406 réponses à ce sujet

[Il Divo]

 

 

 

I can't "know" anything, if by "know" you mean actually be certain that it is true. But I can have a very reasonable basis to believe things that is quite often right and not often right enough that it allows me to achieve things in the world. And it can certainly be a reliable indicator.

It's possible to size someone up in seconds after meeting them. In fact, our basic social cognitive processes are designed to do exactly that. Prediction people's behaviour - and in fact predicting their actions with some degree of reliability - pretty basic to our function. 

 

 

Gotta agree with this. In every day interaction, there are a number of different actions which we don't expect to see because they're socially unacceptable. And our lack of perfect knowledge in predicting human behavior doesn't stop us from abiding by those social cues. They're critical to understanding what is and isn't tolerated.

 

How often do you see someone blatantly crapping in the middle of the street? Or murdering random passers-by "just because"? Hell, for anyone making this very claim, what would you expect to happen as a natural consequence?

[AlanC9]

People are arguing that seeing text and then hearing it will be annoying and disturbing to their experience. OK. I would think a good demonstration of that point, if it's true for more than just a small number of people who should just be ignored   (yes that was rhetorical) , is find me some people saying that those who are seeing & then hearing text in DX:HR is annoying or frustrating represent a large portion of the game's player base. I bet you don't have that data. When I asked you the last time, you responded with the anecdote of your own experience, which you know just as well as I do doesn't answer your questions above. 

I wouldn't be likely to register as a data point under this methodology myself. Unless someone's trying to do a straight-up DX:HR/ME comparison, I wouldn't be motivated enough to talk about the DX:HR system being slightly suboptimal.

[Sylvius the Mad]

When I see full lines of text I see zero lines that are what I want to say.

This is an important difference.  I don't typically want to say anything in particular.  There are things I can say (while maintaining character coherence), and things I cannot say (without breaking character coherence), but there's nothing I specifically want to say.  So it's never the case that there's nothing there I want.  I don't want anything.  So dialogue selection is merely a process of finding the possible ways to deliver the lines available until I find some that don't break my character.

 

I wasn't expressing a value judgment. Sorry for being offensive.

 

What's the difference between a difference and a disability, anyway? Depending on what diagnostic test you use I'm either just inept at recognizing faces or outright prosopagnosic. Meaning that what you can't do with tones, I can't do with faces. I actually do tolerably well on the Cambridge, though failing to recognize Barack Obama and Stephen Colbert, among many others, was a bit shocking. But on unknown faces I'm abominable - fifth-percentile or so. How bad does it have to be to be a disability? Or is the term so squishy that it should never be used?

Calling it a disability suggests that at one end of the continuum there is function, while at the other end there is a lack of function.  Calling it a difference says only that the functions are relevantly dissimilar.

[Sylvius the Mad]

The game purports to simulate, as closely as possible, real human interactions. That's as much a part of the setting as the basic assumption of causality. You can certainly close- your mind to the possibility that the game-world responses to any rules that purported to say that effects follow from causes, but that doesn't mean that the game-world responds to cause-effect.

I still think that the silent protagonist games mimicked real world conversations just about perfectly.

 

Our disagreement lies primarily in our difference of opinion about how the real world works, and the level of information we have regarding it.

 

 

In terms of the falsification, two problems: (i) falsification is on pretty shaky ground, logically (because it presupposes you can isolate hypotheses from assumptions, which is false);

Falsification presupposes no such thing.  It is often described so, but there is no such requirement.

 

We don't falsify hypotheses.  We falsify sets of propositions, which include both assumptions and hypotheses.

 

 

(ii) the actual data has an embedded problem of perception. Our subjective experiences can't count as evidence alone. There needs to be something more to the methodology. And when we measure how people react to speech, we can tease out their reaction to tone. Moreover, language (and writing, particularly in novels) reflects a structure that privileges presentation and content. All of this suggesting that the how matters.

Suggesting, yes.  But unless we can describe the mechanism by which this occurs, we cannot reasonably conclude that it does.

[Sylvius the Mad]

Firstly, when you said "'I'm claiming that there is no reaction I could have which would cause you to think that you'd said the line differently" I thought you were using the general "you" (in the sense that we couldn't imagine any reaction that would lead me, the speaker, to think I'd said the line differently). So I acknowledge that it wouldn't really be your reaction, and it wasn't my intent to suggest it.

I didn't think you were.  Honestly, your approach to language is so alien to me that I often don't foresee your interpretations.

 

Which, I think, only serves to reinforce my point.  We don't all use language the same way, and as such we cannot reasonably expect to be able to predict the reactions of someone else.

 

 

Secondly, your non-response would absolutely make me thing I failed say what I wanted. In the original example, if I made a really cheerful greeting and you didn't react (or reacted absently), I'd think you didn't get my meaning (or weren't paying attention). The intended outcome of the line wouldn't have it's effect, so that would tell me there's a mismatch and I need to try something else to get the effect I was attempting to get.

Those are two different things.  Are you saying that you would question whether you said what you meant to say, or that you would want to try to say something else to achieve the same results?

 

I'm talking about the former, not the latter.

 

 

The key thing to appreciate is that, again, it doesn't matter to me whether you actually perceive my meaning as I internally perceive it. That what is sarcasm to me is cheerfulness to you is irrelevant if - for all possible observable and practical purposes - you respond to my sarcasm as I predict and expect. It's a purely instrumental theory. To use a science analogy (which Bas Van Frassen likes quite a bit) even if all our theories about science are completely wrong, it doesn't change the fact we build very good ipods.

It's for this reason that Huygens's description of diffraction has always bothered me.  The math works, sure, but the explanation is laughable.

 

The bold is plainly wrong. The whole point of a shared culture is that we can do this, because we all internalize similar reactions to cues, stimiuli, etc. You're confusing "reliably" with "perfectly". 

It's a difference of scale.  That I can predict with some accuracy how a population will behave does not grant me the ability to predict the behaviour of an individual, particularly exceptional individuals.

[Sylvius the Mad]

As to why this matters, because it's all about what the point of dialogue is supposed to be. If I'm using dialogue to do things in the world, instead of as my opportunity to monologue at an audience, then whether or not the dialogue option I pick attempts to do what I want matters.

I'm only ever trying to monologue.  How people react to me informs my opinion of them, but changes nothing about whether I was successful in saying what I wanted to say.

 

 

It's the exact same issue as the paraphrase. You want your character to say [X]. If he says [Y], that's a problem. I want my character to say [X] in manner [G]. If I don't have any information on how manner [G] comes up, and if the NPC acts totally inconsistent with the message being delivered in manner [G], then I have a problem, because absent actual information about the tone I can't know whether the disconnect breaks my character - because actually manner [G] doesn't exist and isn't supported in-game - or because the character just doesn't react well to manner [G].

And the mistake I think you're making is in supposing that you have any idea what relationship there is between G and the NPC reaction.

 

 

Let's make this concrete. Alistair actually likes it when you reply sarcastically to his lines. It nets you approval. But if you pick a line you think is sarcastic and the actual delivery is straight, and Alistair disapproves, you might think that what Alistair does is dislike sarcasm.

My impression of Alistair was that he liked being scolded.  See?  Reasonable people can disagree about what human behaviour means.

 

 

Of course, this problem can come up IRL. But IRL you can hear yourself say the line. So you know whether you tried to say it sarcastically. In game, you have no idea.

You could in the game, as well, if only you would decide that you could.

[Sylvius the Mad]

If "imagining" is so great, then why have dialogue at all? We could just have one-word options for what the consequence will be, and you can imagine an almost infinite number of ways your character could say something. Just put it [Negotiate] and then the NPC can just react to that general proposition with you being able to imagine tone and content.

That's how I've always played the silent protagonist games.  I thought it worked quite well.

 

 

Except for the fact that the writers have already chosen that how the line is meant to be delivered.

That choice makes no material difference.  It may as well not be there.

 

 

Like I said: if imagining is so wonderful, we could just have tone indicators and then you could imagine the literal content of the line. Except, of course, that's totally unsatisfying because it removes 50% of the conversation.

I'd rather have keywords indicating topic and leave the tone entirely to the player.  Like the classic Ultima games.