Instead of considering only the current element to determine prefixes and formatting, DITA2Go also looks up one level in the current topic in your DITA document. If the parent element does not have text of its own, and if the current element is the first child of that parent, DITA2Go processes the parent to determine what its format would be if it did have text. This process also identifies any applicable prefix. Then DITA2Go processes the current element, and determines its format; however, a prefix has already been identified from the parent. That parent prefix becomes the first prefix for the current element.
For example, if your document uses the <note> element as a wrapper for other text elements, <note> itself might have no text content, and therefore DITA2Go would not automatically look for an output format for <note>. However, if another text element is nested in <note>, DITA2Go checks <note> for possible format prefixes to apply to the nested element.