Words have to be distinguished from the Characters with which they are written. Each word could potentially be written with different Characters, but in practice most Words have just one string of one or more Characters that is used to write it. And such a string of one or more Characters on the other hand is used to write many different Words. In the TLS, such a list of Words can be seen in the attribution floater, when a character in the text is selected. To identify which of the many different possible Words is used at a specific location is the key to understand that text segment, and this understanding is made explicit by identifying it and marking it as an attribution.