Score Item 9 "Best Language" from what the person can do in Mandarin, not from how fluently the interpreter relays it — the examiner asks the question in Mandarin through the interpreter, reads the answer as interpreted, and credits the language the person names and prefers in daily life. If the response is ambiguous, contradictory, or switches language midway, the examiner probes once through the same interpreter and follows the Interviewer's manual before committing a code. The interpreter's competence is never the score. The score describes the person, and the person's best language may well be one the examiner does not speak at all.
The complication is that Item 9 sits on a boundary. It asks about language, and the interpreter turns language into something the examiner hears indirectly. So the sections below work through the routing, the jurisdiction difference, the scoring mechanics, the probe script, the failure cases, the cost arithmetic and the documentation trail. Each part answers a different question the same examiner will ask on the day.
Routing the Interview Before Item 9 Is Reached
Item 9 only makes sense inside the sequence it belongs to. In the U.S. naturalization interview under the N-400, "Best Language" is the question that captures whether the applicant can read, write and speak English, and whether they will accept an interpreter or an exemption. In an ESL placement or an intake screening, it captures the person's preferred language for instruction. In an FOIA or emergency-intake scenario, it captures the language in which information should be delivered. The order of questions is not arbitrary; it establishes which language the rest of the interview will run in.
For the N-400, the officer must first test English eligibility before "Best Language" is asked. An applicant who qualifies for a 50/20 or 55/15 exemption, or who has a documented disability exception under the appropriate ADA accommodations process, does not take the English portion at all, so Item 9 may be routed through the exemption. The 50/20 exemption applies to applicants who are 50 or older and have lived in the U.S. as lawful permanent residents for 20 years; the 55/15 exemption applies at 55 and 15 years. Confirm the current formulation on the official naturalization guidance before applying either.
If the applicant does not qualify for an exemption, the interview is conducted in English, and the language on the form is the language the applicant listens in. A Mandarin-speaking applicant who answers Item 9 by saying "ao, wo shuo pu tong hua" ("I speak Mandarin") has given the examiner a truthful answer to a question the examiner cannot score from sound alone. The score depends on how the answer is captured in the field and on whether the English portion of the test can be satisfied through translation. It cannot. That has to be said plainly: an interpreter can translate Item 9 and the surrounding language questions, but English proficiency and civics knowledge are tested in English, and the interpreter cannot substitute for either.
The practical routing rule is this: if the person needs an interpreter for Item 9, the case is in one of four positions — an English-test applicant who nevertheless uses a language accommodation, an exemption applicant, a disability-accommodation applicant, or a screening interview that is not an English-proficiency test. The examiner should identify which before the interpreter connects. The wrong routing is the most durable error in this item, and no amount of careful scoring recovers from it.
What "Best Language" Measures in Different Interview Systems
"Best Language" does not mean "best" in the sense of which language the person is proudest of. It means the language the person can most readily understand, speak and be understood in during the interview in front of them. In the N-400, it is a short field with limited codes; in an ESL placement, it is often the trigger for the whole test battery; in an investigative interview, it is a custody-and-comprehension decision. Three systems, one field, three different sets of consequences.
For U.S. naturalization, the field is paired with questions about reading, writing, speaking and understanding English, and with the interpreter-or-exemption routing. The officer's job is to establish which language governs the rest of the interview. A person who answers Item 9 with Mandarin, Cantonese, Toishanese or another variety must be allowed to give the actual variety, and the field should reflect that wording even if the form's dropdown shows "Chinese." The difference matters for the interpreter assignment, not the candidate's identity.
For an ESL or educational placement, Item 9 routes the student to the right assessment. A Mandarin speaker who reports English as their best language but who is placed in English-language instruction after an initial screening has not "passed" the item; they simply have a different route. The school language usually distinguishes between the language of instruction and the language of the home, and Item 9 captures the language the student wants to be evaluated in first.
For an investigative or emergency intake, Item 9 is a comprehension and rights decision. The person is read the information in the best language, the interpreter is instructed not to summarise, and the intake worker records the language in the same field as the interpreter used. The scope boundary here is serious: if the person may be making statements with legal consequences, the general education in this article is not a substitute for the local jurisdiction's interpreter-access policy, and a qualified local professional and the controlling agency guidance should be consulted before the record is finalized.
Short Answer for the Item 9 Code Itself
When the examiner reaches Item 9 with a Mandarin-speaking person on a video call and a professional interpreter already connected, the score is the language the person names as their best language, recorded exactly as stated. On the N-400, the dropdown often lists English, Spanish and "other"; the officer records "Chinese" or the specific variety in the free-text note and moves on to the language-access routing. On an ESL screening, the score is the language code the student names. On a legal or emergency record, it is the language in which the interview will continue. The common rule across all three is that the score describes the person, not the interpreter.
Three scoring boundaries matter. First, a person who answers in Mandarin before the interpreter has finished rendering a question has answered in Mandarin — the interpreter's English relay is not the answer and does not modify the code. Second, a person who names a language for family or home but who names English as the language they prefer for testing has two different facts, and the field must record both because Item 9 has to survive scrutiny later. Third, a person who cannot understand the interpreter at all has not answered Item 9; the officer must stop, reschedule with a different interpreter or modality, and document why the item could not be scored.
Scoring is not the same as interviewing. If Item 9 is going to be read into the record at a later hearing, the officer's note has to make sense to someone who was not on the call. Record the language named, the exact words used by the person, whether the question was asked once or after a probe, and whether the interpreter rendered the answer in first person. "Interpreter confirmed" is not a note; it is a placeholder that will look like padding at the hearing.
Reading Item 9 Through an Interpreter Without Scoring the Interpreter
An interpreter does three jobs on the call: renders the question into Mandarin, renders the answer back into English, and manages the floor on overlap. None of those jobs is the construct Item 9 measures. The construct is the person's self-reported best language and, where the system asks for it, the person's ability to function in that language during the interview. A strong interpreter on a weak candidate can make the interview sound successful. A weak interpreter on a strong candidate can make it sound like a failure. The score sits closer to the candidate than the audio stream suggests.
To keep the interpreter out of the score, the examiner should use a consistent rendering standard. Render the question in full, wait for the interpreter to finish, wait for the person to answer, and only then let the interpreter relay back. If the person answers before the question finishes, do not credit the answer as a full comprehension of the question. If the person's answer is a restatement of the interpreter's words, treat it as repetition, not as spontaneous speech, and probe. The pipeline is: examiner question → interpreter → person answer → interpreter → examiner. Any skip in the chain is a probe, not a score.
A language-identification pitfall needs naming here. Mandarin ("putonghua," "guoyu," "huayu") is not the same as Cantonese ("Yue"), Hokkien ("Minnan"), Toishanese or Shanghainese, and a person who says "Chinese" is telling you less than a person who says "Mandarin." Common Code misuse: recording "Chinese" as if it meant Mandarin at the same time as the interpreter is speaking Cantonese. The field and the interpreter assignment disagree, and the disagreement surfaces later. Ask for the language the person actually prefers and note it, even when the form has one lumped "Chinese" option.
The second diagnostic is dominance. Ask, through the interpreter, which language the person uses at home, at work, with children, and when they think out loud. A person whose answers in Mandarin are spontaneous and unhesitating while their English answers are halting has told you something about dominance. Dominance is not the same as proficiency, and the two are not interchangeable on the form. If the form asks for the best language, dominance is the better proxy; if the form asks for English ability, measure that separately.
The Jurisdiction Difference the Form Doesn't Show
A Mandarin-speaking applicant in California on a video call is in one system. A Mandarin-speaking applicant in Ontario is in another. The federal N-400 and the federal immigration scheme in Canada both put language testing at the centre of the interview, but the composition of the test, the exemption categories and the interpreter-access rules differ in ways that change the score. The Canadian citizenship language requirement is generally described as Canadian Language Benchmark 4 in speaking and listening for applicants aged 18 to 54, with exemptions for older applicants and for people with a disability; the citizenship knowledge test is administered separately. Confirm both on the Government of Canada's current citizenship guidance before use.
The practical difference for Item 9 is where the interpreter sits in the sequence. In the U.S., the interpreter is a language-access tool and the English test is still administered in English; the interpreter does not create an exemption. In Canada, the applicant may submit third-party evidence of language ability at the outset, and the interview's language questions are routed through that evidence rather than re-tested in a single sitting. For the same Mandarin-speaking person on a video connection, the U.S. officer is asking "can I complete this interview in English, and does this person want an interpreter for the rest of it?" The Canadian officer is asking "does the file already show the language evidence, and does this interview need further language testing?"
Limits of the U.S. exemption categories should be stated exactly. The 50/20 and 55/15 exemptions change the English test and the civics-test language, not the underlying facts about residency or age, and the applicant must still meet the physical-presence and continuous-residence requirements. A disability exception under the appropriate ADA accommodations process is separate again and requires documented medical evidence acceptable to the reviewing authority. When the case is a mixed one — say, an applicant eligible for a 55/15 exemption who nevertheless wants to demonstrate English — the interpreter-assisted route is not automatic; the officer must confirm what the applicant is choosing to do and document that choice.
How the Interpreter's Presence Changes the Examiner's Judgement — and Shouldn't
On a video call, the examiner sees the person's face at a delay, hears the interpreter's English in the foreground and the person's Mandarin in the background, and often has the interpreter's notes rather than the person's words. That mixture shifts the exam's centre of gravity from the person to the interpreter, and the shift is systematic. Studies of remote interpretation consistently find that turn-taking and overlap increase in video-delivered sessions, and that the interpreter's voice carries more confidently than the second-language speaker's. The examiner who is not careful will score the confident English rather than the evidence of ability.
Three practical adjustments help. Take notes in the person's first language where the interpreter makes it available, rather than only in the interpreter's rendering. Ask the interpreter to work in first person ("I speak Mandarin at home") so the record shows a person's voice, not the interpreter's summary. Use the person's name, not the interpreter's, when addressing the question, so the person is the one being asked. These three points cost nothing and make Item 9 read as a person's answer when the file is reviewed later.
A second adjustment concerns the interpreter's role. The interpreter should not be told the answer to Item 9 or coached to lead the question. A professional interpreter is briefed on the setting and the parties; the answer itself is the applicant's. A Mandarin-speaking applicant who is asked through one interpreter and answers in English, or who answers with an interpreter who speaks a different Chinese variety, has generated a factual problem for the record, not just a scoring one. The note should say what the person said, in whose voice, and through whom.
The third adjustment is time. Video interpretation is slower than in-person, and the delay between the end of a question and the beginning of the answer gives the examiner time to think about the score before the answer arrives. The examiner should not pre-score the answer during the gap; the answer is the person's next utterance, and the examiner should wait for it. The item's purpose is understanding, and the examiner's judgement about a person's abilities should not run ahead of what the person actually says.
Video-Call Failure Modes and How Each Changes the Coding
A video interpreter call can fail in ways that change the score if the examiner does not separate equipment failure from language ability. Audio feedback on the platform can make the person sound halting; a dropped connection can lose the interpreter's rendering of a question; and a video quality drop can hide the person's mouth during the answer. The examiner's default when equipment fails is to stop, fix the equipment, and restart the item — not to infer language ability from the degraded audio. If the interpreter returns and the item is re-asked, the earlier attempt is a first attempt, not a score.
Another failure mode is the mismatched interpreter. If the interpreter assigned speaks Cantonese but the person's stated language is Mandarin, the recorder must flag the mismatch. A person who says "wo shuo pu tong hua" to a Cantonese interpreter will hear something else, and the answer may be counted incorrectly; the safe course is to stop, request the correct interpreter, and re-ask. A person whose stated language is Mandarin but whose real dominance is Cantonese will reveal it under a correct interpreter in the first thirty seconds. Note that shift in the file; it is evidence about dominance, not evidence about proficiency.
A third failure is the unscored prerequisite. Item 9 is not supposed to be scored without first asking the routing questions; if the interview has not yet asked whether the person requires an exemption or intervention, the officer should not record a "Best Language" that is really a guess. The safer sequence is to ask the routing question first, ask Item 9 as a second question, and record both answers. A person who is routed to an exemption has answered Item 9 in the sense of "what is my best language?" but the answer does not affect the English test because there is no English test.
A fourth failure concerns the interface. If the platform used for the video call has an interpreter agent, request a "translation-only" session for the Item 9 block unless the examiner has a specific reason to use a "bilingual confirmation" integration. A translation-only session will render the answer in English without a corroborating model guessing at the intent, and the examiner can then ask the follow-up through the same interpreter. Treat any model-generated paraphrase as unreviewed and use only a human interpreter for Item 9 and its probes.
Recording the Answer, Handling Cold Probes, and Avoiding Common Mistakes
The record for Item 9 is a person's self-report plus the routing facts. The note should include: language name as stated, whether it is the person's dominant language, whether the person understands the interpreter, whether the person spoke before the interpreter finished, the result of any probe, and whether the person was routed to an exemption or English test. A person who is not asked at least one routing question before Item 9 has not been routed, and the field is likely to be wrong. A "not entirely clear" answer is not a score's friend; the record should show which part of the answer was clear and which part was probed.
The probe script for an ambiguous Item 9 runs through the same interpreter and is asked in the same language. First, ask the person to state in one sentence which language they understand best when they are tired or sick. Second, ask in which language they can explain a process — say, what to do at a pharmacy — without switching. Third, ask which language they would want read to them before signing a document in a legal or emergency setting. The three answers together support the code; the officer should quote them rather than restate them, because "understood" is not a note.
Common mistakes on this item are stable and worth naming. One is recording the interpreter's language as the person's language. A second is recording "Chinese" when the person named Mandarin, or "Spanish" when the person named a specific variety. A third is scoring Item 9 after the interview rather than during it, from memory rather than from the answers. A fourth is scoring the English-test refusal rather than the best-language statement; the two are different fields. A fifth is scoring after a probe that led the answer — a leading probe will often produce a "correct" answer that corrupts the record. Each mistake is recoverable if the note says what was actually asked and what was actually said.
Practical arithmetic is where the item becomes concrete. A 30-minute interpreter-assisted interview with a five-minute, federally funded language-access slot uses five of the thirty minutes for the language-exchange portion and twenty-five for substantive questions. At a per-minute video-interpreter rate of $2 to $4 per minute for a professional Mandarin-English interpreter, that five-minute slot costs $10 to $20; the whole interview costs $60 to $120 in interpreter time. Confirm the rate on the contract with your language-access provider, as rates vary by market, modality and the interpreter's credentials. If the exam needs two interpreters — one for the cold probe and one for the substantive portion — the cost roughly doubles for the overlapping minutes, and the case should be scheduled so the same interpreter covers both.
FAQ
What is Item 9 on the N-400 actually asking?
It asks which language you communicate in best, so the officer knows whether to route you to an interpreter, an English test, or an exemption category. The field is not an English-proficiency score and should not be described as one in the record. A person who answers "Mandarin" truthfully has answered it even if the officer also later tests English.
Should the score be reduced if the interpreter, not the applicant, spoke most of the time?
No. The interpreter's speaking share is a property of the session, not of the applicant, and the officer should score the applicant's own answers. If the person cannot understand the interpreter at all, the item should be suspended and rescheduled rather than scored low. The correct low score would need evidence about the person's ability, and the interpreter's volume is not that evidence.
Does speaking Mandarin make someone ineligible for a 50/20 or 55/15 exemption?
No. The 50/20 and 55/15 exemptions are based on age and years as a lawful permanent resident, and they do not depend on the applicant's language or on whether an interpreter was used. A Mandarin-speaking applicant who meets the age and residency facts can claim the exemption even if the interpreter-assisted session ran normally. Confirm the current exemption wording on the official naturalization guidance.
How do you handle a person who says "Chinese" when the form allows only one code?
Record the lumped code for the form and write the stated variety — Mandarin, Cantonese, Hokkien, Toishanese or another — in the free-text note. The form code is administrative; the note is what the next reviewer will use. A note that says only "Chinese" has thrown away the information that decides which interpreter to assign next time.
What does a Canadian citizenship interview do with Item 9?
The Canadian language requirement is generally expressed as Canadian Language Benchmark 4 in speaking and listening for applicants aged 18 to 54, with age and disability exemptions, and is usually evidenced through third-party test results submitted with the file rather than re-tested in one interview. Confirm the current requirement and the acceptable evidence on the Government of Canada's citizenship guidance. For a Mandarin speaker, the field is a routing fact, and the interview does not substitute for the evidence.
What if the professional interpreter and the applicant disagree about the language?
The applicant's statement controls the field, and the interpreter's observation goes in the note as a separate fact. If the interpreter believes a different Chinese variety is being spoken, the examiner should stop, reschedule with that variety, and re-ask the item. Recording both facts is better than recording a single code that hides the disagreement.
How many attempts at Item 9 are appropriate before it is marked incomplete?
One full question plus one non-leading probe is the normal allowance; a second probe is justified only if a new fact — such as a change in dominance or a different language variety — emerged. Multiple leading or repeated probes turn the item into a coaching exercise. If the person still cannot give a reliable answer after that, mark the item incomplete, state why, and reschedule.
Is scoring Item 9 different if the person is being interviewed for asylum or a protection claim?
The construct is still the person's best language, but the stakes of a misinterpreted answer are higher, and the interpreter-access standards in the local jurisdiction should be followed exactly. This article is general educational information and not legal advice; a qualified local professional and the controlling agency guidance should be consulted before a protection interview record is finalized.
The scoring rule this article has followed is not a legal opinion, a clinical finding, or a determination of immigration status. It is a general summary of how the item is supposed to be read, and it cannot replace the current official guidance, the local language-access policy, or advice from a qualified immigration lawyer in your jurisdiction. Anyone facing an interview or a protection-related proceeding should get that advice. If there is an immediate risk of harm, contact local emergency or crisis services.