Several times recently, I saw discussions of a certain way in which dynamic characterisation of IF protagonists can take place. The idea is this: the game keeps track of several variables that describe the psychology of the protagonist. Actions early in the game have an effect on those variables, such that, say, stealing a purse will decrease your trustworthiness but increase your ruthlessness, while solving a puzzle through violence will increase your "violent" variable. Then, later in the game, certain actions will be made available or unavailable to you based on the value of these variables. If you have been very violent throughout the game, then you are allowed to be (or are required to be) violent at the end. Thus, a personality is established and reinforced through play, which would purportedly bring us to new heights of characterisation.
I very much doubt that it would.
Let me quickly note the explanatory barrenness of the psychology that is presented here to us. Could there be a more unconvincing explanation of somebody's acting violent than the statement that he has often acted violent before? (Compare: "This stone falls because all previous stones have fallen as well." Explanation and prediction have been confused.) And the explanation hardly becomes better when we introduce the notion of a "character trait" called "violent" and say "This person is violent, and that is why he has acted violently in the past and will act violently now."
But worse, for its use in interactive fiction, than the explanatory barrenness of this psychology is its existential and therefore artistic barrenness.
We must deal with the consequences of our past actions, and therein lie possibilities for profound existential tragedy. ("Because I have been violent in the past, this girl has no father anymore. How do I cope with that?") We must even deal with the consequences of things we have not done but have been subjected to, and therein lie even more possibilities for profound existential tragedy. ("Because I was an orphan in the slums of Mexico City, I had to rely on violence to survive. Things got out of hand, and now I am on death row. How can the world cope with that, and how can I?") But there just are no moments in our lives where we must face the fact that we cannot take a certain action because we have created in ourselves the wrong character traits. "If only I had been more courageous earlier on in my life, I now would have the courage to join the Resistance--but alas!" That is not a possible fact that we have to face; that is simply bad faith (and I totally mean this in the Sartrean sense).
Quite in general, our possible actions are not constrained by our "character traits"; that would be to make ourselves into things rather than persons. An interactive fiction that constrains my possible actions by only allowing me to do things which are like the things I did earlier cannot but reinforce the myth that we are determined beings, and that we therefore do not have to try to change. Can we seriously believe that good characterisations in interactive fiction will be achieved by denying one of the most fundamental facts of what it means to be a person? I do not think so.
Let me end by saying that anything that can be used to reinforce a myth can also--by subtle reversals--be used to cast doubt on it. So there might be some subversive uses of the technique described here, and they might be worth experimenting with. But as for the general usefulness of character traits--I doubt whether anything good will come from it.