This is something I've struggled with a lot designing my own persuasion mechanic for Balzu. It seems that, no matter what you write, the player character's response is going to sound wrong in a lot of players' ears. The fact is, the nature of the problem is very complex.
In general, dialogue choices can be created on a scale that ranges from high textual specificity to high textual generality. There is a strong, positive correlation between high specificity and high immersion and low specificity and low immersion. For example, given a quest by the emperor, one of the player's dialogue choices might be: "You are Uriel Septim, the emperor. What choice do I have but to bend to your will?" This is an example of a highly specific textual response. A different designer might have instead provided: "[Accept]", which is a highly generic response.
As is immediately apparent, this correlation is complicated by a second factor: how well the textual response matches the player's conception of their character. A strong match between the response and the player's concept of the character will result in a significant boost to immersion. A strong mismatch will result in a significant
loss of immersion. This is the dilemma you're referring to, of course. The problem that the player expresses eloquently as "My character would never say that!" The challenge, then, is trying to decide where on that line between specificity and generality you need to place your choices.
One of the ways of getting around this is to provide stock choices (as you mentioned in your post): Do-gooder, self-interested, evil, etc. (Good, neutral, evil.) This increases your chances that highly specific text will still result in some sort of match for the player and prevent a break in immersion. The problem then shifts a little from 'where on this scale should I place the text?' to 'how many options should I provide?', though, of course, you still have to answer the first question in each case.
Alignment-based responses, unfortunately, don't address a third factor that affects the player's decision: opportunity. In the simplest sense, opportunity can be split down the middle as "Can/Can't". "Yes, I can take this quest" or "No, I can't, I'm busy." Note that this choice is different from the "Will/Won't" choice that alignment choices represent but is always a factor in the player's decision to pursue a quest. A player may want to help, but have too many plates in the air already.
Even at it's simplest level, then, an adequate range of choices will have to account for both alignment and opportunity:
- Yes, I can and I will
- No I can't but I want to
- Yes I can and I will if you reward me
- No I can't but I would if you did reward me
- Yes I can but I won't
- No I can't and I wouldn't anyway
In practice, however, even this range of choices is not going to satisfy a great number of players. For example, I often want a choice which goes something like this: "I would like to think about your request, and maybe do some background checks on the people involved before I agree to anything". Or, "I'd be interested in doing it if the reward were greater. Can we negotiate?" These kinds of options are operational choices. They act as modifiers for alignment and opportunity considerations. "I will IF...", or "I won't UNLESS...".
As you can see, the number of possible options is proliferating rapidly, and you're still only considering the answer to the simple question: "Will you help me?"! To make matters worse, the examples above have all used more or less generic responses which do not communicate character or atmosphere. They are
factual replies. Factual replies may resonate with certain players playing certain characters, but they are low on the immersion scale. In order to bump up the immersion, you have to add texture to the responses. For example: "My apologies, sir. Though I have sworn to protect the innocent and bring ne'er-do-wells to justice, I know nothing of this matter and must give it careful deliberation" depicts a certain type of character which will resonate with certain players and provide great immersion. But (ignoring game limitations which prevent textual responses of this length) how portable is this method? What if my character is a spunky, free-wheeling elf? Although my alignment, opportunity, and operational parameters may be the same ("Yes I can and I will, but I need more data") it is completely at odds with how I perceive my character. This is where the personality types in come into play (slacker, perfectionist, etc.). You could write a second response along the lines of: "Something stinks like goblin feet about this whole situation. You can bet I'm going to get to the bottom of this!" The result might be the same, but the texture of the response is completely different. Depending on the character the player is role-playing this response may be either better or worse, either increasing or decreasing immersion. The point is: You don't know, you can't know, and you never will know.
This is one of the key challenges to designing true, open world RPGs. Many RPGs side-step the entire issue entirely by choosing your character's personality for you. If you can identify with the character they give you, having resonant dialogue that is always in character can be a great aid to immersion. If you can't identify with the protagonist, however, you may not even want to play the game even though everything else about it may be fantastic.
The choice you have then, at least as I see it, is between selecting from a range of stock characters and providing the best dialogue you can (hoping that the player will identify with at least one of them) and using generic, factual responses which 'get the job done' at a small cost to immersion. My own position at the moment is to go with the generic responses. I'm replacing specific dialogue text with options like: "[Accept]", "[Postpone]", "[Ask for more information]", "[Refuse]", "[Negotiate]", etc. Which is better? I don't know. I think that depends on the designer and the objectives of the mod.
Thanks for broaching the subject. It's been rattling around in my head for weeks now. :bonk: