A non-linear text is a directed graph where each vertex is labeled with a string. In this paper, we introduce the longest common substring/subsequence problems on non-linear texts. Firstly, we present an algorithm to compute the longest common substring of non-linear texts G 1 and G 2 in O(|E 1||E 2|) time and O(|V 1||V 2|) space, when at least one of G 1 and G 2 is acyclic. Here, V i and E i are the sets of vertices and arcs of input non-linear text G i, respectively, for 1≤i≤2. Secondly, we present algorithms to compute the longest common subsequence of G 1 and G 2 in O(|E 1||E 2|) time and O(|V 1||V 2|) space, when both G 1 and G 2 are acyclic, and in O(|E 1||E 2|+|V 1||V 2| log |Σ|) time and O(|V 1||V 2|) space if G 1 and/or G 2 are cyclic, where, Σ denotes the alphabet.