Necessary vs sufficient conditions: avoid reversing the rule
In 'if P, then Q', P is sufficient for Q, and Q is necessary for P. Knowing P lets you infer Q; knowing Q does not generally let you infer P. The equivalent contrapositive is 'if not Q, then not P'. To test a proposed conclusion, try to find a case in which the stated rule is true but the conclusion is false.
An IMAT logic question may describe a rule that sounds familiar while quietly reversing its direction. Translate the rule before deciding whether the conclusion follows.
Required is not the same as enough
Suppose an invented reading club has this rule:
If someone is a club mentor, they have completed the training.
Let M mean “is a mentor” and T mean “has completed the training”. The rule is M → T.
Being a mentor is sufficient to establish completion of the training. Completing the training is necessary for being a mentor. But the rule allows trained members who are not mentors.
These terms describe a logical relationship within the stated rule, not a claim that becoming a mentor somehow causes earlier training. OpenStax explains necessary and sufficient conditions.
Four conclusions that look similar
| Additional information | Proposed conclusion | Does it follow from M → T? |
|---|---|---|
| This member is a mentor | They completed the training | Yes |
| This member has not completed the training | They are not a mentor | Yes |
| This member completed the training | They are a mentor | No: other members may be trained |
| This member is not a mentor | They have not completed the training | No: they may still be trained |
The first valid pattern is modus ponens. The second is modus tollens. The tempting invalid patterns are affirming the consequent and denying the antecedent. OpenStax discusses valid deductive inference.
“If” and “only if” point in different directions
Use this translation table for categorical rules:
| Wording | Translation |
|---|---|
| P if Q | Q → P |
| P only if Q | P → Q |
| P is sufficient for Q | P → Q |
| P is necessary for Q | Q → P |
| P if and only if Q | Both P → Q and Q → P |
Only mentors can lead a club session means leading a session requires being a mentor: L → M. It does not state that every mentor leads a session.
Combine that with M → T and you obtain L → T. Anyone leading a session has completed training. You still cannot infer that every trained member leads sessions.
For “P unless Q”, the usual inclusive logical reading is if not Q, then P. For example, “the session runs unless the room is unavailable” gives room available → session runs. It does not by itself settle what happens when the room is unavailable; perhaps another room can be used. Follow any extra qualifications in the question rather than adding an unstated “if and only if”.
Reverse and negate together
The contrapositive of P → Q is not Q → not P. It is logically equivalent to the original implication.
For M → T, that means not trained → not a mentor. Reversing without negating gives the converse, T → M, which is not established. Negating without reversing gives not M → not T, also not established.
A useful scratch-work habit is to write the original arrow and its contrapositive beside each other. When you compare an answer option, check both the order of the statements and the presence of negation.
Build a counterexample instead of guessing
Consider the claim: “All mentors have completed training, so all trained people are mentors.” Imagine three members:
- Amira is a mentor and has completed training.
- Ben has completed training but is not a mentor.
- Cara has not completed training and is not a mentor.
Every mentor in this miniature world is trained, so the premise holds. Ben makes the conclusion false. One such possible case is enough to show the inference is invalid; you do not need to know how common it is in real life.
Also, a universal rule does not establish that mentors exist. “All mentors are trained” alone cannot justify “some trained people are mentors”. That needs an existence premise.
Self-check: follow the chain
Original practice rules: Every reviewer is trained. Only reviewers may approve a draft. Lina has approved a draft, in accordance with these rules.
What must follow? Lina is a reviewer, and therefore trained: approve → reviewer → trained.
Omar is trained. Must he be able to approve a draft? No. The rules do not establish that he is a reviewer, nor that every reviewer has permission to approve.
When an answer fails, record the exact reversal rather than “logic mistake”. Use the error-log method to practise the same inference structure with a fresh scenario. If the question instead asks how likely an event is, switch to conditional probability; a probability statement is not automatically a categorical implication.
Sources
Or keep going in the iPhone app.
Found a mistake? Write to hello@imatlearn.com