One 90-minute lesson for true beginners, false beginners and heritage speakers

Objective 1 (all learners): by the end of the lesson, each student can order three items from a menu using 我要 plus a measure word, with at least two of the three measure words correct in a recorded 60-second role-play. Objective 2 (beginners and false beginners): each student can recognise and read aloud 8 of the 12 food nouns in characters, with pinyin available only on a help card. Objective 3 (heritage speakers): each student can sustain a 2-minute order-and-negotiate exchange, adding one clarification question and one price question, and can write three of the food nouns from memory.

The objectives share the same task and text, which keeps the class unified, but the success criteria differ in amount and demand. Beginners are assessed on accurate production of a short frame; heritage speakers are assessed on interaction length and literacy, which addresses their specific gap. The false beginners sit between the two, which is realistic. Making the criteria explicit prevents the heritage learners from finishing in three minutes, because the task requires an extension they have not yet automated.

0–10: whole-class warm-up with a menu image and six food items; students point and repeat, then sort items into 主食/饮料/菜. This activates schemata and gives beginners comprehensible input. 10–25: presentation of 我要 + measure word with three model orders; teacher checks understanding with quick whole-class choral and individual responses. 25–40: guided practice in mixed triads (one beginner, one false beginner, one heritage speaker) using a menu information gap; each student orders twice, heritage speakers model but must also read the characters. 40–55: station rotation - beginners do character matching on phones, false beginners do a price-list reading task, heritage speakers do a short written order with three characters from memory; the teacher circulates and the co-teacher supports the beginner station during her 30 minutes. 55–75: role-play rehearsal in the same triads with a checklist (order three items, use a measure word, ask a price question); heritage speakers must add a negotiation move. 75–85: two triads perform; class gives feedback against the checklist. 85–90: exit ticket, one sentence order.

The sequence moves from input to guided to free production, and grouping is deliberately heterogeneous so the heritage speakers' oral strength supports the beginners while the literacy station gives heritage speakers the demand they lack. The station rotation prevents the heritage learners from dominating oral work and addresses their reading gap.

Use a rubric with common criteria but level-specific descriptors. The criteria can be task completion, comprehensibility, vocabulary and structure, interaction, and literacy where relevant. For task completion, a beginner meets the standard by ordering three items with two correct measure words; a heritage speaker meets it by ordering, negotiating and closing the exchange. For vocabulary and structure, beginners are judged on the target frame and food nouns; heritage speakers on range and on avoiding English syntax transferred from family speech. For interaction, beginners are judged on responding to one prompt; heritage speakers on initiating and repairing. Scores should be reported against the student's pathway, not against a single raw total, and the teacher should record the pathway on the rubric before the performance.

The fairness principle is constructive alignment: the rubric must match the objective set for that learner. This avoids the common error of rewarding heritage speakers for fluency they already have while giving beginners no credit for accurate use of a limited frame. The teacher should also moderate with the co-teacher on two recorded performances per level to check that the descriptors are applied consistently.