Modal verb items are deceptive because the trap is rarely a grammar trap. With most Part 5 categories, three options are ungrammatical and one is correct, so elimination works. With modal items, two, three, or even all four options are perfectly grammatical in isolation. She can attend, she could attend, she should attend, she must attend are all legal English sentences. The test is asking something subtler: which degree of modality does the surrounding context require? Answering that question is a reading skill disguised as a grammar skill, and candidates who treat it as pure grammar will keep losing these items.
This guide maps the modal system onto the three axes TOEIC Link actually tests, and gives you a procedure for reading the context that disambiguates them.
The three axes of modality
Every modal item is probing one of three dimensions. Your first move is to identify which.
Axis 1: Certainty (epistemic modality)
This axis runs from "almost impossible" to "logically certain," and it describes how sure the speaker is about a fact.
- must — strong logical certainty: The shipment must have arrived; the tracking shows delivered.
- should — reasonable expectation: The report should be ready by now.
- may / might / could — possibility: The client may want a revision.
- can't / couldn't — logical impossibility: That can't be the final figure; it's missing a column.
The trap on this axis is treating must as obligation when the context signals deduction. The invoice must be wrong is a conclusion drawn from evidence, not a command. Look for evidence words nearby — tracking shows, the numbers indicate, based on — which pull the modal toward the certainty reading.
Axis 2: Obligation and necessity (deontic modality)
This axis describes how strongly something is required.
- must / have to — strong obligation: Employees must complete the training.
- should / ought to — recommendation: You should review the contract before signing.
- need to — necessity, slightly softer than must
- don't have to — absence of obligation (note: this is not prohibition)
The most common obligation-axis trap is the mustn't / don't have to swap. Mustn't = prohibition (you are forbidden). Don't have to = optional (you are not required). Visitors mustn't enter the lab and Visitors don't have to enter the lab mean opposite things, and the test exploits exactly this.
Axis 3: Ability and permission (dynamic modality)
- can / be able to — ability: The system can process 10,000 records per hour.
- could — past ability, or polite/hypothetical: We could increase capacity if needed.
- may / can — permission: Members may access the archive.
The trap here is past-context items where can is illegal because the timeframe is historical, forcing was able to / could. Watch the surrounding tense markers — if the sentence is anchored in the past, present can is usually the trap.
The context-reading procedure
When you hit a modal item, do not pick the modal that "sounds most natural." Run this:
- Identify the axis. Is the sentence about how sure someone is (certainty), how required something is (obligation), or what someone can do / is allowed to do (ability/permission)? Signal words usually give it away.
- Find the calibrating clue. Almost every modal item contains a phrase that fixes the degree: given the deadline, according to policy, based on the data, if the budget allows. This clue is what separates should from must, or may from can't.
- Match the degree, not just the axis. Once you know it is, say, an obligation item, the clue tells you whether it is strong (must), recommended (should), or absent (don't have to).
- Check the timeframe. Past contexts kill present-only forms like bare can in ability senses and force was able to / had to / could have.
This name-the-axis-then-read-the-clue method is the same structured reading discipline the rest of the series applies to other categories — compare the approach in our conditionals and if-clauses treatment, where the trap is also a matter of reading which scenario the sentence sets up rather than which form is legal.
Perfect modals: the high-value sub-category
A meaningful share of harder modal items use the modal + have + past participle construction, which compresses a judgment about the past into a single phrase.
- must have + p.p. — past certainty: She must have left early; her office is dark.
- should have + p.p. — past expectation unmet, often regret: We should have ordered more stock.
- could have / might have + p.p. — past possibility: The error could have come from the import step.
- can't have + p.p. — past impossibility: He can't have approved it; he was on leave.
These map onto the certainty axis but in the past, and they are worth specific drilling because the construction is unusual enough that candidates freeze on it. The clue is the same: evidence in the sentence calibrates the certainty. Treat perfect modals as a named sub-pattern in your error log so they enter your spaced-review cycle rather than being re-puzzled each time, as recommended in our error log design for spaced review cycles.
Common business-context collocations
Some modal-flavored items are really fixed business phrasings:
- would like to (polite want) vs will (future fact)
- would appreciate it if (polite request)
- should you have any questions (formal conditional inversion of if you have)
That last one — inverted should replacing if in formal register — appears regularly in TOEIC Link's email-style items and trips candidates who have never seen the inversion. Should you require further assistance, please contact us means If you require further assistance. File it as a recognized pattern.
The bottom line
Modal items reward reading, not memorizing. The grammar of modals is rarely the obstacle; the obstacle is calibrating the degree of certainty, obligation, or permission against the clue the sentence provides. Identify the axis, find the calibrating clue, match the degree, and check the timeframe. Do that consistently and a category that used to feel like four equally plausible options collapses into one determinate answer.