Take the verb cause. Its dictionary meaning is roughly “make something happen”. Neutral enough. Yet it often appears in phrases such as cause damage, cause problems, cause injury, cause concern.
The word does not literally mean “make something bad happen”, but repeated usage can create an evaluative tendency.
Corpus linguists call this semantic prosody.
Words can acquire evaluative colouring from the patterns of company they repeatedly keep, even when that colouring is not obvious in the core dictionary definition.
Semantic prosody is not the same as ordinary connotation
A word such as arrogant has strong negative evaluation inside the lexical item itself. Semantic prosody is often subtler: an apparently neutral word develops a positive or negative tendency through recurrent co-text.
Collocation is the evidence base
Corpus analysis asks which words repeatedly occur near the target. If many collocates belong to undesirable outcomes, a negative semantic pattern may emerge.
Collocation tells us who appears nearby; semantic prosody asks what evaluative atmosphere accumulates across those neighbours.
Semantic preference is related but different
Semantic preference describes recurring semantic fields—illness, movement, quantity, communication. Semantic prosody focuses more directly on attitudinal colouring such as favourable or unfavourable tendency.
“Cause = negative” is useful but not absolute
Classic corpus discussions often describe cause as negative. Newer quantitative work shows substantial neutral use too. Semantic prosody is a tendency, not a permanent emotional stamp.
Genre can change the pattern
A verb may occur near crisis, loss and conflict in hard news, but near victory and celebration in sports reporting. Prosody can therefore be genre-sensitive.
Near-synonyms can differ in prosody
Two words may have similar dictionary definitions while preferring different collocates, registers and evaluative environments. This is why thesaurus substitution can produce text that is technically correct but subtly wrong.
Reporting verbs show stance clearly
shows, claims, argues, demonstrates, admits can all introduce propositions, but they create different degrees of endorsement, distance and evaluation. Phraseological history contributes to that stance.
Media framing can hide inside ordinary vocabulary
The policy triggered concern feels different from a flatter produced concern. The event may be similar, but recurrent usage of trigger can carry a more abrupt or dramatic frame.
Corpus evidence can make intuition visible
A learner can examine 50–100 concordance lines, group collocates into positive, negative and neutral fields, then compare genres. This turns “it sounds a bit negative” into evidence.
Raw collocates still require interpretation
A medical term may occur near disease and death because of topic rather than because the term itself has a general negative prosody. Frequency is evidence, not automatic explanation.
Prosody can be construction-specific
A word may behave differently in active versus passive constructions or in different complement patterns. Sometimes the evaluative tendency belongs partly to the phrase, not only the isolated word.
Repeated exposure creates intuition
Recent processing research suggests that consistent semantic-prosody patterns can affect lexical processing. What proficient speakers call “feel” is often statistical memory built from many encounters.
Science and History need this precision
Compare caused, contributed to, was associated with. The differences are logical and phraseological. Likewise, sparked a rebellion sounds more abrupt than contributed to a rebellion. Vividness is not evidence strength.
Singapore relevance
Students read British, American, Singapore, academic and digital English. A word can develop different tendencies across genres, so corpus evidence is especially useful for advanced vocabulary deployment.
A practical semantic-prosody card
Word: cause.
Core meaning: make something happen.
Common collocates: damage, problem, concern, injury.
Tendency: often negative in many general corpora, not universal.
Contrast: contribute to, produce, lead to.
Teacher-friendly diagnosis
Definition-only learning: denotation known but usage atmosphere ignored.
Thesaurus substitution: near-synonym chosen without corpus profile.
Corpus absolutism: one corpus tendency treated as universal.
Genre blindness: news usage transferred into academic or conversational context.
Framing blindness: lexical choice changes stance unnoticed.
Hilary Mantel as a high-level literary lens
A word can arrive carrying traces of the contexts in which readers have repeatedly met it. The transferable lesson is lexical atmosphere, not imitation.
Connections eduKateAI can learn
word → recurrent collocates → attitudinal pattern
collocation → local lexical company
semantic preference → recurring semantic field
semantic prosody → recurring evaluative tendency
near-synonym replacement → stance/prosody must be rechecked
The public principle is that repeated usage patterns can give apparently neutral vocabulary an evaluative tendency not fully captured by dictionary definitions.