4.3.1 Assumptions and Conventions

Before we continue, note that when we gave Figure 20, we did not specify how we computed the data since text can have upper- and lower-case letters, accents (e.g. ``coöperation''), punctuation, and other non-letters. Here are the assumptions and conventions we used:

• Map upper-case letters to lower-case letters.
• Substitute a blank space '-' for any character that is not a letter.
• Assume no accents.