What Is a Lexical Chunk?

A lexical chunk is a string of words that behaves like a single vocabulary item. When a fluent speaker says as far as I know or I was wondering if you could, they are not building the phrase from scratch with grammar rules. They pull it from memory whole, the way they would retrieve a single word. Chunks are a large part of why native speech sounds smooth and why learners who memorize only single words often sound stilted.

Where the concept comes from

Pawley and Syder (1983) pointed out a puzzle: native speakers produce long, fluent stretches of speech faster than they could possibly plan them word by word, and they reliably choose the conventional way of saying things out of many grammatical options. Their explanation was that speakers have tens of thousands of memorized sequences. Nattinger and DeCarrico (1992) described these "lexical phrases" for teaching, and Michael Lewis's The Lexical Approach (1993) put them at the center of a method, famously summarized as "language consists of grammaticalised lexis, not lexicalised grammar."

Corpus research supports the idea. Erman and Warren (2000) estimated that roughly half of the spoken and written English they analyzed consisted of prefabricated combinations. Alison Wray's Formulaic Language and the Lexicon (2002) collected the evidence that these sequences are processed as wholes.

Types of lexical chunks

Why chunks matter for vocabulary learners

• Fluency. Retrieving a ready-made phrase costs less mental effort than assembling it, which frees attention for content. This is why a learner with fewer words but more chunks often sounds more fluent than one with a large vocabulary of single words.
• Accuracy. Chunks carry their grammar with them. If you learn look forward to hearing from you as a unit, you never have to wonder whether it is to hear or to hearing.
• Naturalness. Many grammatically correct sentences are simply not what people say. What time is it? is a chunk; What is the hour? is grammatical but odd.
• Comprehension. Knowing chunks lets you predict upcoming words, which speeds up listening in particular.

How to learn lexical chunks

1. Notice them. When reading, look for phrases that feel "pre-packaged," especially at the start of sentences and around verbs.
2. Save the whole phrase. Record it's worth bearing in mind that, not just bear. Sentence mining naturally captures chunks in context.
3. Review the chunk as a unit. Make flashcards with a gap: It's not worth ___ about (worrying).
4. Say them aloud. Shadowing audio trains the rhythm of chunks so they come out as one breath.
5. Swap the slots. Practice frames with new content: I was wondering if you could send / check / call...

Wordie can help with the noticing step: words you save are highlighted whenever they appear again on web pages, so you keep seeing them inside the chunks they typically live in, such as mitigate in to mitigate the risk of. The guide on learning collocations describes a routine that works equally well for longer chunks.

Common mistakes

• Memorizing lists of chunks without context. A list of 200 phrases is quickly forgotten; 20 phrases met in real sentences stick.
• Altering fixed parts. On the other side is not the discourse marker on the other hand; at the end of the day cannot become at the finish of the day.
• Overusing a few favorites. Repeating at the end of the day in every paragraph sounds formulaic in the bad sense. Vary and match the register.

Sources: Pawley, A. & Syder, F. (1983), Two puzzles for linguistic theory, in Language and Communication; Nattinger, J. & DeCarrico, J. (1992), Lexical Phrases and Language Teaching; Lewis, M. (1993), The Lexical Approach; Erman, B. & Warren, B. (2000), The idiom principle and the open choice principle, Text; Wray, A. (2002), Formulaic Language and the Lexicon.