Appendix A: Conditional Logic and Quantifier Reference

Appendix A: Conditional Logic and Quantifier Reference

Appendix A: Conditional Logic and Quantifier Reference

This is your one-stop cheat sheet. The main chapters use conditional and quantifier logic constantly but keep the mechanics light. Here is the full machinery, including the exotic cases you will hit only a few times per test. Use it to look things up, not to read straight through.

Notation: is the conditional arrow. ¬ means “not.” --m--> is a “most” arrow (one direction only). --some-- is a some-relationship (reads both ways).


1. Conditional basics

A conditional is a guarantee. S → N reads “if S, then N”: the sufficient condition S guarantees the necessary condition N. S is enough to force N. N is required for S. No time order and no cause is implied. The arrow is a logical sign, not a story.

  • Sufficient (S): enough on its own to trigger N. It need not happen for N to happen; other paths to N may exist.
  • Necessary (N): must hold whenever S holds. N can occur without S.

The contrapositive

The contrapositive is the only valid transformation: reverse AND negate.

S → N becomes ¬N → ¬S.

The contrapositive says exactly what the original says. If you ever need a shared term to sit on the other side of an arrow, take the contrapositive to move it.

The two classic errors

ErrorLooks likeWhy it fails
Mistaken ReversalN → SReverses without negating. Treats the necessary as if it were sufficient.
Mistaken Negation¬S → ¬NNegates without reversing. Treats the sufficient as if it were required.

A Mistaken Reversal and a Mistaken Negation of the same statement are contrapositives of each other, so both are wrong in the same way. When a stem hands you S → N and an answer asserts N, you can conclude nothing about S.


2. Indicator words

The indicator tells you which term is sufficient and which is necessary. Word order in the sentence does not matter.

Introduces the SUFFICIENT conditionIntroduces the NECESSARY condition
ifthen
when, wheneveronly, only if, only when
all, any, every, eachmust, required, requires
people who, anyone whoneeds, essential, depends on
in order tounless, except, until, without

Note that unless / except / until / without introduce the necessary condition but also require a negation step. See the next section.


3. Tricky operators (translation table)

PhraseRuleDiagram
A only if B”only” marks the necessaryA → B
A only when Bsame as “only if”A → B
Only A are Bthe “only” term is necessary for the otherB → A
A if and only if Bbiconditional, arrows both waysA → B and B → A
Unless B, ¬A”unless” term is necessary; negate the otherA → B
No X are Ythe two cannot coexistX → ¬Y (and Y → ¬X)
None of the X are Ysame as “no X are Y”X → ¬Y

The “only” trap

When you see only or only if, your instinct will be to make the nearby term sufficient, and that instinct is backwards. A only if B does not mean B → A. It means A → B. “Only the funded projects proceed” diagrams as proceed → funded, not the reverse. Everyday check: “Only employees may park in this lot” means a car parked there must belong to an employee (parks here → employee), not that every employee parks there.

The negate-the-other rule (unless / except / until / without)

Two steps, always the same:

  1. The term modified by unless / except / until / without becomes the necessary condition.
  2. Negate the other term and make it the sufficient condition.

Example: “Unless the city raises revenue, the bridge stays closed.” Necessary = raises revenue. Other term (“bridge stays closed”) negates to “bridge opens.” Result: bridge opens → raises revenue. Everyday check: “I will not leave unless the rain stops” makes “rain stops” the necessary term and negates the other, giving I leave → rain stopped. So if you left, the rain had stopped.

”If and only if” is a biconditional

A ↔ B packs two arrows: A → B and B → A, plus both contrapositives. A occurs exactly when B occurs. Triggers: “if and only if,” “when and only when,” “all and only.”


4. Compound contrapositives (the AND/OR flip)

When a term is itself an AND-group or an OR-group, the contrapositive flips AND to OR and OR to AND. This is the single rule that governs all four forms.

OriginalContrapositiveSplits?
A → (B and C)(¬B or ¬C) → ¬AYesA→B, A→C
(A or B) → C¬C → (¬A and ¬B)YesA→C, B→C
A → (B or C)(¬B and ¬C) → ¬ANo
(A and B) → C¬C → (¬A or ¬B)No

Which compounds split

A compound splits into two clean conditionals only when the AND sits on the necessary side or the OR sits on the sufficient side.

  • A → (B and C) splits: if A guarantees both, then A guarantees each. Write A → B and A → C.
  • (A or B) → C splits: if either trigger forces C, then each trigger forces C. Write A → C and B → C.

Do not split the other two. A → (B or C) only promises at least one of B or C, so you cannot pull out A → B alone. (A and B) → C needs both A and B together, so neither A → C nor B → C follows. Splitting these is a common self-inflicted error.


5. Chaining conditionals

Valid links join on a shared term where the term is necessary in one statement and sufficient in the next.

A → B and B → C give A → B → C, so A → C (contrapositive ¬C → ¬A).

If the shared term is on the wrong side, take a contrapositive to align it. Given A → B and C → ¬B, flip the second to B → ¬C, then chain: A → B → ¬C, so A → ¬C.

A → B and A → C do not give B → C.

Two arrows leaving the same sufficient term share no valid path, not even through contrapositives. “All maples are trees” and “all maples are deciduous” do not yield “all trees are deciduous.” Watch for this in Must Be True and Parallel answers.

Group, not individuals

A conditional describes the group as a whole, never each member. “The committee speaks many languages” does not mean each member speaks many languages. Do not distribute a group property down to individuals.


6. Quantifiers: all / most / some / none

QuantifierMeaningRange
all100%100
mostmore than half51 to 100
someat least one1 to 100
some are notat least one is not (“not all”)0 to 99
most are notmore than half are not0 to 49
nonezero0

Note that most and some both allow all as a possibility. “Most students passed” is true even if every student passed.

A horizontal scale from 0% to 100% showing the range each quantifier covers. 'all' is a single point at 100%. 'most' (majority) is a bar spanning 51% to 100%. 'some' (at least 1) is a longer bar spanning 1% to 100%, labeled 'yes, possibly all.' 'none' is a single point at 0%. A footer notes that 'some' is never less than half, and that both 'some' and 'most' can secretly mean all.
The coverage of each quantifier on a 0 to 100 percent scale: "some" and "most" both reach all the way up to 100 percent.

Reversibility

StatementReverses toDiagram
some A are Bsome B are A (valid)A --some-- B
no A are Bno B are A (valid)A → ¬B
most A are Bonly some B are AA --m--> B
all A are Bno reverse inference by itself; with separate proof that A exists, some B are AA → B

Some and none reverse freely. Most and all do not. Reversing a “most” or “all” statement is a Mistaken Reversal. From “most A are B” you may not conclude “most B are A,” only “some B are A.” From “all A are B,” you may not conclude anything backward unless you also know at least one A exists.

Single-statement (down-the-ladder) inferences

One statement implies weaker versions of itself. The implications run one direction only.

all A are B plus proof that at least one A exists ⟹ most A are Bsome A are B

no A are Bmost A are not Bsome A are not B

Upper rungs imply lower rungs once existence is established. Lower never imply upper. “All” gives you “most” and “some” only if the relevant group is known to exist; “some” gives you neither “most” nor “all.”


7. Overlap inferences (two statements, shared term)

When two quantified statements share a term, you sometimes get a guaranteed overlap. Most pairs give you nothing. The table shows when an inference exists and what it is.

Pair (sharing a term)Inference?Result
some + somenevernone
most + somenevernone
most + most (same shared group)alwayssome B are C
some + allsometimesdirection-dependent
most + allsometimesdirection-dependent
all + allalwaysA → C
none + anythingalwaysyields an inference

Direction conditions, briefly

  • most + most delivers only when both “most” arrows lead away from the shared middle term. B <--m-- A --m--> C gives B --some-- C: two majorities of the same set A must overlap. Two majorities pointing in line (one into A, one out) do not combine.
  • some + all works when the “all” arrow points away from the shared term so you can ride the some-link onto it. A --some-- B plus B → C gives A --some-- C. If the “all” arrow points the wrong way, no inference.
  • most + all works the same way: A --m--> B plus B → C gives A --m--> C. Against the arrow, it collapses to nothing useful.
  • all + all is plain chaining: A → B → C gives A → C.
  • none is the strongest input. A → B plus B → ¬C gives A → ¬C every time.

8. Quantifier chaining (the advanced reference)

This is the full system for combining quantified statements. We call it quantifier chaining. Think of each statement as a track and each shared term as a station. You travel from one endpoint to the other across the shared middle. Three rules govern every chain.

Link typeSymbolTravel direction
some--some--either way (reversible)
none / double-not→ ¬either way (reversible)
allfollow the arrow only
most--m-->follow the arrow only (one-way)

You may always ride a some link or a none link in either direction. You may ride an all or most link only in the direction it points. Going against a “most” arrow yields nothing better than the inherent “some” already inside it (a “most” contains a “some” pointing back).

A chain inference is only as strong as its weakest link. Rank strength: some (weakest) < most < all / none (strongest). The result takes the weakest link in the path.

  • most + all → the chain contains a “most,” so the output is at best most.
  • some + all → the chain contains a “some,” so the output is some.
  • all + all → both links are “all,” so the output is all (a clean ).
  • Any chain containing a “some” link can produce at most a “some” inference.

To get an inference at all, you generally need at least one all or none link, because those are the only links restrictive enough to force an overlap. A path made of only “some” links (some + some) forces nothing. A path of “most + most” is the special exception: two majorities of the same set must collide, which is why most + most (arrows leading away) yields a “some.”

Rule 3: relevant negativity

Track negation through the chain, but only the endpoints matter.

  • A negation on an endpoint term carries into the result.
  • A negation on the intermediate (shared middle) term is bypassed; it does not appear in the output.
  • A double-not / none link (→ ¬) always injects relevant negativity into the result, because it negates the far endpoint as you cross it.

So in A → B plus B → ¬C, the middle term B carries no negation outward, but the negation on the endpoint C survives: A → ¬C. Had the negation sat only on B in the middle of the chain, it would have washed out.

Combination table (which pairs yield an inference)

First statementSecond statementInference?Resulting statement
some + somen/anevernone
some + most(either order)never (generally)none
most + mostboth arrows lead away from shared middlealwaysA --some-- C
some + allall-arrow leads away from shared termsometimesA --some-- C
most + allall-arrow leads away from shared termsometimesA --m--> C
all + allarrows chain head-to-tailalwaysA → C
none + anythingnone-link crossablealwaysresult carries negativity, e.g. A → ¬C

Working principles

  • Start at open variables. Begin the chain at a term that appears in only one statement (an endpoint), or at a “some” term. Build outward from there.
  • All and none do the heavy lifting. An additive inference almost always needs an “all” or “none” link, because those are the only restrictive enough to guarantee overlap. Pure “some” and “most + some” chains fail.
  • Recycle inferences. Once you derive a new link, it becomes a track you can use in further chaining.
  • A one-direction inference still counts. If the chain works traveling one way, the inference holds even though the reverse path would not have produced it.

Worked micro-examples

  1. A --m--> B and B → C. Most-arrow then all-arrow, both leading away from the shared middle B. Weakest link is “most.” Inference: A --m--> C. (Most A are C.)

  2. A --some-- B and B → ¬C. Some-link then none-link. Weakest link is “some.” Endpoint C is negated by the none-link. Inference: A --some-- ¬C. (Some A are not C.)

  3. B <--m-- A --m--> C. Two most-arrows leading away from shared middle A. Two majorities of A must overlap. Inference: B --some-- C. (Some B are C.)

  4. A --m--> B and C --m--> B. Two most-arrows pointing into the shared term B. The arrows do not lead away from the middle, so no inference. This is the trap version of case 3.


9. Quick reference card

  • Diagram S → N. Contrapositive: reverse and negate, ¬N → ¬S.
  • Mistaken Reversal N → S and Mistaken Negation ¬S → ¬N are both invalid.
  • “Only / only if” introduces the necessary condition.
  • “Unless / except / until / without”: that term is necessary; negate the other for the sufficient.
  • “If and only if” is a biconditional, arrows both ways.
  • Contrapositive flips AND and OR. Split A→(B and C) and (A or B)→C; do not split A→(B or C) or (A and B)→C.
  • Chain on shared terms. A→B+A→C does not give B→C. Conditionals describe the group, not individuals.
  • Some and none reverse; most and all do not. Most reverses only to “some”; all reverses to “some” only if existence is separately established.
  • Most + most (arrows leading away) ⟹ some. Some + all and most + all work when the all-arrow leads away. All + all ⟹ all. None + anything ⟹ an inference with negativity.
  • Quantifier chaining: ride “some” and “none” either way, follow “most” and “all” arrows only; the result takes the weakest link; only endpoint negations survive.