We present an unsupervised linguistically-based approach to discourse relations recognition,\nwhich uses publicly available resources like manually annotated corpora (Discourse Graph\nBank, Penn Discourse TreeBank, RST-DT), as well as empirically derived data from “causally”\nannotated lexica like LCS, to produce a rule-based algorithm. In our approach we use\nthe subdivision of Discourse Relations into four subsets – CONTRAST, CAUSE, CONDITION,\nELABORATION, proposed by [1] in their paper where they report results obtained with a\nmachine-learning approach from a similar experiment against which we compare our results.\nOur approach is fully symbolic and is partially derived from the system called GETARUNS,\nfor text understanding, adapted to a specific task: recognition of Discourse Causal Relations\nin free text. We show that in order to achieve better accuracy both in the general task and in\nthe specific one, semantic information needs to be used besides syntactic structural information.\nOur approach outperforms results reported in previous papers