We propose in our paper a semantic analysis of Romanian journalistic discourse in the second half of the 19^th century, using a computational system. The corpus analyzed consists in 63 articles with political theme, appeared in four representative publications of that time: "Timpul", "Românul", "România Liberă", and "Pressa". The contrastive research is based on computational linguistics techniques which aim, primarily, lexical-semantic investigations. Processing the corpus with a soft provides further qualitative analysis, contributing to emphasise the semantic specificity of Romanian political discourse of the time and to define elements of modernity that characterize it.