In 1969, Seymour Papert invented the first educational robot called a Turtle. It was an addition to the computer language Logo, which he'd designed in 1965 specifically for educating children. Papert did not simply invent some technology, he offered a revolutionary way of educating children. He gave teachers practical tools to realise constructionist developmental theories in the classroom. We will show that Papert's work forms a Kuhnian Paradigm which has endured for nearly 50 years and provides the foundation for all work with educational robots. The use of educational robots in special needs education was one of many benefits that grew out of the resulting environment. The early robots designs didn't pay attention the needs of this area of education. So early researchers used the available robots and started to ask and seek answers to relevant questions. We analyse this historical research and report on their findings. We find modern research simply confirms the original work. We will introduce the Papert Paradigm and show how it empathised with the changing attitudes towards special needs education. We look at a deepening understanding of the technology provided by the Educational Robot Application Principles. And by combining this information with the Universal Design for Learning ideas we find a set of guidelines to help create better robots for special needs education.