Analytical & Logical Reasoning
Seating puzzles, blood relations, and syllogisms — the skill is drawing a diagram, not holding everything in your head.
What you'll learn
- The four puzzle shapes GATE recycles: seating/ordering, blood-relations, syllogisms, statement-conclusion
- Drawing beats memorising — sketch the situation before deducing
- Why 'all A are B' and 'some B are C' do NOT imply 'some A are C'
- Using a Venn diagram to test syllogism validity in 20 seconds
Before you start
Last lesson promised a puzzle with almost no numbers, where the failure mode is juggling the whole tangle in your head. Here it is. You read once — five people, a round table, “A is between B and C,” “D is opposite A” — and your brain instantly tries to hold all of it at once. That never works. The students who solve these in ninety seconds are not smarter; they simply draw the diagram.
Analytical reasoning rewards externalising the puzzle. A four-clue knot becomes trivial the moment it leaves your head and lands on rough paper, where the eyes can do the bookkeeping the working memory cannot. (It is the same move you make sketching a tricky SQL join or a tangle of if conditions — get the constraints out of your head, then read the answer off the page.)
The four shapes GATE recycles
-
Seating / ordering. People around a table or in a row. Draw the seats; pencil each name in as clues land.
-
Blood relations. “A’s sister’s husband’s father…” Draw a family tree:
- squares for males;
- circles for females;
- lines for marriage and parenthood.
Unwind the phrase left to right, one pencil stroke per link. Start at A, step sideways to a sister, sideways again across a marriage line to her husband, then up one generation to his father — who is A’s sister’s father-in-law.
-
Syllogisms. Two premises and a proposed conclusion — “All A are B; some B are C; therefore…?” — where the only question is whether the conclusion is forced, never whether it sounds right. Test by Venn diagram, never by gut feel.
-
Statement-conclusion. Given a statement, which conclusions necessarily follow? Reject any that need an extra assumption.
A round-table puzzle, drawn out
Five friends A, B, C, D, E sit around a round table. A is between B and C. D is opposite A. Where does E sit?
Walk it through, pencilling each clue onto the diagram:
- Place A anywhere — call it the top seat.
- D opposite A — the bottom seat.
- A between B and C — B and C take the two seats flanking A.
- E is the only person left, and two seats remain: one between B and D, one between C and D.
Without a fifth clue (say “E is to D’s right”), the puzzle is under-determined — the clues allow more than one valid answer, and E could sit in either remaining seat. GATE will always give you enough clues to pin it down, so the feeling “I’m missing a clue” is your cue to re-read for one you skimmed past.
Syllogisms — when intuition lies
Here is the deduction your gut will get wrong every time:
All cats are mammals. Some mammals are dogs. Therefore some cats are dogs?
It looks plausible for half a second. It is wrong, and the reason is worth saying out loud: the “some B” in one premise need not be the same B as in the other.
The mammals that happen to be dogs and the mammals that happen to be cats are two subsets of one bigger set, and nothing in the premises forces them to touch. Picture it — a large mammals circle holding a cats circle and a dogs circle that sit well apart. Both premises hold, the conclusion fails, and that settles it: “all A are B” + “some B are C” never implies “some A are C”.
That single counter-picture is also the general method. Whenever a conclusion feels right, try to arrange three circles so every premise holds and the conclusion does not. Manage it and the syllogism is invalid; fail after honest effort and it is valid.
How GATE asks this
A mix of MCQ and NAT. The MCQs are usually pure deduction puzzles (“who sits where?”, “what is X’s relation to Y?”, “which conclusion follows?”). The NATs are usually numeric (“how many people sit between A and B going clockwise?”, “what is the minimum number of people in the room?”). Two skills, one toolkit — externalise the puzzle, then deduce.
In one breath
Analytical reasoning recycles four shapes:
- seating/ordering
- blood relations
- syllogisms
- statement-conclusion
The one habit that cracks them all is to externalise the clues rather than juggle them in your head:
- a seat diagram;
- a family tree with squares/circles;
- three Venn circles.
The trap that fools careful students is the syllogism “all A are B” + “some B are C” ⇏ “some A are C” (disjoint subsets). Around a round table “opposite” only exists when the number of seats is even.
Practice
Quick check
A question to carry forward
Notice what saved you in every puzzle here: you could externalise it.
- a seat diagram;
- a family tree;
- three Venn circles.
Each turned the words into a picture your eyes could read off. The clues lived in language, and language could be drawn.
The final GA bucket takes the words away and hands you the picture instead — then mangles it. The transformations include:
- a square of paper folded twice and punched;
- a cube flattened into a paper cross;
- a shape held up to a mirror.
There is nothing to deduce in sentences, and often nothing useful to draw, because the puzzle is the transformation itself. You have to mentally perform:
- the fold;
- the rotation;
- the reflection.
Then see what comes out. Here is the thread onward, the last piece of General Aptitude. How do you track the following when the entire puzzle lives only in your mind’s eye?
- a punched hole back through each fold;
- a face around a tumbling cube;
- a figure through its mirror?
Practice this in an interview
All questionsReasoning models are optimized to spend extra inference-time computation on intermediate steps, while test-time compute is the broader practice of allocating more computation during an answer through longer reasoning, multiple candidates, verification, search, or tools. It can improve hard, verifiable tasks, but adds cost and latency and does not fix missing knowledge or correlated errors.
Chain-of-Thought prompting asks a language model to produce intermediate steps before its final answer, usually through worked examples or a step-by-step instruction. Those steps can decompose multi-step tasks and improve accuracy, but they add cost and are not guaranteed to be faithful or useful, especially with modern reasoning models.
Chain-of-thought (CoT) prompting instructs the model to write out intermediate reasoning steps before producing a final answer, which improves accuracy on multi-step arithmetic, logic puzzles, and compositional questions. It is most impactful on models with at least ~10B parameters and on tasks where the answer space is large enough that guessing is hard.