@Lei12: you may have a wrong conception about style usage. According to what you say, your styles describe the appearance of your text (font face and size, colour, line spacing, indents, …) which means they may be used for semantically different things.
You should consider style names (paragraph, character, frame, page, list all the like) as a mark up to convey supplementary information (helping you as an author). With this approach, styles denote the “value” or significance of paragraphs or sequences. You assert that such paragraph is plain text (Text Body), such one is a heading (Heading n), such one is a comment (custom style to be created as Comment), such clause is “important”. Words may need an emphasis (Emphasis) or be written in a foreign language (custom style Foreign), or is an acronym (create also a dedicated style to disable spell-checking on it), … Note that I never describe a style as “red” or “bigger”, …
Then your styles have a single semantic usage, meaning that sequences so styled should always be presented the same way. Usages don’t overlap. Then you can safely modify/update any style to get the expected layout. And you can dramatically change the look of your document without reassigning style names because you are semantically consistent.
A tip: styles can be organised hierarchically where attributes not explicitly overridden are inherited from the ancestor. And when font size is at stake, you can specify it in %-units. You only change the font size in the “master style” (the ancestor of the hierarchy) and all dependent styles are automatically adjusted to follow the percentage constraint. A single parameter to change and the whole document is updated.