ILR Speaking Scale: Levels 0–5 Without False Equivalence

The ILR Speaking scale is a US government proficiency framework with six base levels, numbered 0 through 5, and plus levels from 0+ through 4+. Read it as an ordinal description of functional speaking across varied tasks, accuracy, content and context—not as an automatic conversion to CEFR, an exam band, a course level or an app score.

The Interagency Language Roundtable says its speaking descriptions cover functional ability in current spoken language. A stronger performance in one aspect does not, by itself, justify a higher overall level, and a plus level must substantially exceed one base level without fully meeting the next (official ILR Speaking descriptions). Those two rules prevent many tempting shortcuts.

What the numbers are—and are not

The official scale runs from no functional communicative ability at Level 0 to an exceptionally broad command at Level 5. Each higher base level generally assumes control of the relevant lower-level abilities. The revised document also separates four aspects: functional ability, precision of forms and meanings, meaningful content, and contextual appropriateness (May 2021 speaking revision).

This navigation table is an original planning aid. It does not replace the official descriptions or assign ratings:

Range	Useful planning question	Evidence to collect
0 / 0+	Can any memorised material achieve a basic purpose?	Short, predictable exchanges with the listener's interpretation
1 / 1+	Can the speaker manage common needs beyond one script?	Several everyday tasks with changed details
2 / 2+	Can the speaker sustain routine social and work-related exchanges?	Explanations, questions and complications across familiar topics
3 / 3+	Can the speaker handle broad professional responsibilities?	Extended discussion, support for views and adaptation to audience
4 / 4+	Can the speaker communicate extensively in demanding contexts?	Nuanced, abstract and unfamiliar tasks with varied interlocutors
5	Is performance broadly flexible at the highest described range?	Multiple high-demand samples assessed through an authorised process

Do not promote a learner because one response matches one row. A long monologue may hide weak interaction. Accurate grammar in a memorised speech may not show adaptation. Smooth delivery on a familiar topic may collapse when the audience, purpose or stakes change.

Treat plus levels as their own evidence problem

A plus sign is not a reward for being near the top of a class. The ILR preface defines plus performance as substantially beyond the lower base level without meeting the next base level. That means “2+” should not be inferred from a percentage, an average of unrelated sub-scores or a single feature that appears advanced.

For learning records, write what happened instead:

• completed a routine service call and resolved one unexpected change;
• explained a familiar process but could not defend a recommendation;
• sustained a professional discussion until the topic became unfamiliar;
• adapted wording for one audience but not for another;
• recovered from two misunderstandings without changing the intended outcome.

These statements preserve useful evidence without pretending to issue an official level.

Build a speaking portfolio before making a claim

Use at least three task families. A practical portfolio might include:

1. Exchange: obtain information, negotiate an arrangement or solve a service problem with another speaker.
2. Explanation: describe a process, event or decision so a listener can act on it.
3. Position: support a recommendation, address an objection and consider a changed condition.

Within each family, vary one meaningful condition: topic familiarity, relationship, time pressure, audience knowledge or consequence. Keep the prompt, recording, listener outcome, assessor note and a later transfer attempt. A speaking practice log helps separate observations from impressions.

The portfolio still does not create an ILR rating. Its value is diagnostic: it shows which functions travel and which depend on rehearsal.

A worked example without a homemade rating

Suppose Lena can explain her team's normal reporting process clearly. In a second sample, she answers follow-up questions from a new colleague. In a third, she must defend a change to the process when a manager raises cost and timing concerns.

The first sample demonstrates a prepared professional explanation. The second adds interaction. The third tests support, audience adaptation and unfamiliar pressure. If Lena struggles in the third sample, the useful conclusion is not “she is definitely ILR 2.” It is:

Lena can explain the familiar process and answer routine questions. She needs practice supporting a recommendation when constraints change.

That note generates the next task. Repeat the third scenario with different constraints, then use a record–review–repeat cycle to see whether the improvement transfers.

Do not convert ILR numbers by resemblance

ILR, CEFR, IELTS, TOEFL and other systems were built for different uses. Similar phrases or a tidy online conversion chart do not establish that two scores are interchangeable.

Before using a crosswalk, ask:

• Who produced it?
• Which exact assessment and skill does it cover?
• Was it based on empirical samples or editorial judgement?
• Does the receiving organisation accept it?
• What uncertainty or score range accompanies the mapping?
• Is the comparison current for the test version being used?

Read CEFR spoken interaction and CEFR spoken production within the CEFR framework. Do not relabel those observations as ILR evidence.

Separate the framework from a particular test

The ILR FAQ notes that US government agencies use the descriptions when scoring language proficiency, while individual tests differ in important respects (official ILR FAQ). Therefore a scale description alone cannot tell a candidate which test tasks, administration rules or decision thresholds apply.

For a formal requirement:

1. identify the agency, employer or programme receiving the result;
2. ask which named assessment and skill are required;
3. confirm whether a self-rating or outside certificate is accepted;
4. use the current test instructions;
5. report the original score without an invented conversion.

For ordinary learning, use the fluency assessment tool to compare samples over time. Treat its output as practice feedback, never as an ILR score or government credential.

Limitations and disclosure

LinguaLive is not affiliated with or endorsed by the Interagency Language Roundtable or any US government agency, and it does not issue or predict ILR ratings. The planning table, portfolio design and Lena example are original editorial aids, not official descriptors, prompts or scoring procedures. This guide summarises the scale's structure so readers can avoid false conversions; the current official ILR descriptions control for formal use. An authorised assessment may require more samples, trained raters and procedures not covered here. Language performance also varies with topic, interlocutor, conditions, accommodations and recording quality.

Sources and editorial review

This guide was checked against its primary official or academic reference on 29 July 2026. Language usage can vary by region, relationship, and situation. Review the primary source.