In this chapter, we examine a number of topics that have thus far not been properly addressed. First, we discuss incremental processing in the DLM. We then reflect on the consequences of the massive variation in forms and meanings for modelling. The third section zooms in on the question of how speakers arrive at the semantic vectors that in our model drive production. Following this, we consider whether the model could be made to work for words in utterance context. We conclude with some reflections on human learning as opposed to machine learning.