Well, you can't include someone else's name on a paper if they don't want it to be included. And their acceptance must be affirmative, i.e., there must be no doubt that they are OK with that. Sending them an ultimatum won't be enough if you can't prove that they read it (for example, you can't be sure some e-mail was delivered and read). It might even become a legal case, and laws might be different in your country/state, since in some cases it's required that an action must be done signaling that they are OK – simply not answering the ultimatum isn't enough to say that they agreed.
You can't take authorship for something that you didn't write. So if someone wrote a paragraph, you can't use it literally, but you could rewrite it.
Results, au contraire, are public to who was informed about it, and that person can do anything they want about it. If someone paints a wall blue, anyone can write a paper about the wall becoming blue. If someone tells you "I painted the wall blue and 53% of the viewers said it was better", you can write about that. Of course, you can't say that you did the painting and observation. The exception to that would be some non-disclosure agreement, where you explicitly had to agree that you won't disclose the information you are about to receive.
It's expected that you offer the co-authorship to everyone that was involved in the analysis and conclusions of the research. For example, a laboratory technician that just measured some variable and gave you a table with values won't be a co-author. But if the person isn't interested in doing so, well, use the result and publish your paper.
Also found this piece of interesting information:
But until the results are so widely known and familiar that they have become common knowledge, people who use them are obliged to recognize the discoverer by means of citations