diff --git a/spring-expression/src/main/java/org/springframework/expression/spel/standard/InternalSpelExpressionParser.java b/spring-expression/src/main/java/org/springframework/expression/spel/standard/InternalSpelExpressionParser.java
index cf135d08ed..8d2c0d2a3e 100644
--- a/spring-expression/src/main/java/org/springframework/expression/spel/standard/InternalSpelExpressionParser.java
+++ b/spring-expression/src/main/java/org/springframework/expression/spel/standard/InternalSpelExpressionParser.java
@@ -966,7 +966,7 @@ class InternalSpelExpressionParser extends TemplateAwareExpressionParser {
 		if (desiredTokenKind == TokenKind.IDENTIFIER) {
 			// Might be one of the textual forms of the operators (e.g. NE for != ) -
 			// in which case we can treat it as an identifier. The list is represented here:
-			// Tokenizer.alternativeOperatorNames and those ones are in order in the TokenKind enum.
+			// Tokenizer.ALTERNATIVE_OPERATOR_NAMES and those ones are in order in the TokenKind enum.
 			if (t.kind.ordinal() >= TokenKind.DIV.ordinal() && t.kind.ordinal() <= TokenKind.NOT.ordinal() &&
 					t.data != null) {
 				// if t.data were null, we'd know it wasn't the textual form, it was the symbol form
diff --git a/spring-expression/src/main/java/org/springframework/expression/spel/standard/Tokenizer.java b/spring-expression/src/main/java/org/springframework/expression/spel/standard/Tokenizer.java
index 5ca7b3c30c..8d80fc8a33 100644
--- a/spring-expression/src/main/java/org/springframework/expression/spel/standard/Tokenizer.java
+++ b/spring-expression/src/main/java/org/springframework/expression/spel/standard/Tokenizer.java
@@ -35,7 +35,14 @@ import org.springframework.expression.spel.SpelParseException;
  */
 class Tokenizer {
 
-	// If this gets changed, it must remain sorted...
+	/**
+	 * Alternative textual operator names which must match enum constant names
+	 * in {@link TokenKind}.
+	 * <p>Note that {@code AND} and {@code OR} are also alternative textual
+	 * names, but they are handled later in {@link InternalSpelExpressionParser}.
+	 * <p>If this list gets changed, it must remain sorted since we use it with
+	 * {@link Arrays#binarySearch(Object[], Object)}.
+	 */
 	private static final String[] ALTERNATIVE_OPERATOR_NAMES =
 			{"DIV", "EQ", "GE", "GT", "LE", "LT", "MOD", "NE", "NOT"};
 
@@ -448,8 +455,8 @@ class Tokenizer {
 		char[] subarray = subarray(start, this.pos);
 
 		// Check if this is the alternative (textual) representation of an operator (see
-		// alternativeOperatorNames)
-		if ((this.pos - start) == 2 || (this.pos - start) == 3) {
+		// ALTERNATIVE_OPERATOR_NAMES).
+		if (subarray.length == 2 || subarray.length == 3) {
 			String asString = new String(subarray).toUpperCase();
 			int idx = Arrays.binarySearch(ALTERNATIVE_OPERATOR_NAMES, asString);
 			if (idx >= 0) {