Can. It's by no means typical. Everything from "Zork" through "Jigsaw" follows a pre-set linear plot; either you get it or you don't get it, but there's no straying allowed. Even most ambitious or experimental IF -- "Photopia", for instance -- usually tells just one story. Or if there are multiple paths (as in the case of "I-0"), these are limited to two or three.
What I wanted to do was give the player the impression that the universe was completely open-ended, and that, rather than fumbling among a large number of options to find the one that I had chosen to implement, he could meaningfully do just about anything and get some kind of result. Obviously, it wasn't possible to achieve that: all the possible endings in Galatea are products of my brain, things that I wrote because I anticipated some combination of events that would make them appropriate. But there are 30 or 40 such endings (I've lost track; not all the endings are very distinct from each other), and many hundreds of ways to reach them. So as a story machine, the game is effective at creating scenarios that I haven't specifically thought about. Some of these are better than others, but I've done my best to make sure that anything the game produces at least makes sense. (This is far more difficult than you might imagine, and there are probably still jarring or inappropriate combinations possible, even after extensive debugging and replay.)
This should, I thought, answer (or at least contribute to) the long-running debate over whether it was possible to write true multi-linear or non-linear IF that still felt well-crafted and story-like. Something that would reward multiple playings; something that would respond to the player's own interests. Any given bit of conversation in Galatea might suggest two, three, even four or five further topics for investigation; moreover, actions like KISS, HUG, ATTACK, and SORRY take on different colorations in different contexts. The story goes vastly differently depending on what the player decides is important enough to pursue. I wanted to avoid, as much as possible, the read-the-author's-mind syndrome by abolishing puzzles entirely and leaving the plot open to maximal adjustment.
This approach also creates a special kind of freedom for the author. I had dozens of different ideas milling around when I wrote Galatea -- about art; about feminism and feminist criticism; about friendship and its abuses; about memory, expectation, and desire; about identity and self-definition; and (of course) about interactive fiction and NPC design. Any short story I could have written would have been torn apart by material so diverse. Even a branching, Choose-Your-Own-Adventure kind of presentation couldn't have covered all the ground I wanted to cover. In this medium, however, I had the freedom to put all the threads in there, leaving it up to the player to choose what interested him particularly.
The challenge then is to determine end conditions and write the endings themselves in such a way that they would satisfy the desire for coherence -- tying together and explaining things that have happened earlier, or feeling like a natural outcome of the progress of conversation. I certainly didn't do a perfect job of that, and there are ways to get stories out of the game that would be perfectly dreadful if regarded as static fiction. IF offers a little extra leeway, however: people (at least in my experience) tend to have less rigorous expectations of the plot and pacing than they would if they were dealing with a written story, and if the ending makes sense based on the outcome of the last several turns, they seem relatively content.
What I didn't anticipate was how many people would be made nervous by the multilinearity. Perhaps it's the puzzle-solving nature of IF-players, but much of my feedback has come in the form of the question, "These are the endings I found: have I got them all?"
They don't tend to be reassured by, "No, but it doesn't matter."
Which has led me to wonder. After all the questions about whether we can write real multi-linear IF, perhaps we've neglected another point: is this what players want? Many people seem badly to want to "get it right" or "see it all", and I suppose I can at least partly see their reasoning. For one thing, seeing The Whole Thing is a puzzle, and puzzles appeal to this crowd. But more than that, how do you know you're getting what the author intended you to get if you only see one tiny fragment of the whole work? This goes beyond reader-response criticism. The text IS what the player makes it; responsibility for authorship itself lies somewhere between me and the player. Possibly the player's own sense of narrative cohesion and progression are as important as anything I did in making a session of game-play satisfactory.
And doubtless there are some people who will never be happy as long as they know they've only seen part of the picture. To them I can only say, Sorry. I don't have the whole picture myself. Yes, I generated all the text; yes, I have the source at my disposal. But that doesn't mean I have anticipated everything that could come out of the game. Not really. Even knowing all the endings does not mean knowing all the different ways to reach them, all the possible spins they can carry in different contexts.