Following a feature-based approach to information structure and a new approach to the definition of topics, a structure for different kinds of topics is proposed, showing that the Finnish and Hungarian left peripheries are more similar than previously thought, in line with cartographic models. The most popular definition of topics in terms of aboutness is reanalysed as a consequence of the relationship between the comment of the utterance and support entities: processing effort minimisers, thanks to the application of a principle inspired by Relevance Theory, where comments are expected to be relevant contributions to the discourse. The contrastive feature is proposed to project a head which encodes a true mathematical function, which applies to non-topics, as well. Furthermore, an interface mechanism is proposed for nanosyntax, the feature-based approach being employed.