Proficient vs fluent: what language levels on a CV actually mean

The words are not a scale

"Fluent" and "proficient" are ordinary English words doing work they were never built for. Nothing defines where one stops and the other starts, no institution issues them, and the person choosing between them is the person being evaluated. A candidate who spent two childhood summers with relatives abroad and a candidate who has handled billing escalations in that language for four years may both write "fluent" on the same form. Neither is lying.

Most people place "proficient" somewhere below "fluent" and above "conversational", but exactly where is a matter of temperament rather than ability. Someone who has been told their accent is good rounds up. Someone who once froze in a meeting rounds down, permanently. The label ends up reporting a candidate's self-image and their sense of what a CV is for, mixed with some real ability, in proportions you cannot separate from the outside.

When a screener uses that word as a threshold — fluent advances, proficient does not — the funnel is sorting on vocabulary choice. It is a cut made at the least reliable stage of the process, and everything downstream inherits it.

Why self-reported levels are weak evidence for hiring

Self-report is not worthless in general; it is weak in this specific context, for three reasons that compound. There is no common reference point, so each candidate is silently comparing themselves with a different group — colleagues, classmates, native-speaker friends, the last person they were corrected by. There is no shared unit, so "good enough" has no anchor. And the reader is applying a threshold the writer never knew about.

The incentive runs one way. A CV is a persuasive document, and applicants know that a modest word costs them an interview. That does not make the confident answers false, but it does mean the distribution you are reading is compressed at the top: the people who could genuinely work in the language and the people who overestimated it are wearing the same label, while some capable candidates who described themselves carefully sit below your cut.

None of this is an argument for dropping the question. Ask it — it is useful context, and it tells you how the candidate thinks about their own ability. Just file it where the rest of the CV goes: a claim, to be checked, not a measurement.

CEFR: a shared vocabulary, and what it is for

The Common European Framework of Reference for Languages replaces the adjectives with six levels — A1 and A2 (basic user), B1 and B2 (independent user), C1 and C2 (proficient user) — and defines each one as a set of things a person can do rather than as a quantity of language they possess. B1 listening, in the Europass self-assessment grid, is "I can understand the main points of clear standard speech on familiar matters regularly encountered in work, school, leisure, etc." C1 spoken interaction is "I can express myself fluently and spontaneously without much obvious searching for expressions."

The more useful feature for hiring is that the grid is not one number. It describes five separate competences — listening, reading, spoken interaction, spoken production and writing — and a person is routinely at different levels across them. Someone who has worked years on the phone in a second language can be strong in spoken interaction and noticeably weaker in writing; someone who learned mostly from documentation is often the reverse. A requirement that says "B2" without saying B2 in what has thrown away the part that matters.

It is worth being precise about what CEFR is, though. It is a descriptive framework and a vocabulary, not a selection procedure. A candidate placing themselves at B2 on their own is still self-reporting — better-defined self-report, which is a real improvement, but not evidence. The framework earns its keep when it is used to state the requirement clearly, and something else is used to check it.

How institutions actually write a language requirement

Organisations that have had to make this decision at scale converged on the same shape: name the level, name the skills, and assess them. The US Interagency Language Roundtable scale runs 0 to 5 and defines its levels in terms of work — Level 2, Limited Working Proficiency, is "able to satisfy routine social demands and limited work requirements", while Level 3, General Professional Proficiency, is being able to "participate effectively in most formal and informal conversations on practical, social and professional topics". The unit of description is what the job needs done.

The UK's secure English language test guidance makes the second half of the point. The required CEFR level depends on the route being applied for, and so does the set of skills tested: routes such as Skilled Worker and Student require an assessment of all four skills — reading, writing, speaking and listening — while settlement and citizenship are assessed on speaking and listening only. The requirement is not a word; it is a level, a named skill set, and an actual test.

Neither of these is a hiring policy you should copy. They are worth noting because both refused the shortcut your CV screen is taking, and both refused it for the same reason: the level alone does not say which ability is being demanded.

Conversational fluency and a clear support email are different abilities

This is where the CV label does the most damage, because the interview reinforces it. A phone screen is spoken interaction with a cooperative partner who repeats themselves, slows down, and fills gaps. Someone warm and quick in that setting reads as fluent, and often is — in that competence. It says considerably less about whether their written refusal of a refund lands as firm and courteous or as blunt, because in writing there is no tone of voice to carry the intent and no chance to repair it in the next sentence.

The reverse case is quieter and costs you more people. Careful writers who hesitate on a call — processing the question, choosing a word — come across as weaker than they are, and support and back-office roles are full of them. If the job is 80% written and 20% spoken, an assessment made entirely in conversation has weighted the evidence upside down.

So the honest first question is not what level the role needs but what the role actually does in the language. How much of the work is written and how much spoken. Whether it is drafting or replying under time pressure. Whether the register is casual or has to survive being forwarded to a regulator. Whether there is technical vocabulary that a generally strong speaker would not have. That description is the requirement; the CEFR level is the shorthand for it.

• Written or spoken — and in what proportion across a normal week.
• Composed at leisure, or produced live with a customer waiting.
• The register the work has to hold: informal, formal, or legally consequential.
• Domain vocabulary the role assumes, which general proficiency does not supply.
• Comprehension load: accents, poor phone lines, long documents, or all three.

Turning the requirement into evidence

Once the requirement is written that way, checking it stops being mysterious. Reproduce the demand in a short task: here is the policy excerpt and the customer's message, write the reply in the target language, twelve minutes. Score it against a rubric written before anyone saw a response, with the criteria that actually matter for the job — accuracy against the policy, clarity, tone, and whether errors would impede a reader — rather than a general impression of language quality. A spoken competence gets its own short spoken task if the role needs one.

What you get back is not a certificate and should not be described as one. It is a work sample scored against expert-reviewed criteria, in a single sitting, on one slice of the job — with the limits that implies. It will not tell you how someone performs across a full week, and it is not a language qualification. It will tell you, in the candidate's own words on the page, whether the email you would have to send actually gets written.

One caution worth stating plainly. Language requirements can screen out candidates by national origin, and in several jurisdictions that puts the burden on the employer to show the requirement is job-related and consistent with business necessity — the framework set out in the US Uniform Guidelines on Employee Selection Procedures and the EEOC's guidance on tests and selection procedures, with equivalents elsewhere. Obligations vary by jurisdiction, this is not legal advice, and no product claim of compliance is being made here. The practical discipline is the same one that improves the hire: require the language the job uses, at the level the job needs, and be able to show why.

Sources & further reading

• European Commission, Europass — Common European Framework of Reference (CEFR) self-assessment grids
• Interagency Language Roundtable — Language Skill Level Descriptions: Speaking
• UK Home Office — Prove your English language abilities with a secure English language test (SELT)
• Uniform Guidelines on Employee Selection Procedures, 29 CFR Part 1607 — eCFR