Study guide
3.3 Fill in the Blanks: Word Completion, Context, & Part-of-Speech Clues
Duolingo English Test DET
Track your progress as you read
Finished sections get a check mark.
1 Chapter 1: Introduction, Architecture, & Security
1.1 Test Overview, Format, & Logistics 1.2 Computer-Adaptive Engine & IRT Scoring Mechanics 1.3 Security Protocol, Monitoring, & Proctoring Rules
2 Chapter 2: Scoring Scale & Subscore Framework
2.1 Overall 10–160 Scale & CEFR Benchmarking 2.2 The Four Individual Subscores 2.3 The Four Integrated Subscores
3 Chapter 3: Lexical Recognition & Word-Level Tasks
3.1 Read and Select: Real vs. Pseudo-Word Identification 3.2 Read and Select: Academic Roots, Affixes, & False Cognates 3.3 Fill in the Blanks: Word Completion, Context, & Part-of-Speech Clues
4 Chapter 4: Passage Completion & Reading Reconstruction
4.1 Read and Complete: C-Test Mechanics & Word-Stem Completion 4.2 Read and Complete: Advanced Syntactic & Discourse Clues 4.3 Passage Flow: Reference Words, Cohesive Markers, & Semantic Continuity
5 Chapter 5: Interactive Reading Mastery
5.1 Interactive Reading: Structure, Timing, & Passage Types 5.2 Complete the Sentences & Complete the Passage 5.3 Highlight the Answer, Identify the Idea, & Title Selection
6 Chapter 6: Listening Dictation & Interactive Listening
6.1 Listen and Type: Audio Dictation, Phonology, & Mechanics 6.2 Interactive Listening: Listen and Complete & the 6:30 Shared Clock 6.3 Interactive Listening: Listen and Respond Dialogue Navigation 6.4 Interactive Listening: Summarize the Conversation
7 Chapter 7: Speaking Production & Oral Fluency
7.1 Speak About the Photo: Visual Description & Framing 7.2 Read Then Speak: Structured Argumentation & Timing 7.3 Interactive Speaking: Real-Time Conversation Under a 35-Second Clock 7.4 The Speaking Sample: Graded Long-Form Speaking That Institutions Also Watch
8 Chapter 8: Writing Production & Test-Day Operations
8.1 Write About the Photo: Rich Sentence Structure & Accuracy 8.2 Interactive Writing: Independent Argument & Reactive Follow-Up 8.3 The Writing Sample: Extended Persuasive Response for Admissions 8.4 Test-Day Setup, Technical Readiness, & Pacing Strategy
3.3 Fill in the Blanks: Word Completion, Context, & Part-of-Speech Clues
Key Takeaways Explain This Topic AI
Fill in the Blanks asks you to 'complete the sentence with the correct word', presenting a sentence containing exactly one unfinished word whose ending you must type.
Duolingo's official page gives the task a duration of 0:20, and the frequency table lists 6-9 of these items per test.
Because the opening letters are supplied, the visible stem plus the number of empty boxes fixes the answer's exact length, eliminating candidate words before meaning is considered.
Syntactic position tells you the part of speech and the required inflection: third-person -s, past -ed, plural -s, or an adjectival or nominal suffix.
Fill in the Blanks feeds Reading, so a weak showing here also drags Literacy and Comprehension; Duolingo added the task on 2 April 2024, and it is distinct from the passage-length Read and Complete.
Last updated: September 2026
3.3 Fill in the Blanks: Word Completion, Context, & Part-of-Speech Clues
Quick Answer: Fill in the Blanks gives you a sentence containing one unfinished word and asks you to type the missing letters. Duolingo's official page gives the duration as 0:20 and the frequency table lists 6–9 of these items per test. The opening letters are always supplied, so the task is far more constrained than an open cloze: the visible stem plus the number of empty boxes fixes the word's length exactly. Run a fixed routine — (1) count the boxes, (2) identify the syntactic slot, (3) check collocations and context, (4) verify the inflection — and type.
Anatomy of the Fill in the Blanks Task
The Fill in the Blanks task was added to the DET on 2 April 2024, and it is routinely misdescribed in prep material as an open cloze where you invent a whole word. It is not. Duolingo's own wording is precise:
"Fill in the Blanks asks you to 'Complete the sentence with the correct word.' In this question, you will read a sentence with one unfinished word, then fill in the blanks to complete the word."
In other words, it is a single-word C-test : the beginning of one word is given, the rest is replaced by letter boxes, and you type the missing letters. +-----------------------------------------------------------------------+
- | Complete the sentence with the correct word. |
- | 0:20 |
- | |
- | The committee's final report was primarily conc[ ][ ][ ][ ][ ] |
- | with the economic ramifications of urban expansion. |
- | |
- | [ CONTINUE ] |
- +-----------------------------------------------------------------------+
Work that example in the right order. The stem conc is 4 letters and there are 5 boxes, so the answer is exactly 9 letters . The intact preposition with immediately afterwards narrows the frame to concerned with . And concerned is 9 letters. Stem, count, and collocation all agree — type erned .
Now notice what the count did for you. Without it, conclusive , concurrent , and concentrating are all words a rushed reader might reach for. All three are the wrong length, and the box count eliminates them in under two seconds without any semantic reasoning at all.
How This Differs From Read and Complete
- Fill in the Blanks Read and Complete
- Scope One sentence A short passage of several sentences
- Damaged words Exactly one Typically 6–10
- Time 0:20 3:00
- Frequency 6–9 items 3–6 items
- Context available The single sentence only The whole passage, including intact first and last sentences
That last row is the strategic difference. In Read and Complete you can rescue a stubborn word by reading the intact final sentence. In Fill in the Blanks there is nothing else on screen. All of your evidence is the stem, the box count, and one sentence — which is exactly why the arithmetic carries so much of the load.
Psychometric Behavior
Duolingo's frequency table lists 6–9 Fill in the Blanks items per test, each with its own 20-second timer . When the timer expires the item is submitted as-is and the test advances. Because the test is adaptive, accurate answers pull subsequent sentences toward denser syntax and lower-frequency academic vocabulary. The task feeds Reading , and therefore Literacy and Comprehension .
The 20-Second Countdown Protocol
To consistently solve items within the 20-second window without rushing or freezing, follow this four-phase operational timetable: Seconds 00-03 ===> Phase 0: COUNT THE BOXES. Stem length + boxes = total letters.
- Seconds 03-07 ===> Phase 1: Syntactic Parsing (isolate the required part of speech)
- Seconds 07-12 ===> Phase 2: Collocation & dependent-preposition scan
- Seconds 12-16 ===> Phase 3: Semantic logic & context matching
- Seconds 16-20 ===> Phase 4: Inflectional verification & keyboard entry
Important
Phase 0 is not padding. The letter count is the only hard constraint in the item and the one piece of evidence that cannot be argued with. Candidates who jump straight to meaning routinely type a semantically perfect word of the wrong length and score nothing on an item they understood completely.
Phase 1: Syntactic Parsing (Seconds 3–7)
With the length fixed, determine what grammatical slot the unfinished word occupies. Identify the sentence's subject, finite verb, direct object, and prepositional phrases.
Use the following Positional Clue Matrix to pinpoint the required part of speech instantly. In each row, [WORD] marks the slot holding the unfinished word:
Syntactic Position Formula Required Part of Speech Diagnostic Example (stem + boxes) Target Word Category
Determiner + [WORD] + Noun Attributive Adjective "An unpreced[ ][ ][ ][ ][ ] increase in demand..." → unprecedented Descriptive Adjective
Subject + [WORD] + Direct Object Transitive Verb "The researchers invest[ ][ ][ ][ ][ ] the anomaly." → investigated Action/Analytical Verb
Determiner + Adjective + [WORD] + of Head Noun "A remarkable synth[ ][ ][ ][ ] of two theories..." → synthesis Abstract Academic Noun
Verb + Direct Object + [WORD] Adverb of Manner "The software executed queries flawl[ ][ ][ ][ ][ ]." → flawlessly Modifying Adverb ( -ly )
Verb/Adjective + [WORD] + Noun Phrase Adjective + fixed preposition "The findings were consis[ ][ ][ ][ ] with earlier studies." → consistent Collocation-anchored Adjective
Independent Clause + [WORD] + Clause Subordinating Conjunction "The theory failed beca[ ][ ][ ] evidence was absent." → because Logical Connector
Note
Because the item always supplies the beginning of the word, a pure function word such as a bare preposition is rarely the whole answer — the damaged element is usually a content word or a multi-letter connector. This is a genuine difference from Read and Complete, where short function words are frequently the damaged items.
Phase 2: Collocations and Dependent Prepositions
In academic English, words do not combine randomly. Verbs, nouns, and adjectives govern specific dependent prepositions and form predictable lexical pairings known as collocations .
In many Fill in the Blanks items an intact preposition sits immediately after the damaged word, and that preposition identifies the lemma almost on its own. Memorising high-frequency academic pairings turns these items into instant pattern recognition — and because you already know the target length from the box count, the pairing plus the count usually settles the answer outright:
Essential Academic Prepositional Pairings
Governed by "to" :
- attribute [something] to / ascribe to
- adjacent to / subsequent to / prior to
- adhere to / conform to
- contribute to / lead to
Governed by "on" / "upon" :
- rely on / depend on
- contingent on / premised on
- impact on / reflect on
Governed by "with" :
- coincide with / collaborate with
- compatible with / consistent with
- interfere with / comply with
Governed by "from" :
- prohibit from / deter from
- derive from / stem from
- distinguish from / deviate from
Governed by "of" :
- capable of / conscious of
- deprived of / indicative of
- comprise (takes NO preposition in active voice!) vs. composed of
- Phase 3 & 4: Semantic Logic and Inflectional Verification
Once syntax and collocations have narrowed down the candidate pool, use the sentence's semantic logic to choose the exact word, and then verify its inflectional morphology .
Decoding Discourse Signals
Sentence clauses are linked by semantic signposts that tell you whether the missing idea continues or contrasts with the surrounding context:
Adversative Signals ( although, whereas, despite, however, conversely ) :
The damaged word must carry the opposite semantic polarity to the surrounding clause.
Example : "Although initial trials proved disapp[ ][ ][ ][ ][ ][ ], subsequent iterations yielded promising data." → disappointing
Causal / Resultative Signals ( therefore, consequently, leading to, because, thereby ) :
The damaged word must describe a logical consequence.
Example : "The sudden withdrawal of subsidies caused severe inflation, for[ ][ ][ ][ ] small enterprises into bankruptcy." → forcing
The Final Inflectional Checklist (Phase 4)
Before pressing submit, execute this 3-point grammatical check:
Subject-Verb Agreement : If the damaged word is a present-tense verb, does the subject take a third-person singular -s ? Watch for intervening prepositional phrases: "The collection of rare artifacts exhib[ ][ ][ ] signs of decay." → exhibits (the head noun is the singular collection , not the plural artifacts ).
Tense and Aspect Consistency : Does the sentence describe a completed past action ( simple past -ed ), a continuing state ( present participle -ing ), or an action completed prior to another ( past perfect had been -ed )?
Nominal Plurality : If the word is a countable noun preceded by a plural quantifier ( several , numerous , diverse ), did you type the plural ending ( -s or -es )?
Cognitive Walkthrough: 3 Worked Test-Day Examples
Study how an experienced test taker processes three representative Fill in the Blanks items inside the 20-second limit. Watch the box count do most of the work:
Worked Example 1: Collocation Plus Box Count
Prompt: "The regional environmental council attrib[ ][ ][ ][ ][ ] the sudden resurgence of native bird populations to strict new habitat protection laws."
Phase 0 (Count) : Stem attrib (6) + 5 boxes = 11 letters .
Phase 1 (Syntax) : Subject ( council ) + [WORD] + direct object ( the resurgence... ) + to + noun phrase. The slot is a finite transitive verb.
Phase 2 (Collocation) : The intact to later in the sentence confirms the frame attribute [effect] to [cause] .
Phase 3 (Semantics) : The council is reporting a completed finding, so past tense fits.
Phase 4 (Inflection + Count) : attribute = 9 letters, too short. attributes = 10, still short. attributed = 11 ✓. Type: uted .
Elapsed Time : about 9 seconds.
Worked Example 2: Adversative Pivot
Prompt: "Although the initial prototype performed poorly during preliminary trials, subsequent modifications made the engine remarkably resil[ ][ ][ ][ ] under extreme conditions."
Phase 0 (Count) : Stem resil (5) + 4 boxes = 9 letters .
Phase 1 (Syntax) : made + object ( the engine ) + adverb ( remarkably ) + [WORD]. The slot is a predicate adjective.
Phase 2 (Collocation) : remarkably [adjective] under extreme conditions .
Phase 3 (Semantics) : Although ... performed poorly signals a reversal, so the adjective must be positive.
Phase 4 (Inflection + Count) : resilience is a noun and 10 letters; resilient is the adjective and exactly 9 ✓. Type: ient .
Elapsed Time : about 11 seconds.
Worked Example 3: Agreement Trap
Prompt: "Neither the principal investigator nor any of her laboratory assistants w[ ][ ][ ] able to replicate the anomalous experimental results."
Phase 0 (Count) : Stem w (1) + 3 boxes = 4 letters . This alone decides the item.
Phase 1 (Syntax) : Correlative subject ( Neither X nor Y ) + [WORD] + able to replicate . The slot is a finite past-tense form of to be .
Phase 3 (Grammar) : With neither ... nor , the verb agrees with the nearer subject — any of her laboratory assistants , which is plural.
Phase 4 (Verification) : was is 3 letters and would leave a box empty; were is 4 ✓. Type: ere .
Elapsed Time : about 8 seconds.
Tip
Notice what happened in all three examples: the box count and the grammar independently pointed at the same answer. When they disagree, trust the count and re-examine your grammar, not the other way round. The count is objective; your reading of the sentence may not be.