Parallel Reasoning
Key Takeaways
- Parallel Reasoning questions ask for the same argumentative structure, not the same topic or conclusion wording.
- The fastest filter is to abstract the stimulus into variables while preserving validity, conditional direction, quantifiers, and conclusion force.
- A wrong answer can discuss a similar subject and still fail because it changes a rule, reverses a condition, or strengthens the conclusion.
- On current Logical Reasoning sections, parallel items are often time-expensive, so disciplined filtering protects the rest of the section.
What Parallel Reasoning Asks
Parallel Reasoning questions ask which answer uses the same pattern of reasoning as the stimulus. The subject matter is a disguise. A stimulus about laboratory access can parallel an answer about shipping permits if both arguments move from the same kind of rule to the same kind of conclusion.
LSAC describes Logical Reasoning as testing ordinary-language arguments, including the ability to recognize similarities and differences between reasoning patterns. That is exactly this question type. You are not asked whether the argument is wise. You are asked whether the answer has the same logical design.
A good parallel answer preserves the following features:
- number of major terms or groups
- conditional direction
- quantifier force such as all, most, some, or none
- positive or negative conclusion
- valid or invalid status
- kind of support, such as analogy, elimination, rule application, or causal comparison
Reduce The Stimulus
Do not read five answer choices at full detail before reducing the original. First write a compact skeleton. For a conditional item, if P then Q; P; therefore Q is enough. For a quantifier item, use labels like some A are B; all B are C; so some A are C. For a method item, use a plain-language map: cites a general rule, shows the case meets the rule, draws the required result.
The skeleton should preserve force. If the conclusion says must, cannot, probably, should, or is more likely, keep that strength in the map. Parallel answers often fail by moving from a must conclusion to a likely conclusion or from a recommendation to a factual claim.
High-Yield Filters
| Feature | Ask this first | Why it matters |
|---|---|---|
| Validity | Is the original valid or flawed? | A valid proof should not match an invalid leap |
| Direction | Which condition triggers which result? | Reversals create different structures |
| Quantity | Are the claims all, most, some, or none? | Quantity controls what follows |
| Polarity | Is the conclusion positive or negative? | Contrapositive forms often end negative |
| Terms | How many categories are linked? | Extra or missing categories change the skeleton |
Use these filters before becoming interested in topic. If the original uses three categories and a universal bridge, an answer using two categories and a causal claim is gone even if both talk about science.
Valid Patterns
Many parallel questions use clean valid forms. A common one is modus ponens: if P then Q; P; therefore Q. Another is modus tollens: if P then Q; not Q; therefore not P. A third is a transitive chain: if P then Q; if Q then R; therefore if P then R.
A valid stimulus can have complicated wording but a simple engine. Suppose an argument says that any proposal meeting the budget rule must receive review, this proposal meets the rule, so it must receive review. The correct answer should also apply a sufficient condition to trigger a necessary result.
Quantifier Matches
Quantifier parallel items require restraint. From all A are B and all B are C, all A are C follows. From some A are B and all B are C, some A are C follows. But from all A are B and some B are C, no conclusion about all A follows.
Match the amount claimed. If the stimulus concludes that at least one member of a group has a property, an answer that concludes every member has it is too strong. If the stimulus concludes no member qualifies, an answer concluding merely that one named person does not qualify may be too narrow.
Process Under Time
Parallel items can consume time because each answer is a mini-stimulus. Use a two-pass routine. On the first pass, eliminate by obvious mismatches: wrong conclusion strength, wrong validity, reversed condition, different number of groups. On the second pass, compare the remaining answer or two line by line.
Do not translate every noun into variables if the reasoning is not formal. For analogy or method parallels, state the move: rejects a generalization by giving a counterexample, recommends action by applying a rule, or predicts an outcome by eliminating alternatives. The correct answer should repeat that move.
Common Wrong Answers
- Similar topic, different logic
- Same conclusion type, different evidence type
- Same rule language, reversed condition
- Same flaw family, but this is a regular parallel question with a valid original
- Stronger or weaker conclusion than the stimulus
- Extra premise that does real work the original did not need
Review parallel misses by writing both skeletons next to each other. If the mismatch is visible only after reading an explanation, your skeleton was probably too topic-bound. The goal is not to memorize forms. It is to see the argument as a reusable structure before the answer choices pull you into their stories.
When To Stop Comparing
After the skeleton and force match, do not keep searching for a perfect topic match. A parallel answer can use different people, institutions, or objects. Once the operative rule, evidence pattern, and conclusion type align, the subject matter should no longer matter.
On a current LSAT, there are two scored Logical Reasoning sections, each under a 35-minute clock. Parallel questions are worth the same as easier items. If a parallel item is dense but your skeleton is clear, mark it, bank faster points, and return with a fresh eye.
If a satellite's backup battery fails, its telemetry stream becomes intermittent. The telemetry stream is not intermittent. Therefore, the backup battery has not failed. Which argument most closely parallels this reasoning?