Glossary of language-learning terms

This glossary collects the language-learning vocabulary FlixFluent uses across its product and documentation. Most entries are short definitions plus a one-paragraph elaboration covering the term in practice. Use it as a reference when reading the comparison or learning guides — every term linked from elsewhere on this site has its definition here.

At a glance

  • Definitions for the language-learning vocabulary FlixFluent uses
  • Cross-linked from compare, learn, and guide pages
  • Includes technique terms, linguistic terms, and assessment terms
  • Reference for FSI categories and CEFR levels

How to use this glossary

Each entry stands alone. Read top-to-bottom for orientation, or jump to a term using the table of contents below.

Terms are organised loosely from technique terms (dual subtitles, comprehensible input) to linguistic terms (particle, dictionary form) to assessment terms (FSI categories, CEFR).

Definitions

Dual subtitles

Two subtitle tracks rendered simultaneously — one in your target language, one in your base (native) language. Also called bilingual subtitles, double subtitles, or two subtitles at once.

Used for comprehension support: the base track answers what was said, the target track shows how it was said. FlixFluent renders dual subtitles on Netflix and YouTube via the Chrome extension, using the platform's own track for the base line when your language is available and in-context AI translation when it is not. Full detail: /dual-subtitles/.

Auto-generated captions (ASR)

Captions produced by automatic speech recognition rather than written by a human — YouTube's "auto-generated" tracks are the common example.

ASR captions contain recognition errors and arrive as rolling word fragments rather than sentences, so they are weaker learning material than human-written tracks. FlixFluent detects ASR and auto-translated tracks on YouTube from the player's own metadata, flags them with a visible warning (a banner plus a persistent icon on the subtitle overlay), and regroups the fragments into full sentence cues before display. Netflix has no ASR captions — its tracks are professionally authored.

In-context translation

A translation generated for a word as it is used in one specific sentence, rather than retrieved from a dictionary entry.

The same word can mean different things in different lines, so FlixFluent generates every word translation per word + exact sentence + language pair, passing the surrounding lines as extra context, and caches it for that sentence. A dictionary lookup returns the same gloss everywhere; an in-context translation returns the meaning in play. FlixFluent never takes meanings from dictionary entries; dictionaries are used only to split text into words and show readings. Full detail: /how-it-works/.

Hover translation

A translation that appears when you move the mouse over a word, without stopping the video or leaving the page.

In FlixFluent, hovering a subtitle word shows its translation for that sentence; clicking goes deeper into a full explanation. The same works on ordinary web pages in the any-website mode (beta, /any-website/).

Comprehensible input

Linguistic input slightly above your current ability, where the new vocabulary or grammar is supported by enough familiar context that you can roughly understand it (Stephen Krashen, "i+1").

The core thesis: acquisition happens when you understand most of what you hear or read, with manageable surprise. FlixFluent's tools (hover translation, deconstruction) lower the threshold of comprehensibility for material that would otherwise be too hard, and its catalogue rates videos A1–C2 so you can pick input near your level (/catalogue/).

Lexical coverage

The percentage of the words in a text or video that a learner already knows.

Coverage is a common way to estimate whether material is followable. In an analysis of 88 television programmes, knowing the most frequent 3,000 word families plus proper nouns and marginal words covered 95.45% of the words, and 7,000 word families covered 98.27% (Webb & Rodgers, 2009, Language Learning). FlixFluent's catalogue level filter aims at titles near your tested level for the same reason.

Sentence mining

The practice of harvesting sentences from native content and turning them into spaced-repetition flashcards.

A typical mining loop is: encounter a sentence with one unknown word (i+1), save it as a flashcard with audio + text + translation, review with SRS over time. FlixFluent automates the capture half of this loop: every word you look up while watching is collected into the vocabulary trainer with its original sentence and an episode reference, and productive cards test you with a cloze built from that very sentence. Collections can also be imported from Anki (.apkg) or CSV, and exported as CSV.

Romanization

Writing a non-Latin script in Latin letters — for example Hangugeo for 한국어, sushi for 寿司.

Romanization is a starter scaffold for learners who have not yet learned the target script. FlixFluent ships romanization for Korean (Revised Romanisation), Japanese (rōmaji, with kanji readings resolved per line), Chinese (pinyin with tone marks), Thai, Greek, Cyrillic, and Hebrew when vowel points are present, with a general transliterator for other scripts. Drop it as soon as you have the script.

Furigana

Small kana printed above or beside kanji to show how they are read, for example た over 食 in 食べる (taberu, to eat).

Many kanji have several readings, so the right one depends on the word and the sentence. FlixFluent resolves Japanese kanji readings per line with morphological analysis and shows them as rōmaji in the romanization line and in explanations, rather than as kana above the kanji.

Transliteration

Converting text from one script to another letter by letter or sound by sound, such as writing Russian Cyrillic or Hebrew in Latin letters.

Romanization is transliteration into the Latin alphabet. It only works reliably when the source shows the sounds: unvocalised Hebrew omits most vowels, so FlixFluent romanizes Hebrew only when the text carries vowel points.

Right-to-left (RTL) script

A writing system read from right to left, such as Arabic and Hebrew.

RTL text needs its own layout: lines run right to left and mixed-direction text has to keep each part in order. FlixFluent shows Arabic and Hebrew subtitles right-to-left and keeps Arabic or Hebrew words in the correct direction inside explanations written in a left-to-right language. Its breakdowns also split the prefixes and clitics these languages attach to words from the stem.

Word segmentation

Splitting text written without spaces, such as Chinese or Japanese, into separate words.

Without segmentation, a tool can only point at single characters. FlixFluent segments Chinese with jieba and Japanese with the kuromoji morphological analyser, grouping compound words, so hovering and clicking work on whole words.

Text-to-speech (TTS)

Synthesised speech generated from text — in FlixFluent, the pronunciation audio behind every speaker button.

FlixFluent generates TTS audio with the word's sentence passed as context, so multi-reading words are pronounced the way that sentence requires — Chinese 还 as hái or huán, Japanese 辛い as karai or tsurai — instead of a default reading that can contradict the displayed pinyin or furigana. Each clip is generated once and cached; if generation is unavailable, the browser's built-in voices are used as a fallback.

SRS (spaced repetition system)

Software that schedules flashcard reviews based on how well you remember each card — easier cards reappear later, harder ones sooner.

Anki, JPDB, and Memrise are common SRS tools. FlixFluent embeds its own: the vocabulary trainer in the web app schedules cards with FSRS at a 90% retention target, fed by the words you look up while watching. See /vocabulary-trainer/ for the full mechanics.

FSRS (Free Spaced Repetition Scheduler)

An open-source spaced-repetition algorithm that models each card's difficulty, memory stability, and current retrievability, and schedules the next review for when predicted recall falls to a target retention.

FSRS is the modern successor to the SM-2 heuristic used by classic flashcard apps, and is the algorithm family modern Anki ships as its recommended scheduler. In the open srs-benchmark (roughly 10,000 anonymised Anki collections, ~350 million reviews) FSRS variants predict recall more accurately than SM-2 for the large majority of users. FlixFluent's vocabulary trainer uses FSRS with a fixed 90% retention target.

Retrieval practice (testing effect)

Strengthening memory by actively recalling an item rather than re-reading it. Delayed retention after self-testing reliably beats restudying (Roediger & Karpicke, 2006).

This is why flashcards ask before they tell. FlixFluent's trainer is recall-first at every setting: reveal-then-grade at minimum, typed production at the strictest. The optional "force me to type" mode is retrieval practice in its most demanding form. Video quizzes apply the same principle while you watch, with the answer and an explanation after each question (/video-quizzes/).

Cloze deletion

A fill-in-the-blank exercise where a word is removed from a sentence and must be recalled from context.

FlixFluent's productive cards are clozes built from the exact subtitle sentence where you originally met the word — the context you watched is the context you are tested in. "Fill in the blanks" in the video quizzes is a cloze on a line you just heard, with one to three blanks depending on your level.

Video quiz

In FlixFluent, a short question about the lines you just watched, asked in the gap after a subtitle line.

There are six exercise types (match the clips, vocabulary, fill in the blanks, build the sentence, spot the mistake, and comprehension), mixed and pitched to your CEFR level, with a translation and an explanation after each answer. Questions are saved to the Trainer page of the web app. Full detail: /video-quizzes/.

Interpolated testing

Inserting short tests into a lecture or video rather than only testing at the end.

In two experiments, tests interpolated into online lectures reduced mind wandering, increased note-taking and improved learning (Szpunar, Khan & Schacter, 2013, PNAS). FlixFluent's video quizzes apply the idea to the show you are watching.

Receptive vs productive recall

Receptive recall: see the target-language word, retrieve the meaning. Productive recall: see the meaning, produce the target-language word. Production is the harder, stronger skill.

FlixFluent's trainer starts cards receptive and, in auto mode, graduates them to productive once the receptive direction is established (several successful reviews plus stability thresholds). The direction can also be pinned to receptive-only or productive-only in settings.

Digger

FlixFluent's recursive explanation flow: inside any word or sentence explanation, each highlighted part is itself clickable and opens a deeper explanation of that part.

A back arrow retraces the chain, and there is no depth limit — sentence → word → morpheme → related form, as far as curiosity goes. Available in the extension overlay on Netflix / YouTube and in the web-app trainer.

Study time

In FlixFluent, the minutes you spend actively learning, as opposed to minutes of video watched.

Hovering words, opening explanations and breakdowns, reviewing flashcards, answering quiz questions and taking tests count, with gaps of up to four minutes bridged; watching without interacting does not. The web app shows it for 7 days, 30 days or all time (/web-app/).

Shadowing

Repeating native speech aloud immediately after hearing it, without trying to translate.

Shadowing trains the muscles of pronunciation and rhythm. FlixFluent's auto-pause + 0.5×–1.5× playback support shadowing without rewinding manually. Effective once you can comprehend roughly 70% of a passage at 1× speed.

Dictionary form

The base form of a word as it appears in a dictionary, before conjugation or inflection — e.g. Korean 가다 (go), Japanese 食べる (to eat), Spanish hablar (to speak).

FlixFluent's click-to-deconstruct view returns the dictionary form of any verb (or other inflected word) so you can look it up cleanly. Without it, learners often try to look up conjugated forms (갔어요, 食べました) and get nothing.

Particle

A small functional word, usually attached after a noun or verb, that marks its grammatical role — Korean 은/는, 이/가, 을/를; Japanese は, が, を, に, で.

Particles do most of the syntactic work in Korean and Japanese. FlixFluent labels each particle with its grammatical role on click, which is the single biggest comprehension win for learners of those languages.

Grammar role

The function a word or phrase plays in a sentence — subject, direct object, indirect object, location, time, instrument, modifier, etc.

In English, grammar role is mostly inferred from word order. In Korean, Japanese, German, and many other languages, role is marked by particles or case endings. FlixFluent surfaces the role explicitly.

FSI categories

A US Foreign Service Institute classification of how long it takes a native English speaker to reach professional working proficiency in a target language.

The FSI currently uses four categories, based on the class hours its English-speaking students need: Category I, about 24–30 weeks (600–750 hours) — for example Spanish, French, Italian, Dutch and Portuguese; Category II, about 36 weeks (900 hours) — for example German and Indonesian; Category III, about 44 weeks (1,100 hours) — for example Russian, Hindi, Turkish, Hebrew and Vietnamese; Category IV, about 88 weeks (2,200 hours) — Arabic, Chinese, Japanese and Korean. FlixFluent does not change these constants but reduces friction within them, and its study time card shows how many active hours you have put in.

CEFR

Common European Framework of Reference, a six-level proficiency scale from A1 (beginner) to C2 (mastery).

Roughly: A1 survival phrases, A2 simple conversation, B1 daily life, B2 nuanced opinions, C1 professional, C2 native-like. FlixFluent is most useful from late A2 onwards once basic grammar is in place. FlixFluent's web app measures CEFR level with placement tests for English, Korean, German, French, Spanish, Chinese, Japanese, Dutch, and Arabic (20 questions across grammar, vocabulary, and reading), rates catalogue videos A1–C2, and pitches video quizzes to your level.

Placement test

A test that places a learner on a proficiency scale, here CEFR A1–C2, rather than testing a specific lesson.

FlixFluent's placement tests have 20 multiple-choice questions on grammar, vocabulary and reading, rising from A1 to C2 and taking 5 minutes at most, with a penalty for wrong answers so guessing does not pay. Tests exist for nine languages and can be retaken once a day. The result sets the level of video quizzes and the catalogue filter (/catalogue/).

Auto-pause

A FlixFluent feature that pauses playback at the end of each subtitle line, giving you time to read, parse, or repeat.

Pairs with shadowing — pause, repeat, advance. Disable when you only want to watch.

Click-to-deconstruct

FlixFluent's feature for breaking down a clicked word or sentence into dictionary form, stem, particle, grammar role, and structural parse.

Available on every word and sentence in supported languages. Higher-leverage in morphologically rich languages (Korean, Japanese, German, Russian, Turkish).

Banmal / jondaetmal

Korean informal speech (반말) and formal speech (존댓말). The difference is socially loaded — choosing the wrong register is rude.

Watch for verb endings: -아/어 vs -아요/어요 vs -습니다/ㅂ니다. FlixFluent's deconstruction labels politeness register on every verb.

Subjunctive

A grammatical mood used to express doubt, possibility, necessity, emotion, or hypothetical situations — strong in Spanish, French, Portuguese, less so in English.

FlixFluent labels mood (indicative / subjunctive / conditional) on click for romance languages, making subjunctive triggers visible in real sentences.

Case (Nom / Akk / Dat / Gen)

A morphological marking on a noun phrase indicating its grammatical role — German Nominativ (subject), Akkusativ (direct object), Dativ (indirect object), Genitiv (possessive).

Languages with case (German, Russian, Latin, Finnish) require explicit role tracking. FlixFluent labels case on every clicked noun phrase in supported languages.

Frequently asked questions

Why is comprehensible input the dominant theory of acquisition?
Because the alternatives — pure drilling, pure conversation, pure grammar study — all show diminishing returns past a certain point. Input that you mostly understand, in volume, has the strongest empirical track record.
Do I have to learn the target script before using FlixFluent?
No. Romanization is built in for Korean, Japanese, Chinese, and others. But learn the script — it pays off for years.
How does deconstruction differ from translation?
Translation gives you what a sentence means. Deconstruction gives you why it means that — particles, mood, role, dictionary form.
Is sentence mining better than passive watching?
For long-term retention of specific vocabulary, yes. For listening fluency, passive watching also matters. Most serious learners do both.
Can I use FlixFluent at A1?
You can but the value is lower — most sentences are above your level. FlixFluent shines from late A2 onwards.

Sources & further reading

Install FlixFluent

Add FlixFluent to Chrome and turn Netflix or YouTube into an interactive language lesson. 7-day free trial, 30-day money-back guarantee.

Install from Chrome Web Store