The way autistic people talk to and interact with other people has often been considered ‘wrong’ or ‘impaired’. Previous research has been focused on comparing autistic people with nonautistic people, and in particular how nonautistic people often find it hard to understand autistic ways of talking, interacting, and being. However, autism research is now increasingly being shaped by autistic people, and this research is finding that autistic people are not bad at communicating – they just have their own ‘style’ of talking and socialising, and this style is different to nonautistic people’s. In this thesis, I explore how autistic people talk and socialise, from an autistic point of view. I also look at how different kinds of factors (like social settings, sensory needs, and other people) can affect how autistic people’s experiences of talking and socialising. There are four studies that I talk about in this thesis. The first was a focus group, where autistic people discussed topics about socialising and communicating. They talked about how they naturally liked to talk and socialise, what they found challenging about talking and socialising, and what they felt was different about how they talked and socialised compared to nonautistic people. This study found that autistic people have different experiences of socialising to nonautistic people, and also different expectations about how conversations and social interactions should work. These expectations can sometimes clash with nonautistic people’s expectations and, when this happens, autistic and nonautistic people could find it hard to understand one another and get along. To stop this from happening, autistic people often felt they had to pretend to be nonautistic, which was stressful and unpleasant. However, when autistic people were talking and interacting with other autistic people (or with nonautistic people who were supportive and accommodating), they didn’t have to hide being autistic and could be themselves. These sorts of conversations and social interactions were much more enjoyable and successful than ones where participants had to hide being autistic. The second study was also a focus group. It looked at how autistic people described their experiences of socialising with and talking to other people. The participants talked about what was important to them in social interactions. They also talked about how socialising with and talking to other autistic people was different to socialising with and talking to nonautistic people. This study found that autistic people think being very honest and authentic is important in conversations. It also (like the first study) found that autistic people have different ideas to nonautistic people about what makes socialising with others enjoyable. These differences can lead to misunderstandings between autistic and nonautistic people. The third study looked at how well other people can tell if someone is autistic or not. Specifically, I wanted to know how different contexts might make it easier or harder to tell that someone is autistic. Is it easier to identify autistic people: if you’re autistic yourself, or if you’re nonautistic? If the autistic person is interacting with another autistic person, or with a nonautistic people? If you’re looking at a video of an interaction, or a photo? I found that people are generally good at guessing whether someone is autistic or not, but that all three of these contexts affected how good people are. The most important thing I found was that, when two autistic people are talking to each other, people watching the interaction find it very hard to tell that both those people are autistic. I think this might be because, when people try to guess whether someone is autistic or not, they look for problems with that person’s socialising or talking – but autistic people don’t have many problems socialising with or talking to other autistic people. This makes it difficult to tell that someone is autistic when they’re around other autistic people. I did the fourth study in a daytime support service for autistic adults with high support needs (i.e. adults who need 24/7 care, usually because they have an intellectual disability). I wanted to know how the supported people and staff there talked to one another and socialised, and also what they thought about their relationships with the other staff and supported people in the service. To do this, I spent time in the services, working alongside staff and working with the supported people, and writing down things I saw and what I thought about them. I also interviewed staff and supported people. I found that relationships in the service were nuanced and complicated. Staff and supported people often struggled to understand and relate to one another – but they also often formed strong friendships with one another, and developed unique, meaningful ways of communicating and interacting with one another. This study also emphasised how contexts (like sensory environments, social demands, physical and mental health, and past experiences) can have a big impact on whether autistic people’s social interactions are successful and enjoyable or not. By combining these studies with research from other people, I also developed a new theory. I wanted to explain why autistic people find socialising with each other easy, and nonautistic people find socialising with each other easy, but both groups find it difficult talking to and socialising with one another. I called this theory “Theory of (Un)Known Mind”. It suggests that autistic and nonautistic people have different expectations when it comes to predicting other people’s thoughts and feelings. Autistic people don’t try to predict others’ thoughts and feelings, and don’t expect other people to do so for them, either. Nonautistic people, though, think that predicting other people’s thoughts and feelings is very important in social interactions, and that it’s rude not to do this. This difference in social expectations can cause misunderstandings between autistic and nonautistic people, and can make interactions between autistic and nonautistic people feel difficult or unpleasant. Overall, I argue in this thesis that autistic and nonautistic people misunderstand one another because autistic and nonautistic people have different social styles, and different social expectations – not because autistic people are ‘bad’ at socialising and communicating, like some research claims. Because of this different social style, it’s important that researchers study autistic social interactions in general. This means research shouldn’t focus on what autistic people do ‘wrong’. Instead, it should focus on what autistic people do, how that’s the same as or different to what nonautistic people do, and why these differences can cause misunderstandings. I suggest that, by encouraging nonautistic people to understand how autistic people like to socialise and talk, we can improve social interactions and relationships between autistic and nonautistic people. (A more technical abstract is available at the linked page.)