The present paper aims to investigate the integration of data and data sources in the practice of synchronic and historical pragmatic research. Relying on two case studies in these traditionally corpus-based fields of language use, it is argued that (i) the pieces of information usually regarded as corpus data are plausible statements originated in integrated data sources; and (ii) the methods used in pragmatics have fuzzy boundaries in the actual research practice. It is concluded that even a basically corpus-based research integrates pieces of information from various data sources. Consequently, corpus as a data source cannot be considered exclusive: it is always complemented with further data sources such as the researcher’s intuition, previous research results, theoretical framework, and various inferences.