...After listing 'unnatural dialogue' as a criticism in some form 3/4 times in my last reaction thread update...I decided it might be a good time to try to make another thread explaining what that means and how you can avoid it. I mean last time my advice amounted to 'just copy people', and although I don't think it was necessarily bad advice, clearly something a little more in-depth is needed here.
So usually when I say a story has 'unnatural dialogue', I'm saying
(a) the characters don't sound like people,
(b) they all sound like the same person, or
(c) all of the above.
(a) is the easiest issue to remedy, in my opinion. Because we all know what people sound like, even if we don't think we do. For instance, I doubt many people would mistake the speech of their GPS for that of a real person-- the delivery, tone, pace, and latent emotion is clearly different. These are the qualities that distinguish 'real' speech from 'robotic' speech.
Unfortunately, a lot of writers, even those who seem to have a decent grasp of English, tend to get stuck somewhere in the middle. They can maybe pick up on tone (choosing when to use a comma, a period, or an exclamation mark) and pace (using 'can't' and 'don't' instead of 'cannot' and 'do not' in casual language) but unique delivery and latent emotion are rarely present. None of the words have emphasis; nothing about a character's speech changes with their emotions. Leaving the reader with the impression of a mere 'facsimile' of impactful dialogue...it's the uncanny valley of a character voice.
Now (b) is where things can get tricky...a lot of the time when budding writers DO finally figure out what a character voice is, they end up just using that one voice for everyone in the story. They don't seem to know how to make different characters speak differently, outside of literally having them say different lines.
Or, when they DO try to copy voices from other media, they don't know how to modify them to make them feel unique to their own characters-- I've always found it pretty obvious to tell when 'shonen MC' or 'shoujo MC' voices are just lazily shoved into someone's story, spouting generic anime lines and reactions in the exact same slightly unnatural phrasing that localization can't help but produce...
First of all, your story is coming from you, the original source, so it shouldn't sound 'localized'. Second of all, I don't want to read a story about a reskinned Naruto Uzumaki or Tohru Honda, I want to read a story about YOUR character. Yes, there should be a difference.
What is that difference, though? What makes a human voice; what makes a unique voice? The answer is personality.
Personality will make vastly different voices all sound equally believable and real. Personality will make voices with the exact same word choice sound distinct. The key is to understand your character, and learn how to include that understanding in their dialogue.
Here are some examples of character voices with different personalities. And if you'd like to gauge the distinctness of the character voices in your story, an exercise like this is a good way to do it: write down a brief anecdote in the voice of each character, and see what you come up with.
~~
A: I saw a really cute dog on the way home today. His owner let me pet him! I think his name was Marbles...
B: I saw this dog on my way home today...it was pretty cute. The owner told me its name was Marbles or something...
C: On my way home today I saw THE most adorable dog in the history of dogs! He had spots all over him and little droopy ears and a cute little nose...and his name was Marbles!! I begged the owner to let me pet him; I just couldn't help myself~!
D: I got to pet a cute dog in the park today.
E: On my way home today, I met a lady walking with the sweetest little pup on a leash. She told me its name was Marbles, and allowed me to stroke its nose. It was truly the highlight of my evening.
F: I ran into someone walking their dog in the park today-- he was so sweet, and very well-trained. The dog, I mean...he even let me give him a few scritches under the ears~. You would have loved to meet him.
~~
Hopefully you can tell that, although all these voices are describing the same event, they are not the same. Some descriptions are longer and more detailed, some are short and to the point. Some include tons of exclamation marks, some don't have any. Some include the dog's name and some don't; one describes the owner, and one even describes the listener, to some extent. These inclusions/exclusions point to the unique perspective of each character.
For example: C is clearly the voice of a cheerful, enthusiastic character. But E, despite being much more formal and antiquated, also has similar notes of enthusiasm, thanks to the use of hyperbole ("the sweetest little pup", "the highlight of my evening"). Without a single exclamation point, you can make it clear that Voice E also enjoyed the experience very much.
To give another example: the voices that gender the dog tend to be the more approachable ones. ^^ A, C, and F are all friendlier voices, while B, D, and E are more reserved. Describing a specific gender, especially when talking about an animal or an inanimate object, feels more personal and informal, and will lend that quality to the speaker. Think about how different it feels when someone calls a boat "she" versus "it". ^^
Also, note that D in particular doesn't even give the dog a pronoun, because their description is so succinct. Perhaps the positive experience of getting to pet a fuzzy animal is the focus of their dialogue, and nothing else really mattered to them.
Now, there were some challenges I ran into while preparing these examples: in the beginning, I thought I could give each voice similar phrasing, and have them use a lot of the same words...this quickly turned out to be impossible. ^^; The characters were all too different...the subtle reluctance of Voice B, for example, made this obvious. Unlike Voice A, they wouldn't come right out and call the dog cute; they'd add it as an afterthought. They'd say its name was Marbles "or something", as if it wasn't all that important to them...despite the fact that they clearly took the time to talk to the person with the dog and learn its name, instead of just walking by. ^^
This is clearly the specific voice of a character who wants to appear cool and unbothered, so they wouldn't say things the same way the others would...and that's okay! If it doesn't feel right to make a character use the same word choice or even word order as another character, that's a good sign that they have a distinct personality.
Another issue I ran into was all the characters who didn't make the list: they had unique voices too, but I just couldn't imagine many of the evil villains or self-absorbed pricks or hardened warriors talking about the experience of meeting a cute dog. Either they wouldn't care, they'd rather kick a dog than pet one, or they wouldn't see the point in sharing something like that.
That's okay too! If some of your characters would not talk about a certain experience, or better yet, would simply be unable to have a certain experience due to the way they interact with the world, that's ALSO a good sign that they have a distinct personality!
One last thing to note: when I say 'unique' and 'distinct', I don't mean 100%. Some of these voices could potentially be interchanged: for example, Voice A could say what Voice D said, and Voice F could say what Voice B said. Friendlier characters can easily give less-friendly, more reserved lines in different contexts, and some reserved characters can give less-reserved, more enthusiastic lines in different contexts.
The keyword there is context: something would have to change about that person's mental state or conversation partner to bring out that different line. Not only is this okay, it's vital to think about if you want the characters' emotional changes to feel believable. If you're trying this exercise, try to figure out what it would take to make one character sound like another.
For A to give D's line, I think they'd just have to feel shy: maybe they're with someone they don't feel comfortable with yet, so they don't want to go into detail about things. And for D to give A's line, I think they'd actually have to be younger; a version of themselves that hadn't yet gone through what their present-day character has gone through, and still has that little bit of innocent enthusiasm.
For F to give B's line, I think they'd have to be tired: that intentional 'hesitance' would become an unintentional delay, simply due to lack of energy. They'd say "Marbles or something" not because they want to be aloof, but because they're actually having trouble remembering.
And for B to give F's line, I think they'd have to be with someone they felt particularly safe with. B has a well-established hidden soft side, and their speech could definitely come out in that gentle way under the right circumstances.
And there you have it. ^^ If you struggle with dialogue, or if you just want to get better at it, I hope this little tutorial was helpful in some way. As always, feel free to ask any questions or add your own observations.