Every Japanese learner knows the scenario. You studied consistently for a year or two, hit a decent intermediate level, then life happened — a job change, a hard season, burnout — and you stopped for six months. When you pick it back up, it feels like you’ve forgotten everything.
Have you?
The answer from SLA research is more nuanced — and more reassuring — than the “reset anxiety” most learners feel. But there are real effects too, and the pattern of what you lose first explains a lot about why the gap feels worse than it is.
What the Research Actually Studies
The phenomenon is called language attrition: the loss of previously acquired linguistic knowledge due to reduced use. Monika Schmid, the researcher most associated with this area, spent years tracking German-speaking emigrants in the UK and US, documenting what happened to their L1 as English became dominant. Barbara Köpke has done similar work on French-speaking emigrants. This isn’t exactly your situation — you’re not an emigrant shifting to Japanese as your dominant language — but the patterns they found illuminate what happens whenever a language falls out of active use.
The key finding: attrition is not uniform. Some language knowledge fades quickly; other parts are remarkably stable over years of disuse.
What Goes First
If you stop using Japanese, the first casualties are:
Lexical access — word retrieval
This is the frustrating one. You know you know a word. It used to come automatically. Now you’re staring at a sentence you can construct perfectly in terms of grammar, but you can’t retrieve the specific vocabulary item you need. The word is there — it just takes longer, or doesn’t surface at all.
Research consistently shows that lexical access fades faster than either grammatical knowledge or phonological knowledge. Words you haven’t recently activated slip below the retrieval threshold. This is why returning learners often describe their Japanese as “greyed out” rather than gone — the underlying knowledge is present but access is slow and effortful.
Phonological production — accent and pronunciation
After an extended gap, your speaking pronunciation may drift — toward English phonology, or simply toward a less precise version of the sounds you had acquired. Pitch accent, which requires continuous practice to produce naturally, is particularly vulnerable because it receives almost no corrective feedback in most learners’ practice environments even when they’re active.
Speed and automaticity
Fluency — the ability to produce language quickly and without effort — depends on automatized processes, not just knowledge. Skills automatize through repetition and then de-automatize through disuse. A fluent intermediate speaker who takes a year off may still know what they want to say; they just can’t say it quickly anymore. The knowledge is intact; the motor and processing automaticity has degraded.
What Stays
Here’s the reassuring part:
Core morphosyntax tends to survive. Research on L1 attrition found, counterintuitively, that the grammatical core of a language is more robust to attrition than vocabulary and phonology. For L2 learners, this appears to hold as well: the rules underlying Japanese sentence structure, the particle system (at a conceptual level, if not always at production speed), and core grammatical patterns tend to persist. You don’t “unlearn” grammar the way you lose lexical access.
Reading comprehension can remain functional long after speaking deteriorates. The demanding processing of spoken language (real-time, with no opportunity to re-read) degrades faster than the less time-pressured task of reading. Many returnees find their reading comprehension is much better than their speaking after a gap.
Kanji recognition. Character recognition is treated somewhat like vocabulary: kanji you’ve seen thousands of times remain relatively accessible, while lower-frequency characters become harder to retrieve. But the overall reading ability (pattern recognition at the character level) shows more stability than many learners expect.
Implicit phonological knowledge. Even if you can’t produce Japanese sounds as accurately as you could at your peak, your categorical perception of Japanese phonological contrasts — your ability to hear the difference between certain sounds — tends to persist. This is important because it means the input side of language is more stable than the output side.
The Savings Effect
There’s a phenomenon in language research called the savings effect: relearning an attrited language is consistently faster than the original learning. This is true even when the learner feels like they’ve “forgotten everything.” The underlying representations aren’t gone — they’re not fully accessible at surface level, but they’re there.
In practice, this shows up as the “it’s coming back” experience most returning learners report within days or weeks of re-engaging: kanji that felt gone come back; sentence patterns feel natural again; vocabulary surfaces faster than it did the first time around. The speed of re-activation is predictive of how much was actually retained in implicit memory even during the gap.
The savings effect is why experienced language learners often give this advice: if you’re going to stop, the important thing is to restart. The reactivation speed will tell you more accurately than your gap anxiety what you’ve actually retained.
What Predicts How Much You Lose
Not all gaps are equal. Research on attrition suggests several moderating factors:
Proficiency at the time of the gap. Higher proficiency before stopping predicts more robust retention. This is the threshold hypothesis: once a language is consolidated past a certain depth, it becomes much more resistant to attrition. Beginners lose ground faster; advanced learners retain more over long breaks.
Duration of the gap. Obviously, a two-month gap produces less attrition than a three-year gap. The decay curve is not linear — loss tends to accelerate initially (the first weeks of disuse) then slow as the “surviving” knowledge base stabilizes.
Passive contact during the gap. Even minimal passive exposure — occasional Japanese media, reading without active study — significantly reduces attrition. You don’t need to be studying; you need to be in contact with the language in some form.
Attitudes and identity. Schmid’s L1 attrition research finds that speakers who strongly identify with their L1 culture show less attrition even in extended emigrant contexts. For L2 learners, this probably translates to ongoing motivation and connection to the language community — a reason to maintain identity-level investment even without formal study.
The Japanese-Specific Pattern
Japanese presents a particular attrition profile because of its complex script system. Learners who invested heavily in kanji study find that kanji recognition attrition follows vocabulary: high-frequency characters (common joyo kanji encountered constantly) are mostly retained; lower-frequency characters fade. This can be disorienting because the degradation is visible and concrete — you’re looking at a character you knew and can’t read it.
Pitch accent, as noted, is among the most fragile components — because even active learners rarely get reliable corrective feedback on it. If you weren’t getting feedback on pitch during active study, you’re unlikely to maintain it during a gap.
Listening comprehension sometimes improves with casual exposure during a gap (anime, podcasts, YouTube) even when production is declining. This is consistent with the general pattern that receptive skills are more stable than productive ones.
What to Do About It
Maintain some contact. Even passive exposure — occasional reading, watching without the intention to study — reduces attrition significantly compared to complete absence. You don’t need study sessions; you need exposure.
Prioritize reactivation over distress. The “everything is gone” feeling is real at the retrieval-access level but misleading about the actual state of the underlying knowledge. The savings effect means time spent re-engaging with the language is very well spent. Returning to Anki, reading a chapter of known material, or having a conversation — these reactivate fast.
Don’t rebuild from scratch. A common mistake after a long gap is starting from square one: going back to absolute-beginner materials because things feel “lost.” In most cases, returnees are better served by re-engaging at (or near) their previous level, even if it’s frustrating, than by re-doing beginner content they actually still know.
Vocabulary/lexical access responds quickly to massed review. A week of intensive SRS review of previously-learned vocabulary produces rapid reactivation. If you were using Anki or another SRS system before your gap, resuming reviews — even immediately after a long gap — typically produces faster gains than your initial learning rate.
The Bottom Line
Language attrition is real. A six-month gap without Japanese contact will cost you lexical access speed, some phonological precision, and processing automaticity. It will probably not cost you your grammatical understanding, your reading ability at familiar difficulty levels, or the underlying language knowledge you built.
The research picture is actually more optimistic than the “reset” feeling suggests: what feels like forgetting is often just retrieval lag, and it reactivates faster than it was originally learned. Returning is worth it.
Related: Language Attrition — the SLA concept explained in depth. Fossilization — the related question of permanent stagnation.