For me the conversations still feel jarring and robotic because the hearthlings have to stop everything else they are doing to stand around and talk. I feel like the game really could use a casual conversation animation combination with other activities.
What I mean is conversations that hearthlings can have while crafting or hauling that can be indicated by a simple turn of the head towards who they are talking to and the occasional wave, eye contact, nod, hand gestures, stomping a foot, or shrug of the shoulders. I feel like this should be the majority of all conversations, especially in a busy play-through.
This means that those instances when they forgo work to talk make the subject matter and intensity of the conversation that much more compelling and important for that individual. In other words the personality of a hearthling may dictate how one of them acts because they feel it is more important than the other one.
So you could have one hearthling stopping what they are doing to follow another one around and give them a piece of their mind while they just nod and keep on because they don’t value the conversation nearly as much. So a goat talking about what he ate to a food snob, the food snob just wants to get out of there but the goat keeps standing ahead blocking his path to force the conversation and this affects their relationship and mood. The same idea could apply to many different aspects of the game.
I think visually it would be a lot more obvious what is going on, even with just the thought bubble because you can see the way they feel by how they interact. A hearthling that feels more passionately about a subject will use a lot of hand gestures, one that does not will almost completely ignore you and continue what they are doing or stomp their foot in disagreement.
This means that for most of these conversations, if you are paying attention, you don’t really even need the end emotional conversation bubble to figure out how the conversation went. Limiting emotional responses to actions instead of facial expressions means they need to be kind of generic, but this also makes them universally understood.
I know making these things happen is a lot of work so that the timing of the animation is appropriate, transition seamlessly and interact properly with the correct response.
All in all, I hope conversations will continue to be improved.