Within quantitative methods, it is possible to give a precise, technical answer to whom may question the validity of our results. We could refer to statistical procedures of hypothesis testing and model building, delimiting what we mean be “significance”. Following Ronald Fisher, any however small or peripheral result is significant as long as it’s a sign of a real phenomenon, a verified deviation from randomness. Quantitative epistemology is based on such a definition of significance, and evolved by developing concepts such as reproducibility and experiment design. A parallel quest for a shared and precise definition of significance was pursued all along the 20th century, within the humanities and the social sciences. Defining “the meaning of meaning” would permit the development of a toolkit, comparable to the statistical one, for qualitative research. However, up to today, there is still no formal consensus about what constitutes a significant result in qualitative analysis. Since a few years, the inquiry into formal qualitative significance seems to have been considered inconclusive, and the discipline that was constituted to find an answer to this problem – semiotics – is unpopular in those very fields from which it was born (literary studies, anthropology, linguistics, communication research). Today, digital methods for the study of culture and communication bring these inquiries again on the agenda: researchers have a large battery of algorithms on the one hand and no general guidelines about how to use them for interpretative ends on the other. In this talk we will discuss the potential of computational support to communication research, as a means to blur the threshold between quantity and quality. We will suggest that a general semiotic toolkit can serve as a complement to statistics for the analysis and the interpretation of human and social data.