I can’t agree. If you want the game engine to understand what a character is doing, there is indeed a means to achieve that. I’m posting the link below, but here’s my brief summary of how I had to approach this question before a solution finally appeared.
An interpreter is a computer program which analyses text. Structuralists will say that the text can be chopped up (lexed) and then offered as a candidate to match a defined grammar (parsed). And that’s how Human Talking works.
There is another approach, which we’ll call Phenomenology. This way has fewer names for things. It is interested in action.
A simple example. There is a French word, “Bof”. At this point I wanted to post an authoritative link from the NYT but please don’t read this article it will lead to madness.
Bof is the word you say when you shrug your shoulders. You can say Bof without the shrug. You can shrug without the Bof.
From my perspective, both these things are Speech. If you do both at the same time, you have chosen to emphasize the Speech.
As a writer of Interactive Fiction, I want to present Human Speech to the player. I also need to define the role that Speech plays in the social transaction which is the current context of the story.
The result of that thinking is SpeechMark. Please take a look at it. It’s entirely Open Source. You can use it however you want. MUD or IF, it will give you power.