Few words on the LSAT cause more quiet damage than “some,” “most,” and “all.” They look like ordinary English, and that is the problem: in everyday speech “some” suggests “not many,” and “most” suggests “nearly all.” On the LSAT each of those words has a precise logical meaning, and wrong answers are engineered to exploit the gap between the two.
Master six or seven quantifiers and a surprising number of wrong answers become obviously wrong the moment you read them. This post lays out the exact meaning of each word, what you can and cannot conclude from combining them, and the traps that appear again and again.
| Word | LSAT meaning | Common mistake |
|---|---|---|
| Some | At least one, possibly all | Reading it as “not all” or “a few” |
| Most | More than half, possibly all | Reading it as “nearly all” or “not all” |
| All / every / each | 100%, no exceptions | Treating it as “almost all” |
| None / no | 0%, no exceptions | Treating it as “hardly any” |
| Not all | At least one is not, possibly none are | Reading it as “most are not” |
| Few / a few | Often treated as “some,” check the passage | Assuming it means “fewer than half” |
Suppose a stimulus states: Some athletes are vegetarians. Which of these follows?
(A) Some athletes are not vegetarians.
(B) Some vegetarians are athletes.
(C) Most athletes are vegetarians.
Only (B) follows. “Some A are B” is reversible: if at least one athlete is a vegetarian, then at least one vegetarian is an athlete. But (A) does not follow, because the “some athletes” in the original may in fact be all athletes. And (C) does not follow, because one athlete is not more than half of them.
“Most” is stronger than “some,” but it is weaker than most people assume. “Most A are B” guarantees that more than half of A are B, and it guarantees one extra thing: some A are B. It tells you nothing about B’s other members, and it is not reversible. “Most dogs are friendly” does not mean most friendly things are dogs.
This is the single most tested quantifier inference, and it is subtler than it looks. If most A are B and most A are C, then at least some A are both B and C. Both groups hold more than half of A, so they cannot be completely separate. The overlap is guaranteed even though its size is unknown.
The trap: if most A are B and most B are C, you cannot conclude most A are C. Or even that any A is C. The B’s that are C may all be outside A. Chains that work for “all” do not carry over to “most.”
| Given | Valid conclusion | Invalid conclusion |
|---|---|---|
| All A are B, all B are C | All A are C | All C are A |
| Most A are B, most A are C | Some A are both B and C | Most A are both B and C |
| Most A are B, most B are C | Nothing certain about A and C | Most A are C |
| All A are B, most B are C | Nothing certain about A and C | Some A are C |
| Some A are B, all B are C | Some A are C | All A are C |
When a universal claim meets a “some” claim, the universal claim carries the “some” through. If all A are B and some A are C, then some B are C. The reason is concrete: the A’s that are C are also B’s, since every A is a B. This is one of the few inferences that lets you create a new “some” statement from two older ones.
Reverse the roles, though, and it fails. Some A are B and all B are C gives some A are C, but all A are B combined with some C are B gives you nothing about A and C.
These are the natural home of quantifier logic. The correct answer is usually the one that stays inside what the quantifiers permit, and the wrong answers upgrade “some” to “most” or “most” to “all.” For the broader skill of staying inside what the text supports, see our article on Most Strongly Supported questions.
A classic flaw confuses “some” with “all,” or treats a reversible relationship as one-directional. Many flaw answers described as “overgeneralization” or “improperly infers a universal from a particular” are quantifier errors in disguise.
When the argument moves from “some” to “most” or from “most” to “all,” the missing assumption bridges that exact gap. If the conclusion says “most” and the premises say “some,” you need to find the answer that supplies the extra weight.
1. Circle every quantifier word as you read, including “few,” “many,” “rarely,” “usually,” and “typically.”
2. Translate into a number range. Some = 1 to all. Most = over half to all. All = exactly 100%. None = exactly 0%.
3. Ask what the range allows. If an answer choice requires a narrower range than the stimulus guarantees, it is not supported.
Many, often, usually, typically, and generally are softer than “most.” They describe tendencies, not proportions, and a stimulus that uses them cannot support a hard numerical claim. The reverse also holds: “rarely” and “seldom” signal small proportions without guaranteeing any particular fraction. Treat them as hedges and match answers that hedge in the same direction.