I think that the important thing to keep in mind when conceptualizing what Wolfram|Alpha is is that it addresses questions that are computational or data based. So, the things that the parser understands, generally speaking, are queries that imply some sort of computation or the retrieval of particular items of data, and then perhaps (usually) some computations based on those data. This might be information about a work--its definition, but also various statistical and linguistic information on that word, some of which is computed based on a combination of the data sources. One might argue that possible answers to the sorts of questions that you gave could fall into these categories of analysis--for example by returning relationships between various psychological and developmental theories, or outlining the historical progression of ideas in a field. But these sorts of things would require the curation of a very different set of classes of data (and many of which are either discursive or opinion oriented in many cases) than the sorts of explicitly factual data that are currently curated as inputs into the computations to answer computationally oriented queries.