Complete the eccentric scientist's very strict (and slightly terrifying) lab safety manual by selecting the grammatically correct option.
Under no circumstances _____ the giant red button labeled "DO NOT PUSH."
The correct answer is should you press.
"Under no circumstances" is a restrictive negative phrase. When placed at the beginning of a sentence for dramatic emphasis, it forces inversion between the subject and the auxiliary/modal verb. Therefore, "you should" becomes "should you." We do not use a negative verb ("shouldn't") because "Under no circumstances" already makes the sentence negative!
Inversion
Question inversion vs emphatic inversion: question inversion is basic grammar (Is she ready?) — every learner uses it. Emphatic inversion (Never have I seen…, Not only does she…) is a C1+ rhetorical tool for formal writing and speeches. Same mechanism, different register.
Inversion swaps subject + auxiliary order. Triggered by: questions, fronted negatives (Never, Rarely, Not only), and conditional if-deletion (Had I known…).
Diagnostic: is a negative/restrictive word at the front of a declarative sentence? → inversion required. Is it a question? → inversion is automatic.
Negation
Single vs double negatives: standard English uses ONE negative per clause (I don't see anything or I see nothing). Double negatives (I don't see nothing) are grammatical in many languages and some English dialects, but are non-standard in written/formal English. This is the #1 negation trap for speakers of Spanish, Russian, and French.
Negation = not after auxiliary/modal, or do-support. Negative words (never, nobody, nothing) negate alone without adding not.
Diagnostic: count the negatives in the clause. More than one? → double negative. Fix by replacing one with a positive (anything, anyone, ever).
Modal verb
Must vs should vs might: the most confused modal trio. Must = strong obligation/near-certainty. Should = advice/expectation. Might = possibility. Getting these wrong changes the force of your statement: You must see a doctor (urgent) vs You should see a doctor (advice) vs You might need a doctor (maybe).
Modal verbs are auxiliaries that encode modality: ability (can), permission (may), necessity (must), advice (should), possibility (might), future (will).
Diagnostic: what meaning are you adding? Obligation → must/have to. Advice → should. Possibility → might/could. Ability → can. Future → will.
Word order
English (SVO) vs other patterns: English relies on word ORDER to show who does what (Dog bites man ≠ Man bites dog). Inflected languages (Latin, Russian, German) use case endings and can scramble order freely. In English, changing order changes meaning or requires special constructions (inversion, cleft sentences).
Word order = how English marks grammatical relationships. SVO is the default; fixed adjective order; adverb placement varies by type.
Diagnostic: does your sentence sound "off" even though all word forms are correct? → probably a word order issue. Try moving the element back to default SVO position.
B2 | Upper Intermediate
B2 vs C1: B2 means effective communication on complex topics with some effort. C1 means effortless fluency with precise register control. If you can argue a point but still reach for words and make structural slips under pressure, you're B2.
B2 is the upper-intermediate CEFR level: mixed conditionals, complex passives, reported speech with backshift, participle clauses, and sustained written argument.
Diagnostic: does your writing read as "competent non-native" or "could be native"? The former → B2. The latter → C1.
Hard
Hard vs Medium: Medium tests one rule with realistic distractors. Hard tests interacting rules, edge cases, or context-dependent answers where multiple options seem correct until you think deeply. If you're scoring 80%+ on Medium, try Hard to find your real gaps.
The Hard tag filters for B2+ challenges with layered difficulty: rule interactions, subtle distractors, and contexts that demand genuine grammatical reasoning.
Diagnostic: if Hard questions feel impossible, drop to Medium and master the individual rules first. Hard assumes you already know each rule — it tests whether you can apply them together.