A university-ready redesign for Chinese as a Second Language (CSL)
Combined artifact: (1) Full program blueprint (approx. 12–15 pages equivalent) + (3) fork-style stress test matrix

0) Executive intent (why this exists)

Most university CSL sequences are built around credit hours, textbook coverage, and grading throughput. High proficiency (especially advanced speaking) instead requires a design that guarantees:

Sustained input volume (listening/reading)
High speaking density (minutes per learner per week, not per class)
Spaced retrieval (memory architecture, not “review chapters”)
Explicit cognitive load management (tones + script + meaning cannot all be maxed at once)
Measurement discipline (proficiency outcomes are published, not assumed)

This blueprint converts those requirements into a program that can be staffed, scheduled, assessed, and audited.

1) Program outcomes and targets

1.1 Proficiency targets (ACTFL primary; CEFR secondary; HSK supportive)

Use: ACTFL OPI/OPIc + AAPPL (or equivalent) at fixed checkpoints.
Use HSK as a literacy/vocabulary reference, not as the definition of proficiency.

Targets by year (non-heritage track):

End Year 1: Novice High → Intermediate Low (S/L), Novice High+ (R)
End Year 2: Intermediate Mid (S), Intermediate High (L/R)
End Year 3: Intermediate High → Advanced Low (S), Advanced Low (L/R)
End Year 4: Advanced Low (S) as standard; Advanced Mid (S) for “Advanced Track” cohort

Heritage track targets: similar speaking targets often arrive earlier; literacy targets must be explicit and differentiated.

1.2 What “high proficiency” means operationally

High proficiency is not “knows X characters.” It is:

Spontaneous speaking at length with repair strategies
Listening across registers (classroom + authentic media)
Reading with speed and comprehension (graded → semi-authentic → authentic)
Writing/typing sufficient for real communication (typed production prioritized)

2) The hour model (closing the gap)

2.1 Required weekly dose (minimum)

Class time alone cannot carry the load. The program must require and verify out-of-class work.

Non-immersion semesters (recommended minimum):

Year 1–2: 8–10 hours/week total Chinese time
Year 3–4: 12–15 hours/week total

Where “total Chinese time” includes: class, speaking lab, required media input, extensive reading, retrieval practice.

2.2 The “Input Lab” concept (institutional hack)

Create a 1-credit “Input Lab” each semester (Years 1–3; optional Year 4) that formalizes:

required listening hours
required reading hours
comprehension checks (low-friction)
short oral responses (to connect input → output)

This is the cleanest way to align the university credit system with acquisition reality.

3) Architecture: three systems, one pacing rule

3.1 The three systems (must be designed explicitly)

Sound system: tones, phonology, perception, production
Meaning system: lexicon, grammar patterns, routines, discourse
Script/literacy system: characters, reading fluency, typing, writing

3.2 The pacing rule (anti-collapse)

You cannot max all three at once without overload.
Design rule: In any given 6-week block, one system is the “front,” one is “steady,” one is “maintenance.”

Example:

Weeks 1–6 Year 1: Sound = front, meaning = steady, script = maintenance
Mid Year 2: Script/reading = front, sound = maintenance, meaning = steady
Year 3: Meaning/discourse = front, script = steady, sound = maintenance

4) Program structure (tracks, sections, staffing)

4.1 Two tracks + a gate

Standard Track: designed to produce Advanced Low (S) by Year 4 for successful completers
Advanced Track: requires one intensive summer / immersion or equivalent; targets Advanced Mid speaking
Gate at end of Year 2: placement into standard vs advanced track based on measured proficiency + workload capacity (not “grades”)

4.2 Heritage vs non-heritage: structural separation

Default requirement: separate tracks OR separate labs + assessments, because mixed classrooms predictably mis-serve both.

Minimum acceptable compromise:

shared lecture (culture/content)
separate speaking labs
separate literacy pathways
separate proficiency targets and placements

4.3 Staffing model (realistic)

For each level, you need:

1 main instructor per section (or co-teach)
TAs or trained conversation leaders for labs
graders or tooling for frequent low-stakes checks

If the institution cannot staff speaking labs, it cannot honestly claim it is aiming for advanced speaking outcomes.

5) Year-by-year blueprint

YEAR 1: Sound-first + input-first + typed literacy

Primary goal: prevent tone/phonology from becoming permanent debt while building a habit of input volume.

5.1 Weekly schedule (example)

Class (4–5 contact hours/week)
- 60% listening/speaking tasks, 30% meaning patterns, 10% literacy orientation
Speaking Lab (1 hour/week, small group ≤ 6)
Input Lab (1 credit, 2–3 hours/week)
- Listening: 90–120 min/week graded video/audio
- Reading: 45–60 min/week graded readers or microtexts
Retrieval practice (20–30 min/day)
- spaced vocabulary + tone drills + sentence frames

5.2 Tone training protocol (non-negotiable)

Daily micro-doses:

10–15 min/day tone perception (minimal pairs, identification)
10 min/day guided production (shadowing, slowed → natural)
Weekly:
short tone check (low stakes)
repair clinic: 2–3 high-error patterns per cohort

5.3 Script policy Year 1 (avoid the handwriting trap)

Primary literacy: recognition + typing + reading
Handwriting: limited and strategic
- 150–250 core characters by end of Year 1
- focus on components/radicals for structure awareness
No “copy 30 characters nightly” as the main learning engine

5.4 Pinyin sequencing (adaptive, not ideological)

Weeks 1–6: pinyin heavy for phonology + fast vocab
Weeks 6–12: mixed pinyin + characters
Semester 2: character-forward for reading; pinyin remains a scaffold for new items when needed

Deliverable: by end of Year 1, learners can function without pinyin in familiar texts but still use it strategically.

YEAR 2: Reading fluency + speaking density

Primary goal: prevent “reading outruns speaking” from becoming the default outcome.

5.5 Weekly schedule (example)

Class (4–5 hrs/week): discourse routines + listening + reading integration
Speaking Lab (2 hrs/week) OR (1 hr/week lab + structured partner sessions)
Input Lab (2–4 hrs/week) (bigger volumes)
- Listening: 150 min/week
- Reading: 90 min/week
Retrieval system: cumulative spiral quizzes weekly (not unit resets)

5.6 Literacy scaling: frequency bands + structure

Teach characters in frequency bands, but don’t rely on frequency alone:
- explicit radical/component instruction
- recognition speed targets (not just “knows”)
Introduce “reading speed ladders” (timed readings with comprehension)

5.7 Speaking density requirement (measurable)

Set a minimum:

≥ 45–60 minutes of learner speaking per week (outside of whole-class time)
Tracked via:
lab tasks (recorded clips)
partner sessions (structured prompts + short evaluation rubrics)

YEAR 3: Domain competence + long-form input

Primary goal: move from “classroom Chinese” to “topic competence” and sustained comprehension.

5.8 Structure: content lanes (choose 2–3)

Offer lanes such as:

Media & society
Business & negotiation
Tech/science discourse
Taiwanese/PRC regional register exposure (optional)

Each lane supplies:

a content reader packet
weekly authentic clips
speaking tasks tied to content (summaries + arguments + Q&A)

5.9 Extensive reading becomes core

Reading: 120–180 min/week
Listening: 180 min/week
Speaking: labs + presentations + role plays

5.10 Pronunciation maintenance

Not “tone drills forever,” but:

fluency under speed
listening in noise
targeted repair for persistent errors

YEAR 4: Advanced speaking pathway + verification

Primary goal: make advanced speaking a designed outcome, not a lottery.

5.11 Two-track Year 4

Standard Year 4

Advanced Low speaking expected for successful completers
Capstone: presentation + Q&A + discussion facilitation

Advanced Track Year 4

Requires immersion/intensive experience OR intensive domestic substitute
Targets Advanced Mid speaking
Capstone: debate + field interview + long-form narrative + live interpretation exercises (as appropriate)

5.12 Public-facing capstone portfolio

Each student produces:

2 recorded presentations
2 recorded conversations (unrehearsed prompts)
1 listening/summary task
1 reading/argument response
typed writing samples (email, short essay, notes)

Portfolios enable internal moderation and external review.

6) Pedagogy: what happens inside instruction

6.1 Core classroom pattern (every week, every level)

Input first (listen/read) → meaning extraction → guided output → feedback → retrieval
No week should be “learn chapter → test → forget.”

6.2 Error policy (especially for tones)

Early: tolerate meaning-preserving errors but correct high-impact tone confusions
Mid: tighten accuracy gradually
Late: shift to fluency + repair + register

6.3 Grammar: treat as pattern inventory

Teach grammar as:

high-frequency constructions
response templates
recombination practice
Not as long metalinguistic lectures.

7) Assessment and measurement discipline

7.1 Checkpoints (required)

Entry placement (separate heritage placement)
End of Year 1, 2, 3, 4:
- ACTFL-aligned speaking (OPI/OPIc)
- listening/reading (AAPPL or comparable)
Mid-semester: short diagnostic checks for tones, listening, reading speed

7.2 Publish outcomes (program accountability)

Annually publish:

proficiency distributions by cohort and track
retention + “advanced attainment” rates
impact of immersion participation
differences heritage vs non-heritage

8) Operations: how to implement without fantasy resources

8.1 Minimal viable version (Year 1–2 deployment)

If you can’t do everything at once, do this first:

Speaking labs (small group)
Input Lab credit requirement
Tone micro-dose system + weekly checks
Spiral retrieval quizzes weekly
Script policy: typed literacy prioritized, handwriting capped and strategic
Measurement checkpoints

8.2 Tooling (allowed, but not the point)

Use technology only to enforce:

spaced retrieval schedules
input tracking + comprehension checks
speaking clip submissions + rubric feedback

Avoid turning “apps” into substitutes for contact and input volume.

9) CSL Fork-Style Stress Test Matrix

Purpose: simulate failure conditions that reveal whether the program is truly designed for proficiency or merely for course completion.

Each test has:

Stress condition
Expected failure mode
Detection signal
Mitigation / redesign lever

9.1 Stress Class A — Time starvation (the institutional default)

A1. Remove Input Lab (no required input hours)

Failure mode: reading/speaking plateau; progress becomes selection-based
Signal: proficiency distributions widen; high tail persists, median stalls
Mitigation: reinstate Input Lab or embed required hours into core course grading

A2. Cut speaking labs (class-only speaking)

Failure mode: advanced speaking becomes rare; reading outruns speaking
Signal: speaking outcomes lag 1–2 sublevels behind reading by Year 3–4
Mitigation: restore labs, reduce section sizes, require partner sessions

9.2 Stress Class B — Cognitive overload (CSL-specific collapse)

B1. Handwriting-heavy Year 1 (copy load spikes)

Failure mode: input hours shrink; tone debt accumulates; motivation drops
Signal: tone accuracy stagnates; listening comprehension lags; attrition rises
Mitigation: cap handwriting; shift to recognition/typing; separate literacy drills from tone work

B2. Characters-from-day-1 maximalism (pinyin removed too early)

Failure mode: phonology acquisition slows; vocabulary uptake slows; frustration rises
Signal: poor tone perception under speed; slow oral vocab growth
Mitigation: adaptive pinyin sequencing; protect sound system in first 6–10 weeks

B3. Tone instruction only “embedded” (no dedicated tone micro-doses)

Failure mode: tone errors fossilize; comprehension suffers in authentic speech
Signal: persistent tone confusion patterns; listening drop in noisy/fast speech
Mitigation: daily micro-doses + weekly checks + repair clinics

9.3 Stress Class C — Assessment distortion (“credentialing gravity”)

C1. Grade system rewards neat handwriting, not proficiency

Failure mode: learners optimize for copying; speaking stagnates
Signal: high course grades with low speaking proficiency
Mitigation: reweight grades toward speaking + comprehension + retrieval performance

C2. Chapter tests reset each unit (no cumulative retrieval)

Failure mode: forgetting between semesters; “Year 3 feels like Year 2 again”
Signal: backslide at term start; slow recovery each semester
Mitigation: spiral retrieval schedule; cumulative quizzes weekly

9.4 Stress Class D — Heritage mixing (structural mismatch)

D1. Mixed heritage/non-heritage in same core path

Failure mode: pacing misfits; both groups disengage; literacy suffers for heritage, speaking suffers for non-heritage
Signal: bimodal outcomes; classroom tension; retention drops
Mitigation: split tracks or split labs + separate targets

9.5 Stress Class E — “Immersion myth” (outsourcing the hard part)

E1. Assume study abroad fixes everything; reduce domestic rigor

Failure mode: inconsistent outcomes; SA benefits only those who engage heavily
Signal: huge variance among SA participants; weak speaking gains for some
Mitigation: structure SA with targets, tasks, and accountability; don’t reduce domestic input design

9.6 Stress Class F — Scale stress (what happens when enrollment doubles)

F1. Section sizes grow; speaking minutes per learner drop

Failure mode: speaking outcomes collapse first, then listening
Signal: speaking proficiency distributions shift downward within 1–2 cohorts
Mitigation: protect speaking density with labs, staffing, peer systems

9.7 Stress Class G — Integrity test (the program is “polite”)

G1. Everyone passes, outcomes aren’t measured

Failure mode: the program cannot tell if it works; drift becomes invisible
Signal: no published proficiency distributions; only grades exist
Mitigation: mandatory checkpoints + published dashboard + external moderation sample

10) Implementation roadmap (18–24 months)

Phase 1 (0–6 months): Build the enforcement spine

adopt proficiency checkpoints
create Input Lab syllabus + compliance system
define script policy (typed literacy, handwriting cap)
launch speaking labs
build retrieval schedule templates

Phase 2 (6–12 months): Run the first stress tests

run A1/A2/B1/C2 tests as controlled comparisons (or natural experiments)
publish first dashboard internally
split heritage pathways operationally

Phase 3 (12–24 months): Scale and verify

expand content lanes (Year 3)
standardize capstone portfolio (Year 4)
add external moderation sample for speaking ratings

11) Appendices (ready-to-use templates)

Appendix A — Year 1 weekly template (sample)

Mon: tone micro-dose + listening clip + 10 vocab retrieval
Tue: class speaking task + reading microtext + retrieval
Wed: tone micro-dose + graded listening + 3-minute speaking clip
Thu: lab session + feedback + retrieval
Fri: comprehension check + spiral quiz + short reflection (2 questions, not journaling)

Appendix B — Speaking rubric (minimal)

comprehensibility (meaning gets across)
fluency (pausing/flow)
control of high-frequency patterns
tone intelligibility (priority errors)
repair strategy use

Appendix C — Input Lab compliance (low-friction)

weekly listening: 3–4 items + 6–10 comprehension questions total
weekly reading: 1–2 items + short oral summary
random spot-check interviews