Assembling a corpus of information for the 'Dico Culturel' (Cultural Dictionary) project: Reliable and unreliable sources

In the framework of the “Dico Culturel” research project, we aim to describe and analyze French words and expressions which are rich in cultural and anthropological contents and establish linguistic and inter-cultural parallels with the Spanish used in Colombia. Towards that end, we are going to construct a corpus which allows for cross-comparisons between these two languages, drawn from authentic documents available on the internet. But this approach raises a question. It is how to know whether that corpus shall only consist of information found in reliable sources or also include the use of unreliable ones. To clear up this doubt, we studied a sample of 19 words and expressions in the French language and on that basis, created two corpora in French: one of information drawn from reliable sources and the other from unreliable sources. Then, with the use of a program of textual analysis, we looked at and compared the frequencies, concordances and linguistic and cultural contents of the words and expressions in the sample of each corpus in order to determine their quality and relevance. After analyzing the contents, we concluded that in many cases there were no differences between the corpus of reliable sources and the corpus of unreliable sources. In addition, we noticed that many data complemented each other, since they were found in one or the other of the two corpora, which led us to confirm that the information found in the unreliable sources may be of interest for our research, without excluding their verification in the reliable sources.