The 3-Week Plan

Language courses teach six months of grammar. Your flight leaves in three weeks. That gap isn’t a shortcoming in the traveler — it’s a mismatch between what courses are built to do and what a trip actually requires. This is the plan we built JetPhrase around once we stopped trying to close that gap and started designing for it instead.

Three weeks is a deadline, not a slogan

Three weeks is roughly the window where a trip stops being an idea and starts being a flight number — booked, real, no longer abstract, but not yet the week where you’re packing. It’s also, conveniently, long enough to run three genuinely different phases of learning without asking for a semester nobody has: a week to build coverage, a week to build production, a week to sand down the edges specific to where you’re actually going.

Compress the trip to ten days and the plan compresses with it — fewer situations in week one, tighter spacing in week two — but the shape holds. What doesn’t hold is trying to skip a phase. Coverage without production gets you a traveler who recognizes “la cuenta, por favor” on a menu insert and freezes when a waiter says it back. Production without coverage gets you fluent at ordering coffee and silent at the one moment you actually needed the emergency phrase.

Week one: coverage, not vocabulary

The instinct when three weeks feels short is to cram more words in. The better move is to cram in fewer situations, covered completely. Nearly every trip, regardless of destination, runs through the same short list: arrival, ordering food, asking directions, small talk that smooths a transaction, and one phrase you hope you never need.

What makes this work across languages that otherwise look nothing alike is that the shape of “I’d like something, please” repeats almost everywhere, even when the words don’t:

Language	”I’d like a coffee, please”
Spanish	Quisiera un café, por favor
Italian	Vorrei un caffè, per favore
French	Je voudrais un café, s’il vous plaît
Turkish	Bir çay istiyorum, lütfen (tea, in this case)
Japanese	お茶をください (o-cha wo kudasai — tea, please)

Four unrelated language families, and every one softens a direct want into something closer to “I would like” before naming the thing. Week one isn’t about memorizing five hundred nouns — it’s about drilling that one sentence shape until it’s automatic, so that swapping café for agua, or un café for une bière, costs you nothing new to learn. Coverage of ten situations, done properly, transfers. Vocabulary size, on its own, doesn’t.

The same logic holds for the phrase nobody wants to need. Socorro in Spanish, İmdat in Turkish, 助けて (tasukete) in Japanese, Upomoć in Croatian — four completely different words, and every guide on this site puts one of them in the ten-phrase list anyway, not because it comes up often but because it’s the one situation where “I’ll look it up when I need it” isn’t a plan. Coverage means deciding in advance what you refuse to leave uncovered, not just what’s convenient to drill.

Week two: production, not recognition

Recognizing a phrase on a flashcard and producing it on demand, out loud, to a person who might answer unpredictably, are not the same skill — and most language apps only train the first one. Multiple-choice review builds recognition. A waiter who doesn’t slow down for you demands production.

You don’t learn a language by studying it. You learn it by needing it and surviving the exchange.

Week two is where JetPhrase swaps flashcards for full scenarios — a dialogue, both sides, with an AI partner who answers the way a real waiter, receptionist or stranger on a train platform might, not the way a script does. This is deliberately uncomfortable in a way week one isn’t, because production under mild social pressure is the actual skill a trip tests. Nobody hands you a multiple-choice card at a hotel front desk.

It’s also the week to fix the decisions that otherwise get re-litigated mid-sentence: which register you default to with strangers, whether you lead with the polite form or the short one, how you recover when you didn’t catch the reply. Making those calls once, in week two, is what prevents the hesitation in week three that reads as unsure — the accent was never the problem.

Week three: the edge of your specific trip

By the final week, the plan stops being generic and starts being about where you’re actually going. The regional dish that isn’t in any standard phrase list. The transit system’s specific jargon. The dialect quirk — a Castilian th sound if you’re in Madrid rather than Mexico City, a European rather than Brazilian “a aprender” if you’re in Lisbon — that a one-size guide has to gloss over but your trip doesn’t.

This is also where difficulty notes stop being generic. A phrase that’s trivial for a Portuguese speaker learning Spanish is a genuine snag for a Japanese speaker doing the same, and a plan that pretends otherwise is optimizing for nobody in particular. A German speaker learning Turkish and an English speaker doing the same don’t struggle with the same syllable, the same politeness marker, or the same word order — which is exactly why difficulty notes belong in the language guide for your own starting point, not copied wholesale from a version written for someone else’s.

Week three is small by design: not more material, just the last layer that turns a phrase list into a plan built for your specific dates and destination.

Why not just extend the deadline instead

The obvious objection is that three weeks is arbitrary and more time would simply be better. It would — for fluency. It wouldn’t be for a trip, because the marginal hour in week four buys you vocabulary you’ll use once, while the marginal hour in week two buys you the confidence to actually use the ten phrases you already have. A course optimizes for the language as a system to be understood in full. A traveler needs the language as a set of situations to be gotten through well. Those are different design problems, and most of what makes a grammar course long is solving the first one.

Three weeks is what’s left once you stop trying to teach a language and start trying to prepare a person for a specific, dated, unavoidable set of conversations. If you’ve already got the flight number, that’s not a hypothetical — it’s a countdown that started the moment you booked.