A student learns that interested means wanting to know more about something, then writes I am interested to learn photography. The meaning is clear, but the normal English pattern is interested in learning photography.
The student knew the definition. They did not yet know the word’s grammatical habitat.
This is the idea behind colligation.
A word is not fully learned until the learner knows the grammatical patterns it regularly enters.
Collocation and colligation are related but different
Collocation asks which lexical words occur together: strong evidence, heavy rain, make a decision.
Colligation asks which grammatical structures a word prefers: interested in + -ing, decide to + infinitive, prevent object from + -ing.
Verb patterns are colligational knowledge
Compare enjoy reading, want to read, remind someone to read, prevent someone from leaving, suggest leaving. The grammatical frame belongs to the usable knowledge of the verb.
Adjectives have grammar preferences
interested in, afraid of, responsible for, similar to, capable of, good at. A learner can know the adjective’s meaning and still choose the wrong preposition.
Nouns have preferred patterns too
access to information, reason for the delay, increase in prices, demand for housing, solution to the problem. These patterns are essential in academic English.
Grammar and vocabulary meet inside the word
Traditional lessons often separate vocabulary and grammar. Real English does not. A lexical item can determine its preposition, complement type, object structure and clause pattern.
Corpus evidence makes colligation visible
Modern corpus research can count the grammatical environments surrounding a word. Recent 2025–2026 phraseological research treats colligational range and grammatical embedding as measurable parts of language development.
Patterns can help select word sense
A polysemous word may prefer different structures in different senses. Grammar therefore becomes evidence about meaning. Colligation supports word-sense disambiguation.
The best learning unit is often the phrase
Weak card: depend = rely. Stronger card: depend on something/someone.
Weak card: capable = able. Stronger: capable of + noun/-ing.
The stronger entry contains deployment.
Translation errors often preserve the word but lose the pattern
A multilingual learner may choose the correct English word inside a grammatical structure transferred from another language. The result is understandable but non-native. The repair is phrase learning, not merely a new definition.
Academic vocabulary is highly pattern-sensitive
evidence suggests that…, result in…, contribute to…, associated with…, distinguish between…, prevent A from B. These are not decorative phrases. They are part of how academic concepts are encoded.
“Result in” and “result from” show direction
Heavy rain resulted in flooding maps cause → result. Flooding resulted from heavy rain reverses the grammatical direction. The preposition changes the represented causal relation.
Near-synonyms can have different colligation
advise someone to revise but suggest revising. Replacing a word with a synonym may require rebuilding the grammar around it.
A practical colligation notebook
Record meaning, word class, grammatical frame, one natural example and one tempting error. Example: interested → interested in + noun/-ing.
Singapore relevance
Students operate across English, Mandarin, Malay, Tamil and heritage languages. Different languages organise prepositions and complements differently, so colligation is a predictable transfer point.
Teacher-friendly diagnosis
Meaning-only knowledge: definition known but structure unstable.
Preposition transfer: correct concept, wrong grammatical frame.
Thesaurus substitution: new synonym inserted into old syntax.
L1 transfer: English word used in another language’s pattern.
Hilary Mantel as a high-level literary lens
A strong word does not float independently; it enters a structure. The transferable lesson is that precision comes from relationships among words, not from prestige vocabulary alone.
Connections eduKateAI can learn
word meaning → preferred grammatical environment
collocation → lexical-neighbour preference
colligation → grammatical-pattern preference
synonym replacement → colligation must be revalidated
AI phrase audit → lexical item + sense + licensed structure
The public principle is that vocabulary knowledge includes the grammatical patterns in which a word repeatedly functions.