Consider the sentence:
Arun is tall.
What exactly counts as:
tall?
180 cm?
175 cm?
170 cm?
The answer depends partly on:
- comparison group;
- age;
- context;
- purpose.
Now consider:
Arun sat by the bank.
Which bank?
financial institution?
or:
river bank?
These two problems are not the same.
Tall is usually a case of vagueness.
Bank can be a case of ambiguity.
The central principle is:
Vagueness involves an unclear boundary inside one meaning; ambiguity involves more than one distinct meaning or structural interpretation.
This distinction matters because the wrong diagnosis produces the wrong repair.
Vagueness is about boundaries
Oxford and Cambridge work on vagueness repeatedly uses adjectives such as:
tall;
short;
big;
small.
Why?
Because these words have:
borderline cases.
A person can be:
- clearly tall;
- clearly not tall;
- somewhere near the boundary.
The difficulty is not:
which dictionary sense of tall?
The difficulty is:
where the category stops.
Ambiguity is about competing interpretations
Now:
bank.
English has at least two highly familiar lexical senses:
- financial institution;
- land beside a river.
In:
She walked to the bank,
the reader may need context to choose:
sense A
or:
sense B.
That is not a fuzzy boundary.
It is:
multiple candidate meanings.
A simple diagnostic question
Ask:
Am I unsure because there are two different meanings?
If yes:
ambiguity.
Ask:
Or am I unsure because one meaning has no exact cut-off?
If yes:
vagueness.
This single contrast solves many classification errors.
Vague words can be perfectly useful
Vagueness is not automatically:
bad language.
Consider:
Please arrive early.
How early?
Five minutes?
Fifteen?
Thirty?
The word:
early
has no universal numerical boundary.
Yet the instruction may be completely adequate in an ordinary setting.
Language often works because speakers share:
reasonable contextual standards.
Vagueness becomes a problem when the decision boundary matters
Imagine:
Students who arrive early receive extra time.
Now:
early
controls:
a benefit.
The boundary matters.
A fair policy may need:
Students who arrive before 8:00 a.m.
Precision becomes necessary because:
the rule creates consequences.
Law and policy often replace vague words with operational thresholds
Ordinary language may use:
young;
old;
large;
late.
Institutional systems may need:
age 18;
more than 100 square metres;
after 5:00 p.m.
The exact threshold is not necessarily the “true meaning” of the vague word.
It is:
an operational decision boundary.
“Tall” depends on a comparison class
Sentence:
Mei is tall.
Tall compared with:
The adjective uses:
a contextual comparison class.
Change the comparison class and the same height can receive a different judgement.
Vagueness and context interact
Context can narrow the likely standard.
But even with context fixed, borderline cases can remain.
This is one reason philosophical and linguistic work on vagueness is difficult.
A context can help.
It does not always create:
a perfectly sharp boundary.
Borderline cases are the signature of vagueness
Suppose:
190 cm = clearly tall.
And:
150 cm = clearly not tall
for a particular adult comparison group.
What about:
171 cm?
173?
175?
Some region may feel:
borderline.
Current philosophical work continues to analyse these:
boundary crossings.
The Sorites problem shows why vague boundaries are strange
Imagine one grain of sand.
Not a heap.
Add one grain.
Still not obviously a heap.
Repeat.
At some point:
heap.
But which exact grain created the change?
This is the classic:
Sorites problem.
Vague words can tolerate tiny changes while still having clearly different endpoints.
Students do not need the full philosophy to use the insight
The practical lesson is:
if a word describes a graded property, ask whether there is a genuine sharp threshold.
If not:
do not pretend the language is more exact than it is.
Ambiguity can be lexical or structural
Lexical ambiguity:
bank.
Structural ambiguity:
I saw the man with the telescope.
Who has the telescope?
I do?
or:
the man?
Both are ambiguity because:
two distinct interpretations compete.
Vagueness is different.
This separates the topic from the live sentence-ambiguity article
The eduKateSG sentence-ambiguity page owns:
multiple readings caused by structure or word placement.
This page owns:
the difference between multiple readings and fuzzy lexical boundaries.
That distinction protects both pages.
Vagueness is not the same as polysemy
Polysemy:
one form + several related senses.
Vagueness:
one relevant sense + uncertain boundary/application.
Take:
light.
It can mean:
- illumination;
- not heavy;
- pale colour;
Those are sense questions.
But within:
light bag,
there can still be:
vagueness about how many kilograms count as light.
A word can therefore be:
polysemous overall
and:
vague within one sense.
Underspecification is another nearby concept
Consider:
Mei went to school.
The sentence does not tell us:
That does not make:
went
ambiguous or necessarily vague.
The sentence is simply:
underspecified.
Language routinely leaves irrelevant details unstated.
Underspecification is efficient
If every sentence had to encode:
every detail of the world,
communication would collapse.
Good language leaves out:
information the current task does not need.
So:
missing detail ≠ defective meaning.
Vagueness and imprecision are also related
Sentence:
The room is about 20 metres long.
The word:
about
deliberately creates:
approximation.
This is imprecision.
It can be appropriate when:
exact measurement is unavailable or unnecessary.
Vagueness concerns:
category/application boundaries.
Imprecision concerns:
how exact a stated value is.
“About 20” and “tall” are therefore not identical
About 20 gives:
a numerical centre + tolerance region.
Tall gives:
a context-sensitive gradable predicate.
Both can be fuzzy in ordinary communication.
The mechanisms differ.
Prototype theory and vagueness can interact
The live prototype-theory article explains:
central vs peripheral members.
Vagueness asks:
where category application becomes uncertain.
These can overlap in everyday categories.
But:
typicality is not identical to vagueness.
A penguin can be an atypical but clearly valid bird.
No membership vagueness is required.
This matters in Science
Scientific language often tries to reduce:
ordinary-language vagueness.
For example, a laboratory procedure may replace:
heat strongly
with:
heat to 80°C for 5 minutes.
Operationalisation converts:
vague instruction
into:
repeatable procedure.
But Science also uses legitimate vague terms
Terms such as:
approximately;
significant;
high;
low
can be meaningful when:
the discipline defines their context.
The goal is not:
remove every vague word.
It is:
know when the boundary affects reproducibility or inference.
This matters in Mathematics
Mathematics often replaces vague everyday concepts with:
formal definitions.
That is one reason mathematical categories feel sharper.
But word problems can still contain:
vague ordinary language.
Students must know which part is:
formal
and which part is:
contextual.
This matters in History
Words such as:
wealthy;
powerful;
large;
rapid
can be vague.
A strong historical argument should ask:
compared with what?
and:
measured how?
Otherwise the adjective can hide:
an undefined standard.
This matters in argumentative writing
Claim:
Teenagers spend too much time online.
The phrase:
too much
contains a standard.
Too much for:
The writer should define:
what harm or threshold makes the amount excessive.
Vague language can be rhetorically convenient
Advertising:
Get results fast.
How fast?
Policy:
Applications will be processed promptly.
How promptly?
Vagueness can allow:
flexibility.
It can also avoid:
accountability.
Critical readers should notice the difference.
Ambiguity usually needs disambiguation
Sentence:
Meet me at the bank.
If two senses are plausible and action depends on the choice, the message needs:
more context.
Example:
Meet me at the DBS branch.
or:
Meet me by the river bank.
Ambiguity repair usually means:
select one interpretation.
Vagueness repair usually means setting a standard
Sentence:
Submit early.
Possible repair:
Submit by 5 p.m. on Friday.
The meaning was not split into two senses.
The boundary was too loose for the task.
That difference creates two repair algorithms
Ambiguity:
- list competing interpretations;
- choose intended one;
- rewrite to exclude the others.
Vagueness:
- identify the fuzzy dimension;
- decide whether precision is necessary;
- state a threshold, comparison class or operational criterion if needed.
Singapore relevance
Singapore students move between:
- ordinary conversation;
- school rules;
- Science procedures;
- exam questions;
- public policy;
- formal writing.
The same vague word can be acceptable in one setting and unsafe in another.
Come early may work between friends.
A national examination reporting time needs:
an exact clock time.
A practical vagueness test
Take:
large.
Ask:
If yes:
vagueness is likely.
A practical ambiguity test
Take:
bank.
Ask:
If yes:
lexical ambiguity/polysemy/homonymy issue.
Parent-friendly practice
Ask:
Is 175 cm tall?
Let the child answer.
Then ask:
For whom?
Now:
What does “bank” mean?
Let the child give two meanings.
Explain:
tall has one relevant meaning with a fuzzy border;
bank has competing meanings.
Teacher-friendly diagnosis
Vagueness/ambiguity collapse: every unclear sentence labelled ambiguous.
Polysemy/vagueness confusion: multiple senses and fuzzy boundaries treated alike.
Missing-detail error: underspecification treated as ambiguity.
Precision absolutism: all vague language treated as bad writing.
Operationalisation failure: consequential rule uses vague threshold without definition.
Context blindness: comparison class for gradable adjective not identified.
Hilary Mantel as a high-level literary lens
A word can be exact in atmosphere while remaining deliberately imprecise in measurement.
The transferable lesson is:
uncertainty has different shapes, and a careful writer knows which shape the sentence contains.
The value is diagnostic precision, not imitation.
AI and lexical vagueness
AI systems often treat:
vague
as:
unknown.
Or they invent a threshold:
“tall means above 180 cm.”
That can turn context-sensitive language into false precision.
A useful public audit prompt is:
Classify the uncertainty as ambiguity, vagueness, imprecision or underspecification. If vague, identify the comparison class or dimension and do not invent a sharp threshold unless the relevant domain defines one.
The larger lesson
Tall can be unclear.
But it is not unclear in the same way as:
bank.
Tall has:
a fuzzy boundary.
Bank has:
competing senses.
And:
Mei went to school
can leave details unstated without being either vague or ambiguous in the relevant sense.
The mature vocabulary learner asks:
What kind of uncertainty is this?
That question comes before repair.
Connections eduKateAI can learn
vagueness → one meaning + fuzzy application boundary
ambiguity → multiple competing interpretations
underspecification → detail omitted because task may not require it
imprecision → approximate value or tolerance
vague predicate → comparison class/context may shift standard
AI precision → do not invent thresholds where language supplies none
The public principle is that uncertainty in language has several distinct mechanisms, and accurate reading begins by distinguishing a vague boundary from an ambiguous meaning or merely omitted detail.
Related eduKateSG resources
Research basis
This article draws on the Oxford Handbook of Experimental Semantics and Pragmatics chapter on vagueness, the Cambridge Handbook of Formal Semantics discussion of vague gradable predicates, and 2025 Oxford work on borderline cases and vague boundary crossings. Sources include https://academic.oup.com/edited-volume/38644/chapter-abstract/335403789, https://www.cambridge.org/core/books/abs/cambridge-handbook-of-formal-semantics/vagueness/7C3B9429AB56D91A10557CAD20B6A3F8, and https://academic.oup.com/analysis/article/85/4/848/8327105.