Where to begin. You may start with the Preliminary first step — The Art of Grammar — if you want a foundation in the parts of speech before logic. Or you may begin with the introduction to logic, What Is Logic · An Orientation. Or you may start with the two Study sections below, which teach without keeping score. Or you may go straight into the exercises: choose a difficulty, then select an exercise.
What to do in an exercise: each correct answer earns points and each error takes away points; a set is complete at 100 points. All moods and figures of the scholastic account are possible here — including the fourth figure and the weakened (subaltern) moods.
Where schematic terms are possible, a small switch on the exercise page chooses letter or English terms — capital letters stand for general terms, lower-case letters and proper names for singular. Letters are the easier dress: finish a set in letters, and it will invite you to attempt the same in English.
Among the unscored options: The Diagram Workshop lets you mark Venn diagrams freely — shade regions, place ×s — and shows what your marks assert. The Square of Opposition sets two terms in the four traditional relations and displays any immediate inference — converse, obverse, contrapositive, contradictory — drawn from whichever corner you choose.
A general term is a universal: one word standing over a host of individuals. Picture it looking down on them — the term above, and beneath it every single thing of which it can be truly said:
Every circle in these diagrams is such a universal seen from above, its individuals gathered inside. Two moves say everything: an × reaches down and picks one individual out at random — “some dog” — without saying which; shading declares that a region holds no individuals at all. When two or three circles overlap, their regions sort the individuals beneath — and the whole logic of propositions and syllogisms can be read off the picture. Use the buttons below to see any form drawn for you, or mark the regions yourself and let the readout say what your diagram asserts.
Shading declares a region empty; an × declares that something lives there. Each categorical proposition makes exactly one mark:
“All S are P” — shade the part of S outside P (no S escapes P).
“No S are P” — shade the overlap (nothing is both).
“Some S are P” — place an × in the overlap.
“Some S are not P” — place an × in S outside P.
With three circles, a universal proposition empties a lens of two cells — one inside the third circle and one outside it. A single × placed in one cell asserts several “some”-propositions at once, as the readout below will show. In syllogism mode the readout also states what your marks force about the two outer terms — the conclusion, if one follows.
Every simple categorical proposition has a quantity — universal or particular — and a quality — affirmative or negative. Cross these two and exactly four forms result, named by the vowels of affirmo (“I affirm”) and nego (“I deny”):
A — universal affirmative, All S are P. E — universal negative, No S are P. I — particular affirmative, Some S are P. O — particular negative, Some S are not P.
The square sets these four at its corners — A and E along the top, I and O beneath — so that the logical bonds between them may be seen at a glance. Four such bonds hold:
Contradictories (the diagonals, A–O and E–I) can never share a truth-value: of each pair, exactly one is true and one false. Contraries (A–E) cannot both be true, though both may be false. Subcontraries (I–O) cannot both be false, though both may be true. Subalterns (A above I, E above O): truth descends from the universal to its particular, and falsity climbs from the particular to its universal.
Why it is used. Given the truth or falsity of any one corner, the square lets you read off at once what must follow for the other three. It is the oldest engine of valid inference from a single premise — and the ground of the immediate inferences below, which you can try out on any pair of terms in the square at the foot of this page.
Conversion exchanges subject and predicate. E and I convert simply, and the result is equivalent — truth passes both ways: “No fish are birds” ⇄ “No birds are fish”; “Some singers are poets” ⇄ “Some poets are singers”.
A converts only per accidens — the quantity drops: “All foxes are animals” yields “Some animals are foxes”. Why the drop? The A proposition places the foxes wholly inside the animals, but it says nothing about the whole of the animals — “All animals are foxes” would claim far more than was given (the illicit conversion). Yet since there are foxes (every term is non-empty on the traditional account), that part of the animals which contains them must be foxes: hence “some animals are foxes”. Truth is preserved downward only — you cannot climb back from “Some animals are foxes” to “All foxes are animals”.
O has no converse at all: “Some animals are not foxes” is true, while “Some foxes are not animals” is false.
Obversion changes the quality and replaces the predicate with its complement. It works on all four forms and is fully truth-preserving — the obverse is equivalent to the original: “All foxes are animals” ⇄ “No foxes are non-animals”; “Some poets are singers” ⇄ “Some poets are not non-singers”.
Contraposition returns, from an A or an O, an equivalent with both terms changed places and replaced by their complements — but it is really three familiar steps in succession, not a single leap: obvert, then convert, then obvert again. Take “All foxes are animals.” (1) Obvert — change the quality and negate the predicate: “No foxes are non-animals.” (2) Convert — an E converts simply: “No non-animals are foxes.” (3) Obvert again: “All non-animals are non-foxes.” That is the contrapositive, equivalent to the original (whatever fails to be an animal certainly fails to be a fox). Beware the tempting shortcut of converting first — “All animals are foxes” — for an A does not convert. E contraposes only per accidens, to an O; I has no contrapositive.
The contradictory is not a truth-preserving inference but a truth-reversing pairing: A with O, E with I, each always taking the opposite truth-value of the other.
In sum: simple conversion (E, I), obversion (all), and contraposition (A, O) are equivalences — truth passes both ways. Conversion of A and contraposition of E hold in one direction only, the universal weakening to a particular by existential import. And O cannot be converted, nor I contraposed, at all: “Some animals are not dogs” is true, yet its converse “Some dogs are not animals” is false — the O gives no purchase on the predicate beyond the part it names.