With these rules, I keep getting the space characters grouped into the text of the WORD tokens. text_line : SPACE* WORD (SPACE | WORD | DIRECTIVE)* ( NEWLINE | EOF ) ; WORD : ~( DIRECTIVE_IND | SPACE | NEWLINE ) TEXT_CHARS* ; DIRECTIVE : DIRECTIVE_IND TEXT_CHARS+ ; SPACE : ( ' ' | ' ' )+ ; NEWLINE : ' '? ' ' ; fragment DIRECTIVE_IND : '@' ; fragment TEXT_CHARS : ( ~( ' ' | ' ' | ' ' | ' ' ) )+ ; This tokenizes "abcd 1234" to: "abcd" (WORD) " 1234" (WORD) Why not ? "abcd" (WORD) " " (SPACE   More>>

			

			text_line : SPACE* WORD (SPACE | WORD | DIRECTIVE)* ( NEWLINE | EOF ) ;
			

			

			WORD  : ~( DIRECTIVE_IND | SPACE | NEWLINE ) TEXT_CHARS* ;
			

			

			DIRECTIVE : DIRECTIVE_IND TEXT_CHARS+ ;
			

			

			SPACE  : ( ' ' | ' ' )+ ;
			

			

			NEWLINE  : '
			

			'? '
			

			' ;
			

			

			fragment DIRECTIVE_IND : '@' ;
			

			

			fragment TEXT_CHARS : ( ~( ' ' | ' ' | '
			

			' | '
			

			' ) )+ ;