But software is represented as, and legally protected as, text.
The mechanisms and arguments used to protect the freedom to work with the text of the software apply to the text qua text. They are therefore available to arguments regarding the freedom to work with text in general.
Speech is encoded in text much the same as software is; both speech and code are “uses” of text. If these uses should both be free for their own reasons then the same arguments and mechanisms for supporting this freedom can be used for both.
The mechanisms that free software uses when protecting software freedom can therefore be used to protect freedom of speech. This despite the fact that software use and speech are very different acts.
Software, as the expression of functionality rather than as functionality itself, can be thought of as a form of speech, and indeed this is the conclusion that the EFF encryption free speech case reached.
Stallman does differentiate between the freedoms that he believes are needed by functional works, works of opinion and works of expression. But I am arguing that his “four freedoms” should be applied to the texts of free speech as well as to the texts of free software for different reasons, not that cultural works are functional.