DutchMouth isn't built on hunches. Every design decision is grounded in cognitive science research on memory and learning.
Dutch verbs are morphologically complex. Unlike English, where most verbs follow simple patterns (walk/walked/walked), Dutch requires learners to master:
Most Dutch courses treat verbs as vocabulary items. DutchMouth treats them as grammatical systems requiring dedicated, focused practice.
Production, not selection
DutchMouth uses typed production exclusively. No multiple choice. No word banks. When you type "gewerkt" from memory, you engage deeper cognitive processing than when you select it from four options.
Research consistently shows that retrieval practice (actively recalling information) produces stronger memory traces than recognition tasks. The act of retrieval itself strengthens memory, independent of additional study time.
"The testing effect demonstrates that the act of retrieval itself strengthens memory." — Roediger & Karpicke (2006), Psychological Science
Review at the optimal moment
The spacing effect shows that distributed practice produces better long-term retention than massed practice. By reviewing material just before you would forget it, you maximize retention per minute of study time.
| Level | Interval | Cumulative |
|---|---|---|
| 0 | 1 day | 1 day |
| 1 | 3 days | 4 days |
| 2 | 7 days | 11 days |
| 3 | 14 days | 25 days |
| 4 | 30 days | 55 days |
| 5 | 60 days | 115 days |
| 6 | 120 days | 235 days |
| 7 | 240 days | 475 days |
The exponential growth in intervals follows the empirically observed forgetting curve. Early intervals are short because new memories decay quickly. Later intervals stretch because consolidated memories are more stable.
Confidence before advancement
A verb form advances to the next interval only after two consecutive correct responses. A single correct answer might be a lucky guess or temporary recall.
Requiring two consecutive successes before promotion reduces false positives and protects against the "illusion of competence"—feeling like you know something because you just saw it, when the knowledge hasn't actually consolidated.
Honest feedback, real progress
An incorrect response resets the verb to interval 0 (review tomorrow). This may feel strict, but errors indicate the memory trace has degraded below the recall threshold.
Continuing at the same interval would likely produce another failure. Resetting allows the memory to reconsolidate with shorter intervals before attempting longer gaps again. This ensures that promoted verbs genuinely represent stable knowledge.
Each verb requires mastery of five specific forms—the minimal set for productive competence:
From these five forms, you can derive all other conjugations through rules. Learning all 15+ possible forms would be inefficient when most are rule-governed.
Dutch perfect tenses use either "hebben" or "zijn" as the auxiliary—one of the most persistent error sources for learners:
The rules are complex: most verbs use hebben, motion verbs with a destination use zijn, state-change verbs use zijn, and some verbs vary by meaning. DutchMouth requires explicit auxiliary knowledge because this distinction often persists as an error source even at advanced levels.
Dutch separable verbs split in main clauses—a major stumbling block for learners:
DutchMouth tracks separability as a core property because word order changes fundamentally with separable verbs, the past participle inserts "ge" between prefix and stem (op-ge-beld), and learners must know which verbs separate to produce correct sentences.
DutchMouth organizes 2000 verbs into seven semantic categories, taught in a specific order designed to maximize speaking ability early:
Categories 1-3 form the grammatical backbone of Dutch. You cannot form sentences without copula verbs (zijn, hebben), modals (kunnen, moeten), and basic motion verbs (gaan, komen). These ~320 verbs unlock basic conversation before you've learned a single "action" verb.
Within each category, verbs are ordered by morphological complexity:
Teaching irregular verbs early forces rote memorization. Teaching them after regular patterns allows learners to understand how they deviate from the norm, making them more memorable.
The most common SRS failure mode is review pile-up. Learners get overwhelmed, skip sessions, and the backlog grows exponentially.
When reviews exceed 30: New verb introduction automatically throttles to 2 per day until the backlog clears.
Reviews always take priority over new learning. Skipped reviews decay faster than new material would benefit from introduction, and the psychological burden of a large backlog reduces motivation. Spaced repetition only works if the spacing is maintained.
DutchMouth focuses exclusively on verb morphology. This is intentional. Verb conjugation is a discrete skill that benefits from focused drilling.
The focus is the feature. By doing one thing well, DutchMouth provides a foundation that general-purpose language apps cannot match for this specific skill.
Ebbinghaus, H. (1885). Über das Gedächtnis: Untersuchungen zur experimentellen Psychologie.
Roediger, H. L., & Karpicke, J. D. (2006). Test-Enhanced Learning: Taking Memory Tests Improves Long-Term Retention. Psychological Science, 17(3), 249-255.
Pimsleur, P. (1967). A Memory Schedule. The Modern Language Journal, 51(2), 73-75.
Cepeda, N. J., Pashler, H., Vul, E., Wixted, J. T., & Rohrer, D. (2006). Distributed Practice in Verbal Recall Tasks: A Review and Quantitative Synthesis. Psychological Bulletin, 132(3), 354-380.