This paper reports on an experiment in which two mobile robots solve the symbol grounding problem in a particular experimental setup. The robots do so by engaging in a series of so-called language games. In a language game the robots try to name an object that is in their environment first by categorizing their sensing, then by naming this categorization. When the robots fail to categorize or name, they can adapt their memory in order to succeed in the future. At the start of each experiment, the robots have no categories or names at all. These are constructed during the experiment. The paper concludes that in the investigated experiment the robots solve the symbol grounding problem as a result of the co-evolution of meaning and lexicon.