Multiword units (MWUs) are sequences of two or more words that are stored and retrieved as pre-fabricated chunks in the mental lexicon, rather than assembled word by word through grammar rules. They range from fully fixed idioms to more flexible collocations, and they are now understood to be a fundamental organizing principle of language — not a peripheral addition to grammar. See also Multiword Unit.

Types of Multiword Units

• Idioms: meanings that cannot be derived compositionally (kick the bucket, bite the bullet)
• Collocations: words that co-occur more frequently than chance (make a decision, strong coffee)
• Formulaic sequences: fixed chunks used in specific contexts (as it were, at the end of the day)
• Phrasal verbs: verb + particle combinations with often non-transparent meanings (give up, look into)
• Binomials: fixed word pairs (trial and error, black and white)
• Sentence stems: partially fixed frames (The reason for this is…, It is worth noting that…)

Why Multiword Units Matter for Fluency

Native speakers store and access a vast repertoire of MWUs. These pre-built chunks are retrieved faster than assembled sequences, which is a key contributor to fluency. Learners who have not internalized MWUs must assemble each utterance from scratch, resulting in slower, more effortful production and a less natural “feel.”

Research suggests native speakers store tens of thousands of MWUs and rely on them heavily — possibly for 50% or more of fluent speech. See Formulaic Language.

Acquisition of Multiword Units

Learners often acquire MWUs first as unanalyzed chunks (I don’t know, What’s your name?) before breaking them down into components. This is particularly evident in children’s L1 acquisition and in adult immersion learners. Over time, learners must balance:

• Learning MWUs as wholes (for fluency)
• Understanding their internal structure (for productivity)

Instructional Implications

• Vocabulary instruction should explicitly teach common collocations, not just individual words
• Learners benefit from exposure to large amounts of authentic input rich in MWUs
• Concordancers and corpus tools help identify which MWUs are most frequent
• Sentence mining and flashcard systems should include MWU examples in context

Related Terms

• Formulaic Language: Fixed and semi-fixed sequences in language
• Collocation: Co-occurrence patterns between words
• Chunking: Cognitive process of grouping items into units
• Phrasal Verbs: Multi-part verbs with idiosyncratic meanings
• Idiom: Non-compositional fixed expressions