'the main cause for' without looking at scope or source, so it flags this fair sentence on the span 'the main cause for'.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

The rule matches only fixed phrases. (1) The same claim with different wording gets through. “Ever since the new coach arrived, nothing has gone right for this team.” makes the same one-cause claim but uses none of the trigger phrases, and the engine does not flag it. Neither does it flag “Thanks to the rain, everything has gone wrong.” or “The coach alone is to blame.” (2) Small changes to a phrase it knows can break it. “The coach is the only reason that we lost.” is flagged, but “The coach is the only reason we lost.” is not, because the rule wants “for”, “why” or “that” after “the only reason”. The phrases it knows are read the same way whatever the cause: “Because of the rain everything has gone wrong.” and “Because of the new bus schedule everything has gone wrong.” are both flagged.

Legal

These rules run only when a scan is set to the legal domain, or to all domains. The checker on this site runs the general rules.

Tier II · Psychological

LEGAL_SANCTIONS_THREAT

It warns that the other side will be punished if it keeps making an argument, so the argument gets dropped instead of answered.

In the rule pack: Sanctions Raised · Legal domain

How it works

A penalty warning makes an argument feel dangerous to keep, whether or not it is weak. The reader or the other side feels the risk and starts to doubt the claim, but nobody has shown why it is wrong. The threat stands in for the reasons that would actually settle the point.

The tier, in plain words

It is Tier II (Psychological) because it works through fear of being punished, not through evidence about the argument itself.

Examples

  1. The roofer's claim that the second coat was never paid for is baseless, and sanctions will be sought (marked: Sanctions raised) unless it is withdrawn by Friday.

    Honest version: The roofer's claim that the second coat was never paid for is wrong: the bank record from June 3 shows the full payment, and we will file it with our answer.

  2. If the garden club keeps pressing its fence-line argument, we will be seeking sanctions against (marked: Sanctions raised) it before the spring hearing.

    Honest version: We disagree with the garden club's fence-line argument because the March survey puts the post two feet inside our lot, and we will explain that at the spring hearing.

A fair use

The judge's order in the bus-shelter repair case explained when sanctions are warranted (marked: Sanctions raised) for a filing, then declined to impose any.

It reports what the court said about sanctions and what it actually decided. It does not threaten anyone or push anyone to drop an argument.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

The rule looks for particular words, such as “sanctions will be sought”, “warrants sanctions”, or a referral to a disciplinary board. It names no court rule or statute, so a threat made by a rule's number alone gets past it: “If the garden club keeps pressing its fence-line argument, we will move under Rule 11 before the spring hearing.” is not flagged. The same threat also gets past it in two other ways. First, it can be hinted at without those words: “You may want to think hard about what judges do to people who waste their time before you file that bake-sale refund claim again.” The engine did not flag this. Second, the penalty can be named in plain terms instead of legal ones: “Keep pushing the phone-upgrade fee argument and we will ask the court to make you pay every dollar of our legal fees.” The engine did not flag this either. The rule also cannot judge whether the challenged argument is really frivolous. It flags the words, not the purpose behind them.

Tier II · Psychological

LEGAL_STRAW_MAN

It restates the other side's argument in a bigger or cruder form than they gave it, then knocks down that weaker version.

In the rule pack: Opposing Position Restated in Absolute Terms · Legal domain

How it works

The reader usually hears the other side's position only through the writer's summary, so a stretched summary takes the place of the real one. Words like “every,” “never,” or “essentially arguing” make the stretched version sound extreme, and an extreme claim looks foolish without any proof against it. The reader ends up agreeing that the claim is wrong, but nobody has shown that the claim the other side actually made is wrong.

The tier, in plain words

This is a Tier II (Psychological) move because it works on how the reader feels about the other side, making them look unreasonable or extreme, and gives no evidence about what they really said.

Examples

  1. Plaintiff claims that every (marked: Position made absolute) shingle on the roof was installed wrong, which is hard to square with a roof that has not leaked in two years.

    Honest version: Plaintiff claims that three shingles on the north slope were installed loose; our inspector's report found those three shingles fastened to code, and the roof has not leaked in two years.

  2. The garden club is essentially arguing that we should (marked: Position made absolute) never be allowed to use the shared water line again.

    Honest version: The garden club asks that the shared water line be used only on weekday mornings; we ask for evening hours too, because our beds dry out by noon.

A fair use

Plaintiff argues that no (marked: Position made absolute) refund was issued for the returned phone, and attaches the store receipt showing the return date.

The summary is accurate. The plaintiff's claim really is that no refund was issued, and the sentence points to the receipt as evidence. Here “no” is simply what the claim says, not an overstatement. The rule can't tell the two apart, because it matches the words and never checks them against what the party actually said.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

1) The rule looks for two fixed wordings. The first is a formal party label (plaintiff, defendant, petitioner, respondent, appellant, appellee, opposing or other party, side, or counsel), then a verb like “argues” or “claims,” then an absolute word such as “every,” “no,” or “never.” The second is a reducing frame such as “essentially arguing” or “really asking,” followed by “we,” “this court,” or “the plaintiff” and then “should,” “must,” or “have to.” A loose paraphrase that uses neither wording gets through even when it is just as distorted. For example, “In other words, the homeowner wants the contractor to rebuild the entire porch for free” and “The bake sale committee wants every family to bake four dozen cookies, which no working parent could manage” are not flagged. “Homeowner” and “committee” are not on the party list, “wants” is not on either verb list, and neither sentence uses the reducing frame. 2) The rule is loaded only when the engine runs in the legal domain (or in all domains). The same sentence, “Plaintiff claims that every shingle on the roof was installed wrong,” is not flagged in the default general domain.

News and media

These rules run only when a scan is set to the media domain, or to all domains. The checker on this site runs the general rules.

Tier II · Psychological

MEDIA_WEASEL_QUANTIFIERS

Vague crowd

It says an unnamed and uncounted group agrees, so a claim seems widely held when nobody can check how many people actually hold it.

In the rule pack: Vague Quantifiers · News and media domain

How it works

Phrases like “many people say” or “it is widely believed” suggest that a crowd already agrees. The reader feels the weight of that crowd but never finds out who is in it, how many there are, or what they based their view on. Since no number or name is given, the claim can't be checked, and it can't be shown to be wrong either.

The tier, in plain words

This is a Tier II (Psychological) move because it plays on the wish to agree with others, so a reader goes along with the claim because “many” seem to believe it, not because of any evidence.

Examples

  1. Many people (marked: Vague crowd) say the garden club's spring plant sale was the best in years, and some experts (marked: Vague crowd) think sales will double next year.

    Honest version: The garden club's treasurer said spring plant sale receipts were the highest in five years, and she expects sales to double next year based on early pre-orders.

  2. It is widely believed (marked: Vague crowd) that the new city bus schedule saves riders time, and there is growing support (marked: Vague crowd) for adding a Sunday route.

    Honest version: The transit office's ride logs show trips on the new city bus schedule average six minutes shorter, and 140 riders have signed a petition asking for a Sunday route.

A fair use

At the garden club meeting, 18 of 24 members voted to hold the plant sale on Saturday; some people (marked: Vague crowd) wanted Sunday instead, and several people (marked: Vague crowd) abstained.

The sentence gives the real count first (18 of 24 in a named vote). “Some” and “several” only split up the six members who were left over. They don't suggest a larger agreement than the tally shows. The engine still flags it, because the rule counts two quantifier phrases and does not look at the numbers around them. A citation doesn't stop it either: this rule has no citation check, and a similar bake-sale sentence with “(Table 2)” was also flagged.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

1) The rule only fires after two vague phrases. One phrase on its own gets through: “Many experts think the new phone upgrade is worth the price.” is not flagged. 2) Any wording outside the rule's fixed word lists also gets through. Contractions do too: the rule looks for “it is” and “it has been”, so “it's” is missed. “It's widely believed the new phone upgrade fixes the battery problem, and it's generally thought the camera is better too.” is not flagged, while the same sentence with “it is” in both places is. This rule also flags nothing in “Most fans think this is the team's best season in years, and a lot of folks expect a title.” (“a lot of” and “folks” are not in its lists), though another rule, Consensus as proof, marks “Most fans think”.

Tier II · Psychological

MEDIA_EMOTIONAL_LEAD

Emotional lead

It opens a story with a strong feeling word, so the reader reacts before learning what happened.

In the rule pack: Emotional Lead · News and media domain

How it works

The first words of a story set the mood for everything after them. A word like 'shocking' or 'heartbreaking' tells the reader how to feel before they know the facts, so the facts that follow get read through that feeling. The word carries no evidence of its own. It only reports the writer's reaction, or the reaction the writer wants you to have.

The tier, in plain words

It is a Tier II (Psychological) move because it works on the reader's feelings, setting a mood of alarm, anger or sorrow before any fact arrives.

Examples

  1. Shocking (marked: Emotional lead) changes are coming to the Route 9 bus: starting Monday, buses will run every 20 minutes instead of every 15.

    Honest version: Starting Monday, Route 9 buses will run every 20 minutes instead of every 15, according to the schedule the city transit office posted this week.

  2. Heartbreaking (marked: Emotional lead) news from the school bake sale: this year's sale raised $450, down from $600 last year.

    Honest version: This year's school bake sale raised $450, down from $600 last year, according to the parent group's treasurer.

A fair use

Damaging (marked: Emotional lead) winds of up to 60 mph are possible Tuesday evening, the county weather office said in its storm advisory.

'Damaging winds' is a standard forecast term for wind strong enough to break branches or bring down lines. The sentence gives the speed and names the source, so the word reports a measured risk. It is not there to set a mood. The engine flags it anyway because it matches the word itself and does not look at how the word is used.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

The rule checks a fixed list of words in the first sentence of the text. (1) A charged word that is not on the list gets through. “Nightmare commute ahead: starting Monday, Route 9 buses will run every 20 minutes instead of every 15.” is not flagged. (2) A listed word after the first sentence also gets through: in “Starting Monday, Route 9 buses will run every 20 minutes. Riders called the change shocking.” the word “shocking” sits in the second sentence, so it is not flagged. The same goes for a scene-setting opening with no charged word at all, such as “The rain had soaked through her coat by the time the Route 9 bus finally came, twenty minutes late.” It still leads with feeling, and the rule does not flag it.

Tier II · Psychological

MEDIA_BURIED_QUALIFIER

Late qualifier

It opens with the story people will remember and puts the fact that weakens it, such as a denial, a correction or “no evidence,” at the end.

In the rule pack: Late Qualifier · News and media domain

How it works

Readers form their view of a story from its opening lines, and many stop reading before the end. When the correction or denial comes last, it is technically there, but it arrives after the reader has already made up their mind. Nothing false has to be said. The order of the facts does the work that evidence should do.

The tier, in plain words

It is a Tier II (Psychological) move because it works on how people hold first impressions: the feeling the opening creates stays with the reader even after the correction arrives.

Examples

  1. Last night's storm left fallen branches on nearly every street in town, flooded a dozen basements and sent a long line of neighbors to the hardware store, and by breakfast most of them were blaming the old storm drains on Elm Street. However, the town engineer found no evidence (marked: Late qualifier) that the drains failed.

    Honest version: The town engineer found no evidence that the old storm drains on Elm Street failed in last night's storm, though by breakfast most neighbors were blaming them after the storm left fallen branches on nearly every street, flooded a dozen basements and sent a long line of neighbors to the hardware store.

  2. Owners who installed this week's phone update flooded the help forums with posts saying their batteries now die before lunch, and the most shared thread, titled "This update ruined my phone," drew thousands of angry replies within a day. In the phone maker's lab tests, though, the drain was not confirmed (marked: Late qualifier) on any model.

    Honest version: The phone maker's lab tests did not confirm a battery drain on any model, though owners who installed this week's update flooded the help forums with posts saying their batteries now die before lunch, and the most shared thread, titled "This update ruined my phone," drew thousands of angry replies within a day.

A fair use

Starting Monday, Route 4 buses will run every 15 minutes instead of every 20, Route 9 will add a stop at the public library, and the last evening bus on both routes will leave the depot at 11:10 p.m. instead of 10:40 p.m. However, the start date for the new Sunday service is not confirmed (marked: Late qualifier) yet.

The main news is stated plainly and is not changed by the last sentence. The caveat is about a different detail, the Sunday service, and it is said clearly. Nothing earlier in the text sets up an impression that the caveat has to undo, so the order is simply the natural order of importance.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

1) Wording outside its short list. The rule only fires when a “however / but / though” is followed within about 150 characters by one of a few fixed phrases (“no evidence,” “not confirmed,” “denied,” “could not be confirmed,” “remains unclear,” and so on). A buried denial worded as “has not been confirmed,” “could not confirm,” or “did not confirm” goes unflagged. So does one with no transition word, like “The town engineer's report, for its part, found the drains worked as designed.” All of these were tested and none was flagged. 2) It counts characters, not reading order. If the frame is set in under 200 characters, as in a headline or a short alert followed right away by the qualifier, the rule stays silent. It also stays silent when the qualifier comes more than about 150 characters after the “however,” even though that pushes the qualifier deeper into the text.

Tier II · Psychological

MEDIA_SPECULATIVE_FRAMING

Guess as likely

It uses forecast words like “is expected to” or “could well lead to” to make a guess about the future sound nearly settled, without saying who expects it or why.

In the rule pack: Speculation Presented as Likely · News and media domain

How it works

Words like “expected”, “poised”, “set to”, and “signs point to” suggest that someone has weighed the evidence, but they never say who did or what they found. Each phrase is hedged, so none of them can be proven wrong. Stack two or three together and the reader walks away treating the outcome as close to certain.

The tier, in plain words

This is a Tier II (Psychological) move because it works on what readers hope for or brace for, and that expectation replaces a claim they could check.

Examples

  1. The new city bus schedule is expected to (marked: Guess as likely) cut wait times, and it could well lead to (marked: Guess as likely) fuller buses by spring.

    Honest version: Transit planners estimate the new city bus schedule will cut average waits from 15 minutes to 10, based on a three-month trial. They have not projected how many more riders it will bring.

  2. The home team is poised to (marked: Guess as likely) win the title this season, and signs point to (marked: Guess as likely) a sold-out final.

    Honest version: The home team leads the league by six points with four games left, and 9,000 of the 10,000 seats for the final have been sold.

A fair use

The posted timetable says the new city bus schedule is set to (marked: Guess as likely) start on June 1, and the Route 4 bus is slated to (marked: Guess as likely) run every 12 minutes.

This is not a guess. It reports a published plan, names where the plan comes from, and gives exact dates and times that a reader can check. Here “set to” and “slated to” are ordinary scheduling words, not a forecast made to sound certain.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

1) A single phrase passes. The rule needs at least two matching phrases, so “The new phone upgrade is expected to sell out on the first day.” raises no flag, and neither does the same guess followed by a sentence without a listed phrase (“...is expected to sell out. Stores are bracing for long lines.”). 2) Wording outside the fixed list passes. “Could face delays” on its own (the rule only catches “could soon/eventually/ultimately/well face”), “will likely”, and contractions such as “It's expected to” or “the storm's likely to” do not match. So “The garden club's spring show could face delays, and it will likely draw a smaller crowd.” is not flagged. (Also note that the rule runs only when a scan is set to the media domain or to all domains. In a general scan, the first example is not flagged.)

Finance

These rules run only when a scan is set to the financial domain, or to all domains. The checker on this site runs the general rules.

Tier II · Psychological

FIN_ANCHORING

Chosen reference point

It measures a number against a hand-picked high or low, so the same figure looks like a bargain or a boom.

In the rule pack: Chosen Reference Point · Finance domain

How it works

The first number a reader sees becomes the yardstick for the numbers after it. A price “down 45% from the record high” feels cheap because the record high is the comparison, even if that high was unusual and today's price is ordinary. The reader comes away with a sense of value or momentum but never learns whether the reference point was a fair one to pick.

The tier, in plain words

It is Tier II (Psychological) because it works on the feeling of a bargain or a missed chance, not on evidence about the number itself.

Examples

  1. Season tickets are down 45% from the record high (marked: Chosen reference point), so this is the year to finally buy them.

    Honest version: Season tickets cost $330 this year, about the same as last year's $320; three seasons ago they peaked at $600.

  2. Last year's phone is now selling at just 40% of its peak price (marked: Chosen reference point), which makes it a steal for anyone ready to upgrade.

    Honest version: Last year's phone now costs $360, down from $900 at launch, while this year's model costs $950.

A fair use

The river is down 30% from its record high (marked: Chosen reference point) in April, which the county flood office says is normal for late summer.

The reference point is dated and relevant to the reader. A source supplies the context, and the sentence says what the drop means instead of using it to push a choice.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

1) Small changes in wording get past the regex. “Down 45% from their record high” is not flagged because the rule accepts only “its”, “the”, or no word before the anchor. “Less than half of what they did at the peak” is not flagged because there is no percent figure. 2) The most common retail anchor has no percent and no peak/high wording at all: “Was $900, now only $360” and “40% below the list price” both pass unflagged. The rule also runs only when a scan is set to the financial domain or to all domains. In the default general domain it flags nothing, even on the examples above.

Tier II · Psychological

FIN_PROJECTION_AS_FACT

Forecast as fact

It states a guess about the future, like a price or a sales total, as if it were already settled.

In the rule pack: Projection Stated as Fact · Finance domain

How it works

A forecast is an estimate, and it can be wrong. Writing “will reach” in place of “is expected to reach” drops the words that tell readers they are looking at an estimate. Readers take in a confident number as a known one and never ask what it rests on.

The tier, in plain words

This is Tier II (Psychological) because it works on how readers feel about uncertainty: sounding sure gives comfort or urgency where the evidence gives neither.

Examples

  1. The school bake sale opens at nine on Saturday, and sales will reach $600 (marked: Forecast as fact).

    Honest version: The school bake sale opens at nine on Saturday. Last year it took in $540, so the committee expects sales of about $600.

  2. The new phone goes on sale in March, and the price will fall (marked: Forecast as fact) by half before summer.

    Honest version: The new phone goes on sale in March. The maker's last two models dropped about 40% within four months, so the price may fall by half before summer.

A fair use

On July 1, under the fare change the transit board approved in May, bus ticket prices will reach $2.50 (marked: Forecast as fact).

This isn't a forecast. The new fare was already voted on and has a start date, so “will” describes a scheduled fact, and the sentence says where it comes from.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

The rule only matches a short list of money words followed by a short list of verbs. So “sales will climb to $600” passes, because “climb” is not on the verb list. “The garden club's plant sale will bring in $900” also passes, because the subject is not on the list. The rule also runs only when the text is checked in the financial domain (or all domains). In the default general domain, the same bake-sale sentence is not flagged at all.

Tier II · Psychological

FIN_RECENCY_EXTRAPOLATION

Streak extended

It takes a short run of good results and treats the run as if it will keep going, without saying why.

In the rule pack: Trend Extrapolation · Finance domain

How it works

A recent streak is easy to picture, so the reader carries it forward without being asked to. The writer never states or defends the prediction. They only point at the streak, or say nothing suggests it will end. Because the forecast is never spelled out, the reader is never shown the evidence it would need, such as how long the run has lasted, what caused it, or what might change.

The tier, in plain words

It is a Tier II (Psychological) move because it works on hope and the fear of missing out, not on evidence.

Examples

  1. The school bake sale has raised more than the year before for three springs in a row, and there's no reason to think this will stop (marked: Streak extended).

    Honest version: The school bake sale has raised more than the year before for three springs in a row. Three years is a short run, and turnout and donated goods change each spring, so we are planning for a total near last year's.

  2. After six straight wins, the home team continues its winning ways (marked: Streak extended), so fans are already planning the championship parade.

    Honest version: The home team has won six straight games. Half the season is left, and four of the remaining games are against the top teams in the league, so the title is still open.

A fair use

The Route 9 bus has consistently delivered (marked: Streak extended) on-time arrivals 95 percent of the time over five years of monthly logs, but the schedule changes in January, so that record may not hold.

It reports a measured past record, names the source and the time span, and says plainly that the record may not continue. It does not extrapolate.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

1) It only matches a few fixed phrases. A sentence like “The new phone sold out in its first week, and the trend shows no sign of slowing” makes the same move and is not flagged. The regex needs “no sign” to be followed directly by “this/it/the ... will stop/slow”, so “no sign of slowing” gets through. The same goes for “keeps winning”, “is on a roll”, or “has outperformed three years straight” when no word like “consistently” or “always” appears. 2) When the conclusion is left for the reader to draw, as in “The garden club doubled its membership last spring. Imagine how big it will be in five years.”, no trigger phrase appears and nothing is flagged. Also, the rule is financial-only: it runs when a scan is set to the financial domain or to all domains. In a general scan it never fires, even on the exact example text.

Back to the contents

Tier III

Institutional

Moves that borrow the voice or weight of an office, a credential or an organization, so that official standing takes the place of a reason.

General

These rules run on every scan, including the checker on this site.

Tier III · Institutional

CREDENTIAL_AS_PREMISE

Credential as proof

It offers a title, a school, an award or years of experience as the reason to believe a claim.

In the rule pack: Credential Given as the Premise

How it works

A credential tells a reader who is speaking, not whether the claim is right. Put where the reason should be, it asks the reader to trust the speaker instead of weighing the evidence. The same credential can come with the evidence, and then it is background, not proof.

The tier, in plain words

This is a Tier III (Institutional) move. It borrows the standing of a school, an award or a profession to carry a claim, in place of the evidence for it.

Examples

  1. A Wellsmere-educated nutritionist says (marked: Credential as proof) the new school lunch menu is healthier, so the board should approve it.

    Honest version: The new school lunch menu has a third less added sugar than the old one, by the nutrition office's count of the recipes, so the board should approve it.

  2. As a baker with thirty years of experience, I can tell you (marked: Credential as proof) the bake sale should only sell bread.

    Honest version: The bake sale should only sell bread: last year the bread sold out by noon, and half the cookies were left at closing.

A fair use

As a nurse, I know (marked: Credential as proof) the new visiting hours are a relief for patients' families, and the ward survey in May found 41 of 50 families preferred them.

The credential comes with the evidence: a survey with a month and a count that a reader can check. The job title says who is speaking; it does not replace the reason. The engine flags it anyway, because it matches “As a nurse, I know” and does not read the rest of the sentence.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

A credential in other words gets past it. “A well-known baker says sourdough keeps longer.” and “Our baker, who has worked here for thirty years, says sourdough keeps longer.” are not flagged, while “As a baker with thirty years of experience, I can tell you” is. The rule reads a school or an award only in a few shapes (a name before “-educated”, “-trained” or “-winning”, or before “winner” or “recipient”) and a status word only from a short list (“renowned”, “leading”, “respected” and a few more), so “A professional baker says sourdough keeps longer.” is not flagged either. It also cannot tell a credential given as background from one given as proof: a sentence that gives its evidence too is flagged the same way.

Tier III · Institutional

INSTITUTIONAL_POSITION_AS_SETTLED

Finding with no reasons

It says a body has concluded or confirmed something, and gives no reason, data or source.

In the rule pack: Institutional Finding Stated Without Its Evidence

How it works

“Has concluded that” sounds like the end of a careful process. The reader learns that a finding exists but not what it rests on, so the body's standing does the work the evidence would do. A conclusion whose grounds are left out cannot be checked.

The tier, in plain words

This is a Tier III (Institutional) move. It leans on a body's authority to settle a question, in place of the reasons the body had.

Examples

  1. The transit board has concluded that (marked: Finding with no reasons) the new timetable is working.

    Honest version: The transit board reported that 95 of 100 weekday trips were on time last month.

  2. The garden club has found that (marked: Finding with no reasons) the new compost works.

    Honest version: In the garden club's test beds this summer, tomatoes grown with the new compost gave about a fifth more fruit than those grown without it.

A fair use

The transit board has concluded that the new timetable is working, based on arrival logs that show 95 of 100 weekday trips on time last month.

The finding comes with what it rests on, in the same sentence: the arrival logs and a count a reader can check. The engine does not flag it, because it reads “based on” as a reason given.

Does the rule mark it? No. The rule leaves it unmarked.

What it misses

1) Other verbs and tenses get past it. “The club's committee concluded last week that the compost works.” (plain past tense) and “The garden club has decided to move the plant sale to Saturday.” are not flagged, while “The club's committee has now concluded that the compost works.” is. 2) Any reason word in the same sentence turns it off, whether or not a reason follows: “The transit board has concluded that the new timetable is working, based on nothing.” is not flagged. 3) It reads any subject the same way, a person or a plain noun as well as an organization, so it cannot tell a body's formal finding from a remark. Citation checks use the first case-insensitive occurrence of each matched phrase, within 120 characters on either side. The rule stays quiet only when every one of those windows contains a citation shape; otherwise citation suppression removes none of its matches. A cited earlier copy can hide an unsourced repetition much later. The Method page explains this inherited limit.

Tier III · Institutional

NEUTRALITY_CLAIM

Neutrality claimed

It calls its own view neutral, independent or objective instead of showing the evidence.

In the rule pack: Neutrality Claimed

How it works

A label of neutrality asks the reader to lower their guard: if the speaker has no side, there seems to be nothing to check. Calling a view objective does not make it so, and the label stands in for the evidence that would show it.

The tier, in plain words

This is a Tier III (Institutional) move. It claims the standing of an impartial referee for a view, in place of showing how the view was reached.

Examples

  1. As a neutral third party (marked: Neutrality claimed), the phone store finds the upgrade fair.

    Honest version: The phone store, which sells both plans, compared their monthly costs: the upgrade costs four dollars more and adds a second line.

  2. Our review of the kettle reflects the evidence objectively (marked: Neutrality claimed).

    Honest version: We tested six kettles for boil time; the new one boiled a liter in three minutes, the fastest of the six.

A fair use

As an independent contractor (marked: Neutrality claimed), I set my own hours.

Here “independent” names a kind of job, not a claim to have no side, and the sentence argues nothing. The engine flags it anyway, because it matches “as an independent” and a noun and cannot tell the job from the stance.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

Neutrality claimed in other words gets past it. “As a completely neutral observer, I think the kettle is the best.” (a word between “a” and “neutral”), “As a fair-minded observer, I think the kettle is the best.” and “Speaking as someone with no stake in it, the phone upgrade is fair.” are not flagged. The rule reads a short list of phrases (“neutral”, “independent” or “nonpartisan” right after “as a”, “objectively and without bias”, “reflects the evidence objectively” and a few more), so the same claim in other words goes unmarked.

Tier III · Institutional

BUREAUCRATIC_OBSCURITY

Fog language

It wraps a simple claim in stiff, official-sounding words, so the claim sounds more settled than it is, or so its real meaning is harder to see.

In the rule pack: Dense Bureaucratic Phrasing

How it works

Formal legal and office phrases sound like a rule or a procedure is speaking, not a person, so readers tend to accept the sentence rather than question it. The dense wording also slows readers down, and a plain point, often an unwelcome one like a cut, a delay or a cost, can slip past them. No evidence has been added; only the tone has changed.

The tier, in plain words

This is a Tier III (Institutional) move because it borrows the voice of an office or a legal document, so the claim leans on how official it sounds, not on reasons the reader can check.

Examples

  1. Pursuant to the implementation (marked: Fog language) of the revised service plan, Route 9 buses will arrive every 20 minutes instead of every 15.

    Honest version: Route 9 buses will now come every 20 minutes instead of every 15. That means a longer wait at each stop.

  2. Notwithstanding the foregoing (marked: Fog language), the contractor says the leak under the sink falls outside the scope of the warranty.

    Honest version: Even so, the contractor says the warranty does not cover the leak under the sink.

A fair use

It should be noted that (marked: Fog language) the garden club survey covered only 40 households (Table 2), so the result may not hold for the whole town.

Here the formal phrase points to a real limit in the evidence and says where the numbers are (Table 2). It makes the claim weaker and easier to check, not more official. The rule matches the words “It should be noted that” no matter what follows, and this rule is not switched off when a citation is present, so the engine flags the sentence anyway.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

1) The rule only catches a fixed list of phrases, so newer office jargon gets through. For example, “To right-size service levels, the transit team will leverage stakeholder alignment across Route 9” hides a service cut in the same way, and the engine does not flag it. 2) The buzzword check needs exactly one word between the buzzword and the noun. “The phone plan will effectuate a streamlined trade-in mechanism for all customers” has two words in between, so it is missed. Long noun phrases with no listed words are also missed, such as “A service frequency optimization initiative has been undertaken for Route 9.”

Tier III · Institutional

VAGUE_INSTITUTIONAL_APPEAL

Vague authority

It says important organizations back a claim but never says which ones.

In the rule pack: Vague Institutional Appeal

How it works

Words like “leading” or “key” make readers picture serious, trusted bodies standing behind the claim. No names are given, so the reader cannot check who said it, what they said, or whether anyone said it. The claim gets the weight of authority without anything a reader can test.

The tier, in plain words

This is a Tier III (Institutional) move. It borrows the standing of institutions to carry a claim, in place of the evidence those institutions would supply.

Examples

  1. Leading organizations recommend (marked: Vague authority) watering tomatoes at dawn, so the garden club is moving its watering hour.

    Honest version: The garden club is moving its watering hour to dawn. The county extension office's vegetable guide recommends morning watering so the leaves dry before evening.

  2. Industry leaders have long recognized (marked: Vague authority) that a roof should be replaced every twenty years, so we booked a crew for spring.

    Honest version: The label on our shingles lists a twenty-year lifespan, and the roof is now twenty-two years old, so we booked a crew for spring.

A fair use

Two key groups support (marked: Vague authority) the new bus schedule: the riders' council, which voted for it in June, and the school board, which asked for an earlier first bus.

The sentence names both groups and says what each one did, so a reader can check it. The engine still flags it. It matches “key groups support” and finds nothing nearby that it counts as a citation, because it does not read the names that come after the colon.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

1) Small changes in wording get past it. The rule knows a fixed list of adjectives, nouns and verbs, in one word order. None of these are flagged: “Well-known organizations recommend watering tomatoes at dawn” (“well-known” is not on the list), “Watering tomatoes at dawn is recommended by leading organizations” (passive order), and “Top experts back the new watering schedule” (“back” is not on the list; “Top experts recommend the new watering schedule.” is flagged). 2) A citation shape can silence the rule, whether or not it is a real citation. “Leading organizations recommend watering tomatoes at dawn (see Table 2), so the garden club is moving its watering hour.” is not flagged, because “(see Table 2)” looks like a reference. Citation checks use the first case-insensitive occurrence of each matched phrase, within 120 characters on either side. The rule stays quiet only when every one of those windows contains a citation shape; otherwise citation suppression removes none of its matches. A cited earlier copy can hide an unsourced repetition much later. The Method page explains this inherited limit.

Tier III · Institutional

TOTALIZING_HARM_LANGUAGE

Total harm

It describes harm as total and final, with everything ruined and nothing left, without saying what was actually damaged or how much.

In the rule pack: Totalizing Harm Language

How it works

Phrases like 'completely destroyed' or 'nothing left' give the reader a finished verdict instead of a measurement. A picture of total loss is vivid and hard to argue with, so readers tend to accept the scale without asking what was counted. Because no scope is given, there is nothing to check, and a partial loss can pass as a total one.

The tier, in plain words

BiasClear files it under Tier III, Institutional, because it speaks with the finality of an official damage report and settles the question of scope the way a signed assessment would, but with no assessment behind it.

Examples

  1. Last night's hailstorm completely destroyed (marked: Total harm) the vegetable garden.

    Honest version: Last night's hailstorm flattened the tomato and squash beds in the vegetable garden; the herb beds and the fruit trees lost leaves but no branches.

  2. The basement flooded over the weekend, and everything is ruined (marked: Total harm).

    Honest version: The basement flooded over the weekend with about two inches of water; the carpet and three boxes of books are soaked, but the furnace and the washer are dry and working.

A fair use

The backyard shed was completely destroyed (marked: Total harm) in Tuesday's fire, but the house ten feet away had only a scorched fence post.

The sentence names one object, the shed, says what happened to it, and marks where the damage stopped. That makes 'completely' a checkable fact about a bounded thing, not a sweeping verdict. The engine flags it anyway because the rule matches the words 'completely destroyed' and cannot see that the scope is limited.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

1) Plain past tense gets through. “Last night's hailstorm destroyed the whole vegetable garden.” and “Last night's hailstorm wiped out the whole vegetable garden.” are not flagged. The rule catches “the whole” or “the entire” only after -ing verbs like “destroying”. It catches a past-tense word like “destroyed” or “ruined” only right after an intensifier such as “completely”, or right after “everything” or “everything is”: “Last night's hailstorm completely destroyed the vegetable garden.” is flagged. 2) A reference mark can silence the rule. “Last night's hailstorm completely destroyed the vegetable garden [1].” and “Last night's hailstorm completely destroyed the vegetable garden (see p. 2).” are not flagged. Bracketed numbers, page numbers and similar reference marks can count as citation shapes. That mark says nothing about whether the whole garden was lost. Citation checks use the first case-insensitive occurrence of each matched phrase, within 120 characters on either side. The rule stays quiet only when every one of those windows contains a citation shape; otherwise citation suppression removes none of its matches. A cited earlier copy can hide an unsourced repetition much later. The Method page explains this inherited limit.

Legal

These rules run only when a scan is set to the legal domain, or to all domains. The checker on this site runs the general rules.

Tier III · Institutional

LEGAL_SETTLED_DISMISSAL

It calls a rule “settled” to end a discussion, without showing how the rule answers the point someone raised.

In the rule pack: Settled Law Dismissal · Legal domain

How it works

The word “settled” tells readers that people with authority answered the question long ago, so asking again feels pointless or rude. The sentence never names the rule, quotes it, or shows how it fits the case at hand. The reader gets a verdict and a closed door, not a reason they can check.

The tier, in plain words

This is a Tier III (Institutional) move because it uses the weight of past rulings and rulebooks, the authority of the institution, where an argument the reader could test should be.

Examples

  1. The garden club's plot rules are settled law (marked: Settled law), so the new member's question about the waiting list needs no answer.

    Honest version: Rule 3 of the garden club's bylaws gives open plots to members in the order they joined; the new member joined in May, so she is fourth on the waiting list.

  2. It is well settled that (marked: Settled law) a rained-out game counts as a tie, so the coach's request for a makeup game fails.

    Honest version: League rule 7.2 says a rained-out game counts as a tie and is not replayed, so the coach's request for a makeup game does not fit that rule.

A fair use

It is well settled that (marked: Settled law) rained-out games count as ties: league rule 7.2 says so on page 4, and the board applied it to three games last season.

The sentence names the rule, says where to find it, and points to past cases where it was applied, so a reader can check the claim. Here “well settled” sums up evidence that is given instead of standing in for it. The engine still flags it because this rule has no citation check and matches on the phrase alone.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

The rule only looks for a short list of stock phrases: well settled (with or without a hyphen), clearly or plainly established, settled, or erroneous, black-letter law, hornbook law, settled law, principle, precedent or authority, and “controlling authority is clear”. Other wording that makes the same move gets through. For example, “That question was answered years ago, so the new member's complaint about the waiting list needs no reply” is not flagged, and neither is “It is firmly established that a rained-out game counts as a tie, so the coach's request fails.” The rule also runs only in the legal domain (or all domains). In a general scan, “It is well settled that a rained-out game counts as a tie, so the coach's request for a makeup game fails” raises no flag at all.

Tier III · Institutional

LEGAL_MERIT_DISMISSAL

It calls an argument meritless or frivolous so that nobody has to answer what the argument actually says.

In the rule pack: Merit Dismissed Without Engagement · Legal domain

How it works

Words like “plainly meritless” or “wholly without merit” sound like a finished judgment, so the reader assumes someone has already checked the argument. But the sentence gives no fact, rule, or record showing why the argument fails. The strong label does the work that a reason should do.

The tier, in plain words

It is Tier III (Institutional) because it borrows the tone of a court's final ruling, and that official weight stands in for actually engaging with the argument.

Examples

  1. The homeowner's claim that the contractor skipped the second coat of sealant on the deck is plainly meritless (marked: Merit waved off), and the court should dismiss it.

    Honest version: The homeowner's claim that the contractor skipped the second coat of sealant on the deck should be dismissed: the job log and dated photos from both days show two coats going on.

  2. The garden club's challenge to the new plot fees is wholly without merit (marked: Merit waved off) and should be rejected.

    Honest version: The garden club's challenge to the new plot fees should be rejected because the club's own bylaws, in section 3, let the board set fees each spring without a member vote.

A fair use

The small-claims judge found the fence claim had no merit (marked: Merit waved off) because the survey map placed the fence two feet inside the owner's own lot.

It reports a judge's finding and gives the reason right away: a survey map anyone can check. The reader gets the evidence along with the verdict.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

The rule only knows a fixed list of phrases. Plain synonyms get past it: “baseless,” “devoid of merit,” “lacks merit” (without “any”), “meritless on its face,” “without any merit,” and “frivolous” on its own with no “clearly,” “obviously,” or “patently” in front all go unflagged. The rule also runs only when the text is checked in the legal domain (or all domains), so the same sentence checked as general writing is never flagged.

Tier III · Institutional

LEGAL_WEIGHT_STACKING

It treats a big, unnamed pile of authority as if its size settled the question, and never shows why the claim is right.

In the rule pack: Authority Weight Stacking · Legal domain

How it works

Hearing that “every court” or “the weight of authority” agrees makes a point feel closed before the reader asks what those sources actually said. The reader gets a sense of size, not a reason: no source is named, nothing is counted, and the reasoning that decided the matter is left out. Checking a whole stack of sources feels like too much work, so most readers just take the summary.

The tier, in plain words

This is Tier III, Institutional, because it leans on the combined standing of courts and other official sources and asks readers to trust that weight instead of judging the argument.

Examples

  1. The weight of authority (marked: Weight of authority) is against moving the bake sale outside, so it stays in the gym.

    Honest version: The school's event handbook says food sales must be held indoors, so the bake sale stays in the gym unless the committee changes that rule.

  2. Every court that has addressed (marked: Weight of authority) a leaky-gutter dispute sided with the homeowner, so the roofer must redo the job for free.

    Honest version: Two small-claims rulings on the same warranty wording sided with the homeowner because the warranty covers gutter leaks for two years; this leak began at eight months, so the roofer must redo the job for free.

A fair use

The weight of authority (marked: Weight of authority) on gutter warranties is evenly split, four small-claims rulings each way, so this case turns on the wording of the roofer's contract.

It reports what the authority actually shows, gives the count on each side, says plainly that there is no consensus, and points to the contract wording as the thing that decides the case. It is not using the pile of rulings to win the point.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

Small changes in wording get past it. “Every court to have considered a leaky-gutter dispute sided with the homeowner” is not flagged. The rule accepts “every court to consider” and “every court that has considered”, but not “to have considered”. A plain list also gets through: “Three past rulings, two consumer guides, a trade handbook and a repair pamphlet all side with the homeowner” stacks up authority without any of the stock phrases the rule watches for. (The rule also runs only when text is checked in the legal domain or all domains. The same flagged sentences checked in the general domain return no flags.)

Tier III · Institutional

LEGAL_PROCEDURAL_GATEKEEPING

It turns down a complaint on a procedural point, such as a missed deadline, the wrong person asking, or the wrong step, and never says whether the complaint is right.

In the rule pack: Procedural Bar Raised · Legal domain

How it works

A rule about timing or who may speak sounds fair and neutral, so readers accept the refusal without asking about the actual problem. The procedural point may be true, but it answers a different question. The reader comes away thinking the matter is settled when nobody has checked the facts.

The tier, in plain words

This is a Tier III (Institutional) move because it uses an institution's own rules and procedures to decide the question, so nobody has to weigh the evidence.

Examples

  1. The bake sale committee said the parent failed to raise (marked: Procedural bar) her concern about the prices before the sign-up deadline, so it will not look at whether the cookies cost too much.

    Honest version: The parent raised her concern about the prices after the sign-up deadline. The committee is not reviewing it for that reason, so it has not decided whether the cookies cost too much.

  2. Riders who missed the comment period have waived any objection (marked: Procedural bar) to the new bus schedule, so the transit board will not discuss the longer waits on Route 9.

    Honest version: Some riders sent their objections after the comment period closed. The transit board will not discuss them at this meeting, so nobody has yet checked the longer waits on Route 9.

A fair use

The league ruled that the coach failed to raise (marked: Procedural bar) his protest before the final whistle, as the rulebook requires, but it still reviewed the game video and found the goal was clean.

The sentence gives the procedural point and its source (the rulebook). Then it deals with the substance: the league checked the video and gives its finding. The late protest is not the only reason for the decision.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

1) Plain wording gets past the rule. The rule matches set phrases like “failed to raise”, “waived any objection”, “lacks standing to raise” and “procedurally barred”. “You brought this up too late, so we won't be looking at the prices at all” makes the same move and is not flagged by this rule. (Another rule, FEAR_URGENCY, catches “too late” in it.) Neither is “That point was not raised in time, and the board considers the matter closed.” 2) The rule only matches “before the court”, “before this court”, “before the tribunal” or “before this tribunal”. A sentence that names any other body is missed: “The complaint about the roof repair is not properly before the committee, so we will move on” is not flagged. Also note that the rule only runs when a scan is set to the legal domain (or all domains). The same sentences get no flag in the general domain.

News and media

These rules run only when a scan is set to the media domain, or to all domains. The checker on this site runs the general rules.

Tier III · Institutional

MEDIA_ANONYMOUS_ATTRIBUTION

Faceless source

It pins a claim on sources the reader cannot name or check, so the claim sounds reported without being shown.

In the rule pack: Anonymous or Unverifiable Source Attribution · News and media domain

How it works

Phrases like “insiders say” or “people familiar with the matter” suggest that someone close to the facts has spoken. The reader cannot see who they are, how they know, or whether anyone checked. The vague label stands in for the evidence and lends the claim the weight of inside access.

The tier, in plain words

This is Tier III (Institutional) because the claim borrows authority from an implied position inside an organization instead of from anything the reader can check.

Examples

  1. People familiar with the discussions (marked: Faceless source) say the city will drop the late-night bus route next month.

    Honest version: The transit board voted on Tuesday to drop the late-night bus route starting next month, according to the posted meeting minutes.

  2. Unnamed sources (marked: Faceless source) say the new phone upgrade will cost more than last year's model.

    Honest version: The maker's posted price list shows the new phone upgrade costs $50 more than last year's model.

A fair use

Two unnamed staffers (marked: Faceless source) at the bus depot, who are not allowed to speak publicly, showed the reporter the repair logs for the broken buses.

The sentence gives the reason the sources are not named, and it points to evidence the reporter saw with their own eyes (the repair logs). This is careful, legitimate anonymous sourcing, not vagueness standing in for proof.

Does the rule mark it? Yes. The rule marks it, so this is a false alarm.

What it misses

1) Wording outside the rule's fixed list gets through. Bare “Sources say the new phone upgrade will cost more than last year's model.” is not flagged. The rule counts “sources” only after a word such as “unnamed”, “anonymous” or “unidentified”, or before “close to”, “familiar with” or “with knowledge/insight”. “Word around the garden club is that the spring plant sale will be canceled.” is not flagged either. 2) A citation shape can silence the rule, whether or not it is a real citation. “Unnamed sources say the bus will now run every 15 minutes (see Table 2).” is not flagged, because “(see Table 2)” looks like a reference. Also note that the rule only runs when a scan is set to the media domain or to all domains. In the general domain, “Unnamed sources say the bus will now run every 15 minutes.” is not flagged. Citation checks use the first case-insensitive occurrence of each matched phrase, within 120 characters on either side. The rule stays quiet only when every one of those windows contains a citation shape; otherwise citation suppression removes none of its matches. A cited earlier copy can hide an unsourced repetition much later. The Method page explains this inherited limit.

Back to the contents