This paper reconsiders semiotic modelling in light of recent scholarship on Charles Peirce, particularly regarding his concept of proposition. Conceived in the vein of Peirce’s phenomenological categories as well as of his taxonomy of signs, semiotic modelling has mostly been thought of as ascending from simple, basic sign types to complex ones. This constitutes the backbone of most currently accepted semiotic modelling theories and entails the further acceptance of an unexamined a priori coherence between complexity of cognition and complexity of (...) signification. Following recent readings of Peirce’s post-1900 semiotic, we engage in a discussion as to what are the limits of this approach. From Stjernfelt’s conception of the dicisign in nature, we derive a perspective that affords understanding the practice of modelling as a reciprocal interplay between decomposition of complexity and recombination into further complexity. This discussion is facilitated by the recent extrapolation of the constructivist concept of scaffolding in biosemiotics research. Cognition, we argue, begins with a fundamental irritation of trying to make sense of a structure that is more complex than what can directly be derived from experience and, in so doing, urges meaning-seeking processes. The yet unknown object is decomposed into more tangible objects and is subsequently reassembled from these more manageable conceptions of the object. In support of our argument, we discuss the notions of semiotic competences and resources in light of such a naturalized account of meaning-making. (shrink)