Myhill–Nerode theorem

In the theory of formal languages, the Myhill–Nerode theorem provides a necessary and sufficient condition for a language to be regular. The theorem is named for John Myhill and Anil Nerode, who proved it at the University of Chicago in 1957 .

Statement

Given a language <math>L</math>, and a pair of strings <math>x</math> and <math>y</math>, define a distinguishing extension to be a string <math>z</math> such that

exactly one of the two strings <math>xz</math> and <math>yz</math> belongs to <math>L</math>.

Define a relation <math>\sim_L</math> on strings as <math>x\; \sim_L\ y</math> if there is no distinguishing extension for <math>x</math> and <math>y</math>. It is easy to show that <math>\sim_L</math> is an equivalence relation on strings, and thus it divides the set of all strings into equivalence classes.

The Myhill–Nerode theorem states that a language <math>L</math> is regular if and only if <math>\sim_L</math> has a finite number of equivalence classes, and moreover, that this number is equal to the number of states in the minimal deterministic finite automaton (DFA) accepting <math>L</math>. Furthermore, every minimal DFA for the language is isomorphic to the canonical one .

Generally, for any language, the constructed automaton is a state automaton acceptor. However, it does not necessarily have finitely many states. The Myhill–Nerode theorem shows that finiteness is necessary and sufficient for language regularity.

Some authors refer to the <math>\sim_L</math> relation as Nerode congruence, in honor of Anil Nerode.

<!-- old proof

If <math>L</math> is a regular language, then by definition there is a DFA <math>A</math> that recognizes it, with only finitely many states. If there are <math>n</math> states, then partition the set of all finite strings into <math>n</math> subsets, where subset <math>S_i</math> is the set of strings that, when given as input to automaton <math>A</math>, cause it to end in state <math>i</math>. For every two strings <math>x</math> and <math>y</math> that belong to the same subset, and for every choice of a third string <math>z</math>, the automaton <math>A</math> reaches the same state on input <math>xz</math> as it reaches on input <math>yz</math>, and therefore must either accept both of the inputs <math>xz</math> and <math>yz</math> or reject both of them. Therefore, no string <math>z</math> can be a distinguishing extension for <math>x</math> and <math>y</math>, so they must be related by <math>\sim_L</math>. Thus, <math>S_i</math> is a subset of an equivalence class of <math>\sim_L</math>. Combining this fact with the fact that every member of one of these equivalence classes belongs to one of the sets <math>S_i</math>, this gives a surjective function from states of <math>A</math> to equivalence classes, implying that the number of equivalence classes is finite and at most <math>n</math>.

In the other direction, suppose that <math>\sim_L</math> has finitely many equivalence classes. In this case, it is possible to design a deterministic finite automaton that has one state for each equivalence class. The start state of the automaton corresponds to the equivalence class containing the empty string, and the transition function from a state <math>X</math> on input symbol <math>a</math> takes the automaton to a new state, the state corresponding to the equivalence class containing string <math>xa</math>, where <math>x</math> is an arbitrarily chosen string in the equivalence class corresponding to <math>X</math>. The definition of the Myhill–Nerode relation implies that the transition function is well-defined: no matter which representative string <math>x</math> is chosen for state <math>X</math>, the same transition function value will result. A state of this automaton is accepting if the corresponding equivalence class contains a string in <math>L</math>; in this case, again, the definition of the relation implies that all strings in the same equivalence class must also belong to <math>L</math>, for otherwise the empty string would be a distinguishing string for some pairs of strings in the class.

Thus, the existence of a finite automaton recognizing <math>L</math> implies that the Myhill–Nerode relation has a finite number of equivalence classes, at most equal to the number of states of the automaton, and the existence of a finite number of equivalence classes implies the existence of an automaton with that many states.

-->

Use and consequences

The Myhill–Nerode theorem may be used to show that a language <math>L</math> is regular by proving that the number of equivalence classes of <math>\sim_L</math> is finite. This may be done by an exhaustive case analysis in which, beginning from the empty string, distinguishing extensions are used to find additional equivalence classes until no more can be found.

For example, the language consisting of binary representations of numbers that can be divided by 3 is regular. Given two binary strings <math>x, y</math>, extending them by one digit gives <math>2x + b, 2y + b</math>, so <math>2x + b \equiv 2y + b \mod 3 </math> iff <math>x \equiv y \mod 3 </math>. Thus, <math>00</math> (or <math>11</math>), <math>01</math>, and <math>10</math> are the only distinguishing extensions, resulting in the 3 classes. The minimal automaton accepting our language would have three states corresponding to these three equivalence classes.

Another immediate corollary of the theorem is that if for a language <math>L</math> the relation <math>\sim_L</math> has infinitely many equivalence classes, it is regular. It is this corollary that is frequently used to prove that a language is not regular.

Myhill–Nerode theorem

Statement

Use and consequences

Generalizations

See also

Notes and references

Bibliography

Further reading