Week 8: ICAO Level 4 as regulation

A rating scale for human language, with careers attached to it, and the plain-language fallback for when phraseology runs out

26 April 2027

Everything so far has been about the language. This week is about the people required to speak it, and the machinery that decides whether they may.

Why a scale exists at all#

Phraseology covers the routine. The moment a situation is not routine, the phrase book runs out and the speakers fall back on plain language — and plain language is where proficiency stops being a formality.

This is the argument that produced regulation. Charkhi Dadri, from week 6, is one of its sources: fifteen recommendations, several of them specifically on English language proficiency. Avianca 052, from week 7, is another — a crew who said true things in a form that did not carry their urgency.

So ICAO built a rating scale, six levels across six criteria: pronunciation, structure, vocabulary, fluency, comprehension, and interactions. Level 4 is operational, the minimum to work internationally. Below it you do not fly international. Level 4 requires periodic re-testing; Level 6 does not.

Note the shape of the thing. It is not a test of English. It is a test of whether you can handle the unexpected in English, which is a much stranger thing to measure.

What Level 4 actually says#

Read the descriptors rather than the number. Level 4 does not require an accent-free speaker, or a fluent one, or one who never makes an error. It requires, roughly, that pronunciation is intelligible to the aeronautical community, that errors do not interfere with meaning, that the speaker can paraphrase when a word is missing, and that comprehension holds up when the exchange takes an unexpected turn.

That last criterion is the one that does the work, and the one hardest to assess in a booked twenty-minute interview.

The uncomfortable part#

Priya assesses this professionally, and this lecture is co-taught so that the uncomfortable part gets said by the person who does it rather than about them.

A rating scale for human language has consequences for people’s careers, and those consequences do not fall evenly:

• The scale is applied to non-native speakers in practice. A native speaker with genuinely poor radio discipline is rarely the one being re-tested.
• Native speakers are a hazard the scale does not measure. Idiom, speed and local shorthand from a native speaker are a well-documented comprehension problem for everybody else on frequency — and there is no rating that catches it. Week 7’s catalogue was mostly native-speaker behaviour.
• Testing is uneven across states and providers, and the incentive to award Level 4 sits with people who are not neutral.
• Accent is not the criterion, and gets marked anyway. Which is why this course refuses to mark it, and says so on every assessment page.

Plain language, and doing it well#

The practical skill: when the phrase book runs out, say the thing plainly and slowly, and structure it. MAYDAY first if it is one. Nature of the problem, intentions, assistance required. Short sentences. One idea each.

BAW123   Mayday, mayday, mayday, Speedbird one two three, engine failure,
         one two thousand feet, turning left towards the coast,
         request immediate descent and vectors to nearest suitable.

Standard where the standard exists, plain where it does not, and no attempt to sound calm at the cost of being clear. Sounding calm is not a criterion either.

The thing to take away#

The scale is a real answer to a real problem, and it is also a blunt instrument pointed mostly at one group of people. Both. If your instinct is that it should be abolished, say what replaces it for the crew who cannot handle a non-routine exchange. If your instinct is that it should be stricter, say what catches the native speaker.