Skip to main content

AJATT (All Japanese All The Time): The Original Immersion Movement

AJATT (All Japanese All The Time) is an immersion-first self-study philosophy for Japanese. The core idea is to surround yourself with Japanese audio, video, and text for as much of every day as possible, and to acquire the language through use and exposure rather than classroom instruction.12 It matters because it seeded much of the immersion-and-sentence-mining discourse that English-speaking self-learners still use today.32

Overview​

What AJATT is​

AJATT is framed as a method-as-lifestyle, not a fixed curriculum. The site's tagline states the orientation: "You don't know a language, you live it. You don't learn a language, you get used to it."1

The original presentation explicitly rejects grammar-rule study as the main structure for learning. In its own words, "learning grammar rules in order to use a language is like learning quantum physics in order to drive a car."1

AJATT was created and written by an author who published under the single name Khatzumoto. The blog operated under that pseudonym, and this article says nothing about the person beyond what he published.14

Method-as-lifestyle, not a curriculum

AJATT prescribes a way of arranging your day around Japanese, not a fixed syllabus to complete chapter by chapter. The "all the time" in the name is the literal program, not a slogan.1

Origin: Khatzumoto and the ~2006 blog​

Khatzumoto created the AJATT website in 2006 as a personal blog. It was aimed at self-learners (anime fans, hobbyists, "geeks") who wanted to learn Japanese without classes, textbooks as the main structure, or paid courses.32

The method was openly influenced by the Antimoon self-study site and by Stephen Krashen's input-based theory of second-language acquisition. Khatzumoto aimed to spend every free moment listening to or looking at Japanese, even when he did not understand it.25

Khatzumoto self-reports that he started learning Japanese at age 21 and reached near-native ability within roughly 18 months. This is his own published account, so it should be read as a self-report rather than an independently verified outcome.4 In his words: "In a word, my fluency was near-native. When I spoke on the phone, Japanese people assumed I was Japanese."4

At the 18-month mark, he lists, among other things, conducting "a job interview 100% in Japanese," writing "business and personal emails," reading "IT/physical science/computer science expert document[s] (manuals, software docs, academic papers, even legal documents)," and writing "4500 kanji from memory, 90% retention."4

The 18-month figure is one person's self-report

The roughly 18-month timeline is Khatzumoto's description of his own result, on his own blog. It is not an experimentally verified or peer-reviewed outcome, and it should not be read as a typical or promised result.4

Stewardship of the AJATT name later passed on. Khatzumoto stepped back around 2023, after which a successor community (Tatsumoto) continued to publish under the AJATT banner.3 The "~2006" origin is well attested by both that successor community and the secondary wiki summary, so this article treats it as a dateable historical fact.32

The AJATT method, in sequence​

The original Khatzumoto presentation lays out an ordered workflow. The main moves are kanji first, then mass immersion, then sentence mining into a spaced-repetition system (SRS), all wrapped in lifestyle principles.1

The sequence is the heart of the method. Its order is the part most often lost when people use "AJATT" loosely. The diagram below captures that ordered pipeline.

Phase 1: Kanji first via Remembering the Kanji​

The original method starts with kanji before grammar and before learning to read ordinary text. The learner works through James Heisig's Remembering the Kanji, learning to write each character and recall one English keyword. The learner explicitly does not learn the readings at this stage.16

The target in the original overview is the roughly 2,046 general-use (jōyō) kanji via Remembering the Kanji, loaded into a spaced-repetition system. It frames learning the kana, the two basic Japanese syllabaries, as trivial once the kanji are done.1

Remembering the Kanji 1 (James W. Heisig, University of Hawai'i Press, first published 1977) covers about 2,200 kanji and deliberately defers all readings to Volume 2. Volume 1 teaches only the writing and a single keyword meaning per character.67 This meaning-before-reading design is exactly what AJATT front-loads.1

The original guidance is to push through the kanji phase without pausing or starting grammar study on the side.1

Phase 2: Maximize immersion (the "all the time" core)​

The defining AJATT move is to fill every available moment with Japanese input: audio, video, music, and text. This includes long stretches of low-comprehension passive listening ("whether he understood it or not").2

This rests on Krashen's Input Hypothesis: the idea that language is acquired through exposure to comprehensible input ("i+1"), and that minimizing stress and maximizing enjoyment lowers what Krashen calls the "affective filter."5 Comprehensible input means input just slightly above the learner's current level, close enough to be understood with effort.

The "all the time" branding is the literal program. The aim is to spend as many waking hours as possible in contact with Japanese, on the premise that volume of exposure drives acquisition.12

Phase 3: Sentence mining into an SRS​

After the kanji foundation, the learner moves to the 10,000-sentence phase: harvesting whole, real sentences from authentic Japanese material and reviewing them in a spaced-repetition system. The overview recommends an SRS such as Mnemosyne. Anki is the tool most associated with the practice in later immersion communities.18

The stated rationale is that a sentence is "a set of words arranged according to grammar rules with the added benefit of showing the 'sense' in which to use the words." Sentences teach usage and natural collocation, or words that naturally appear together, in a way that isolated word lists do not.8

The original instructions are explicit about the unit of study: "Do not: learn individual words. Learn sentences," and "Do not: translate sentences. Understand them instead."1 This whole-sentence, SRS-based harvesting is what the wider community later standardized as "sentence mining."

The lifestyle principles around the method​

These are Khatzumoto's stated principles. They are reported here as his framing, not as a J-Compass endorsement.

  • Make it fun. Choose immersion material you actually enjoy (anime, games, comics, TV), because enjoyment sustains the volume of input the method needs.2
  • No formal classes; textbooks are not the main structure. The method is built around self-directed immersion rather than coursework, and it explicitly downplays grammar-rule study.12
  • Do it every day; do not pause. Consistency, especially through the kanji phase, is treated as non-negotiable.1
  • Belief as a starting condition. The overview opens with a belief step, while noting that "believing in yourself is essential, but by itself it obviously won't get you anywhere."1

Cultural impact and legacy​

What AJATT popularized​

AJATT was one of the most influential early-internet sources pushing immersion-first self-study, sentence mining into an SRS, and RTK-first kanji (starting with Remembering the Kanji) into mainstream Japanese-learning discourse among English-speaking self-learners.32

It functioned as a bridge between Krashen's academic input theory and a practical, internet-native daily workflow. It popularized the idea that filling your environment with the target language could substitute for classroom study.25

Successors: MIA and Refold​

AJATT directly influenced later immersion frameworks. Matt Colwell ("Matt vs Japan"), an AJATT practitioner, later co-founded the Mass Immersion Approach (MIA). MIA preserved AJATT's input-and-mining core while adding systematic pitch-accent study and more structured intermediate and advanced guidance.2

MIA was later rebranded as Refold, a language-agnostic immersion framework. The lineage commonly described is AJATT → MIA → Refold.2

MIA and Refold are treated here as lineage and concepts, not as J-Compass articles, except where the site index already links to them.

Criticisms and honest trade-offs​

The comprehensible-input objection​

AJATT's early "listen even if you understand nothing" stage draws criticism from mainstream SLA. Krashen's own hypothesis requires input to be comprehensible (i+1). Large quantities of input far above the learner's level (i+many) are not what the theory predicts will drive acquisition. On Krashen's own terms, very-low-comprehension passive listening is therefore weakly supported.59

The broader field accepts comprehensible input as necessary. But it rejects the strong claim that explicit study and output play no role. Immersion-only is not the consensus position.9

A more recent peer-reviewed critique argues that acquisition is active and interactive rather than passive-absorptive. It also argues that a one-size i+1 prescription ignores differences between learners.10

The "hidden prerequisites" objection​

Many immersion-only success reports run alongside unacknowledged scaffolding, or supports that are not fully named: a grammar reference consulted in parallel, prior exposure, or unusually high daily time budgets. "Immersion alone" can therefore understate the supports actually used.9

This is a community-debate observation, not a single finding tied to AJATT by name. It applies to immersion-only narratives generally, not to any one practitioner.

RTK first teaches no readings

Completing Remembering the Kanji teaches keywords and writing but no readings, so a learner can finish hundreds of hours of RTK and still be unable to read ordinary Japanese aloud. Critics question whether meaning-first kanji study is the most efficient foundation. Defenders note that it clears a major barrier before immersion begins.67

Sustainability, intensity, and who it fits​

The "all day, every day" intensity suits high-time-budget learners (students, people on sabbaticals, and highly self-directed learners). It fits poorly with 5-hours-a-week, deadline-driven learners.12

The main variables are available daily hours, tolerance for ambiguity during low-comprehension input, self-direction, and access to enjoyable native material. The method's results scale with the volume of input the learner can sustain.4

That scaling is exactly what makes the 18-month self-report non-transferable as a promise. Per the category's editorial rule, the honest framing is to recommend AJATT for learners it fits, caution those it does not fit, and make no "fluent in X months" claim.4

Good to know​

"AJATT" the brand vs. immersion the general practice​

AJATT is a specific named blog and a specific prescriptive workflow: RTK first, then mass immersion, then 10,000 sentences. Immersion is the general practice of learning through target-language exposure, and it long predates AJATT.12

Many people loosely say "I'm doing AJATT" when they mean "I'm immersing." Keep the brand and the practice distinct; not every immersion routine is AJATT.12

Treating the 18-month figure as a generalizable promise​

The "fluent in 18 months" figure is Khatzumoto's self-report of his own result, not a verified or typical outcome. He published it himself and hedged it, calling his fluency "near-native" and listing things he still could not do at 18 months.4

Stating the figure as a generalizable promise misrepresents the source. Cite it as a self-report each time it appears.4

AJATT predates and is distinct from MIA and Refold​

AJATT (2006) came first. MIA and then Refold are later descendants that revised and systematized it.32 People use the names loosely, but the chronology is fixed: AJATT is the origin, not a synonym for its successors.32

Staying within what is publicly documented about Khatzumoto​

The safe claims are the dateable facts (the 2006 origin, the RTK-first → immersion → sentence-mining sequence, the 2023 hand-off) and the labeled self-report (18 months).43 Speculation about Khatzumoto's identity or biography is out of scope and unsourced, so this article makes no claim about the person beyond what he published.43

See also​

References​

Footnotes​

  1. Khatzumoto. "The Method: An Overview." AJATT | All Japanese All The Time. https://alljapanesealltheti.me/overview/index.html (primary-source mirror of the original alljapaneseallthetime.com / alljapaneseallthetimee site). (limitation: practitioner blog; used only as the primary self-description of the method, not as a verified outcome.) ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9 ↩10 ↩11 ↩12 ↩13 ↩14 ↩15 ↩16 ↩17 ↩18 ↩19

  2. "All Japanese All The Time." Learn Any Language Wiki (Fandom). https://learnanylanguage.fandom.com/wiki/All_Japanese_All_The_Time (secondary summary; corroborates the 2006 origin, the Antimoon/Krashen influence, and the 18-month claim). (limitation: open wiki; used only where it corroborates a primary source.) ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9 ↩10 ↩11 ↩12 ↩13 ↩14 ↩15 ↩16 ↩17 ↩18

  3. Tatsumoto. "What's AJATT?" AJATT (successor community). https://tatsumoto-ren.github.io/blog/whats-ajatt.html (primary source for the 2006 origin date, the 2023 hand-off, and the successor community's restatement of the method). (limitation: prescriptive successor community, not a neutral history.) ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9

  4. Khatzumoto. "Pure Pwnage: How Fluent Was I After 18 Months?" AJATT | All Japanese All The Time. https://alljapanesealltheti.me/pure-pwnage-how-fluent-was-i-after-18-months/index.html (primary self-report; quoted as a self-report, not as an independently verified outcome). ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9 ↩10 ↩11

  5. Krashen, Stephen D. Principles and Practice in Second Language Acquisition. Pergamon Press, 1982. (The Input Hypothesis / "comprehensible input" / i+1 framework that AJATT's immersion phase draws on.) ↩ ↩2 ↩3 ↩4

  6. Heisig, James W. Remembering the Kanji 1: A Complete Course on How Not to Forget the Meaning and Writing of Japanese Characters. University of Hawai'i Press (first published 1977; multiple later editions). Volume 1 teaches the writing and a single English keyword/meaning for roughly 2,200 characters and deliberately defers all readings to Volume 2. ↩ ↩2 ↩3

  7. "Remembering the Kanji." Wikipedia. https://en.wikipedia.org/wiki/Remembering_the_Kanji (used for the publication facts and the meaning-before-readings structure of RTK; corroborates 6). ↩ ↩2

  8. Khatzumoto. "10,000 Sentences: Why." AJATT | All Japanese All The Time. https://alljapanesealltheti.me/10000-sentences-why/index.html (primary source for the sentence-method rationale). ↩ ↩2

  9. VanPatten, Bill, and Jessica Williams (eds.). Theories in Second Language Acquisition: An Introduction. Routledge (2nd ed., 2015), chapter on the Input Hypothesis. (Mainstream SLA appraisal: comprehensible input is widely accepted as necessary, but the strong claim that explicit learning and output play no role is not supported by the field.) ↩ ↩2 ↩3

  10. "Beyond comprehensible input: a neuro-ecological critique of Krashen's hypothesis in language education." Frontiers in Psychology, 2025. https://www.frontiersin.org/journals/psychology/articles/10.3389/fpsyg.2025.1636777/full (peer-reviewed critique that acquisition is active and interactive rather than purely passive-absorptive, and that one-size i+1 input ignores individual learner pathways). ↩