I appreciate that Crawford made the differentiation between reaction, participation and interaction. I may or may not have used reaction and interaction to mean the same things before, and now I know to do better 
. I also liked the point that he made about interactivity existing on a continuum, rather than as a binary property, which in turn allows people to approach it in a subjective manner. In the web that we know today, interactivity exists in some form or another and in varying magnitudes, from the use of simple buttons to complex reciprocal elements. And those that incorporate interactions in meaningful and interesting ways capture attention and attract user traffic. So in this way, one could argue that there now exists a certain threshold for interactivity that applications on the web have to meet to ensure the work gets to the audience.
Building on top of Crawford’s definition of interaction – a cyclic process in which two actors alternately listen (input), think (process), and speak (output) – I think a strongly interactive system could be one that has exposed to the user multiple areas for sending input, listens carefully through those channels, and once it receives input, reciprocates by sending back a meaningful response that satisfies its purpose. It is important that the “speaking” part of the cycle conforms or relates to the “listening” and that it’s not just gibberish; if I ask a local person for directions to the Eiffel Tower in French, I expect exactly what I asked for, and not, for instance, directions to the Shibuya crossing in Japanese. In other words, it is not sufficient to just have all three segments of the cycle present for a system to be strongly interactive, and the content and quality of each is just as important in ensuring user satisfaction.