← All posts

What the Research Says

·3 min read

What the Language Learning Research Actually Shows

Second language acquisition research has produced robust findings about input, output, practice frequency, and fluency development. Most popular language learning advice is poorly calibrated against this evidence.

Where SLA Research and Popular Practice Diverge

Second language acquisition (SLA) is a mature field with decades of rigorous research. The findings challenge many popular language learning approaches: the primacy of grammar instruction, the value of vocabulary lists, and the optimal balance of input versus output. What the evidence actually supports is sometimes counterintuitive.

What Replicates Strongly

Comprehensible input is the primary driver of implicit grammar acquisition. Krashen's Input Hypothesis — that language is acquired through understanding messages just beyond current ability (i+1) — has accumulated substantial empirical support. Naturalistic acquisition studies, immersion research, and developmental sequence data all confirm that high-frequency exposure to comprehensible input develops grammatical competence that explicit grammar instruction alone cannot replicate.

Spaced repetition outperforms massed study for vocabulary retention by a factor of 2–3×. Cepeda et al.'s meta-analysis (254 studies) confirms that distributed vocabulary learning with spaced intervals produces dramatically better retention at delayed tests than equivalent time spent in massed review. Anki and similar SRS tools are the most evidence-aligned approach to vocabulary acquisition.

Output practice (speaking and writing) develops fluency that input alone doesn't build. Merrill Swain's Output Hypothesis shows that language production forces learners to notice gaps in their grammatical competence — gaps that passive comprehension doesn't reveal. Output practice activates procedural memory systems for automated retrieval that input activates less efficiently. Learners who read/listen extensively but don't produce often develop excellent comprehension with poor spontaneous production.

Daily short practice outperforms equivalent weekly sessions for procedural language skills. For speaking and writing, frequency maintains the procedural memory networks that language production requires. A daily 10-minute speaking session activates retrieval pathways more effectively than a weekly 70-minute session with the same total time.

Pronunciation acquires a sensitive period beyond which native-like accent is rare. Neuroscience of critical periods shows that phonological acquisition is most efficient before puberty. Adults can achieve excellent pronunciation with deliberate practice (shadowing, phonetic training) but rarely achieve native-level prosody acquired in childhood. This doesn't preclude good communication — it sets expectations for pronunciation training goals.

What the Research Can't Tell You

Individual acquisition rates, optimal input difficulty calibration, and the most productive balance of input/output vary by learner. Some people acquire vocabulary rapidly through input; others need more explicit study. Tracking vocabulary size growth, comprehension test scores, and speaking fluency ratings over 4–8 week experimental blocks is the most reliable way to identify your highest-leverage approach.

Based on

Language Learning research

Browse the evidence →

Try it yourself

Run your first experiment today

SteadyPractice gives you structured templates, randomized tracking, and plain-English results — so you can find what actually works for you.