When the text interpreter processes source code, it divides the code
into blank-delimited strings, and then calls recognizers to identify
them as words, numbers, etc., until one recognizer identifies
(recognizes) the string; if the string is not recognized, the text
interpreter reports an error (undefined word).
The usual way to deal with recognizers is to just write code that one of them identifies (see Default Recognizers); however, you can also manipulate them (see Dealing with existing Recognizers) or even define new ones (see Defining Recognizers).