Is Łukasiewicz logic only three-valued?

No. The family includes finite-valued systems and an infinite-valued form over [0,1].

Is strong conjunction ordinary multiplication?

No. Its standard formula is max(0,x+y-1).

Does an implication value of 1 prove a causal relationship?

No. It is a formal truth-functional evaluation under an assignment, not evidence that one event causes another.

AI Glossary / Graphs and symbolic AI / Intermediate

Łukasiewicz logic

Definition

Łukasiewicz logic is a family of many-valued logics associated with Jan Łukasiewicz. Instead of restricting every proposition to true or false, these systems allow additional truth values and define precise operations for combining them. The infinite-valued form uses values throughout the interval from 0 to 1; finite-valued forms use specified finite sets of values.

What it means

The numbers are truth degrees under the chosen semantics. They are not automatically probabilities, confidence scores, or percentages of people who agree. Assigning 0.7 to a proposition requires a defined interpretation in the application.
The logic also distinguishes weak and strong versions of conjunction and disjunction. They agree on ordinary Boolean inputs but can differ on intermediate values. Consequently, writing a vague instruction such as combine these conditions with AND is insufficient: the intended connective must be specified. The logic is a formal reasoning framework, not a rule that all uncertain AI outputs should be averaged.

How it works

In the standard infinite-valued semantics, negation maps x to 1-x. Implication from x to y has value min(1, 1-x+y). Weak conjunction is min(x,y), while strong conjunction is max(0,x+y-1). Weak disjunction is max(x,y), while strong disjunction is min(1,x+y).
Evaluate a compound formula by applying these operations according to its structure. A formula's value under one assignment differs from validity under every allowed assignment. In the standard formulation discussed here, 1 is the designated value for satisfaction. Choosing the truth-value set and connectives is part of specifying which logic is being used.

Three practical examples

A logic tutorial can compare Boolean, three-valued, and infinite-valued evaluations of the same formula.
A research prototype can express graded constraints using explicitly selected Łukasiewicz connectives and test the resulting formulas.
An AI explanation interface can distinguish a rule's formal truth degree from an estimated probability, preventing two different interpretations from being silently merged.

Think of it this way

Think of a precisely defined scoring language with rules for how scores combine. Knowing the input scores is not enough; the combination rule determines the result. The analogy is limited because a logic includes formal notions of satisfaction and inference, whereas an ordinary scorecard may have no logical semantics at all.

Worked example

Consider an original demonstration assigning p = 0.8 and q = 0.6 in the infinite-valued system. Negation gives not-p = 0.2. The implication p implies q evaluates to min(1, 1-0.8+0.6) = 0.8. Reversing it gives q implies p = min(1, 1-0.6+0.8) = 1.
Weak conjunction gives min(0.8,0.6) = 0.6, while strong conjunction gives max(0,0.8+0.6-1) = 0.4. Weak disjunction gives 0.8, while strong disjunction gives 1. These are different operators, not competing calculations of the same unspecified connective.
Now examine repeated use of p. Weak conjunction of p with itself remains 0.8. Strong conjunction of p with itself is 0.6. Repeating a premise under the strong connective therefore need not leave its degree unchanged. A software implementation that removes duplicates as though every conjunction were idempotent could change a formula's meaning.
For another contrast, let r = 0.5. Then not-r is also 0.5. Weak disjunction of r with not-r is 0.5, whereas their strong disjunction is 1. It is essential to state which disjunction appears when discussing a many-valued version of a familiar classical formula.
Finally, set p = 1 and q = 0. Both conjunctions become 0, both disjunctions become 1, and p implies q becomes 0. The familiar Boolean behavior is recovered on these endpoint inputs. This arithmetic demonstrates the operations; it does not establish that the chosen input degrees are appropriate for any real application.

Illustrative case study

Fictional case: an educational software team builds an interactive rule calculator. One developer implements AND using minimum, while another uses max(0,x+y-1). Their screens disagree on intermediate inputs but match on true and false examples.
The team labels the operators as weak and strong conjunction, adds intermediate-value tests, and explains why a value such as 0.6 is not being presented as a 60% event probability. The revised lesson makes the mathematical choice visible rather than treating one implementation as a mysterious rounding error.

Common misconceptions and mistakes

Do not equate truth degrees with probabilities without an explicit additional model. Do not substitute product multiplication for the strong conjunction: 0.8*0.6 = 0.48, which differs from the Łukasiewicz result 0.4.
Do not assume every classical simplification remains valid for every connective at intermediate degrees. In particular, strong conjunction is not generally idempotent. Also distinguish evaluating a formula for one assignment from proving it valid. A result of 1 on one example does not show that the formula takes value 1 under all assignments.

Business use

The concept is useful when evaluating research or software that claims to reason with graded rules. It provides a precise vocabulary for asking which mathematical semantics are implemented.
Document the allowed truth values, connective definitions, and interpretation of input degrees. Test intermediate cases as well as Boolean endpoints. If business users see a score, explain whether it represents a formal satisfaction degree, an empirical probability, or a separate operational measure. A correct logical calculation cannot compensate for an unclear meaning assigned to the inputs.

Check your understanding

Why do the two AND-like operators give 0.6 and 0.4 when p = 0.8 and q = 0.6?

Answer: Weak conjunction uses minimum, while strong conjunction uses max(0,p+q-1). Their definitions differ on intermediate values.

Related terms

Fuzzy logic
T-norm
Negation
Satisfiability
Many-valued logic
Truth degree
Logical implication
Boolean logic
Idempotence

Learn next

Use the three-valued set {0, 0.5, 1} and evaluate implication for all nine ordered input pairs. Check that p implies p always has value 1. Then find an intermediate assignment where strong conjunction of p with itself differs from p.

Sources

Primary source: Bofill, Manyà, Vidal, and Villaret, New complexity results for Łukasiewicz logic, preliminary definitions of finite and infinite-valued systems: https://link.springer.com/article/10.1007/s00500-018-3365-9 . Numerical demonstrations and the educational case are original.

Previous: Ugly duckling theorem Next: Ultimate sampling unit