An experiment was conducted to investigate the process that new information was being set into prior knowledge through learning of unknown words. Participants were presented twelve unknown words of Uchinaaguchi (Okinawan dialect of Japanese language) and learned their meanings by reading corresponded Japanese common words and example sentences (in four sessions). Half of participants were presented single example sentence for each word repeatedly three times (1-sentence condition) and the others were presented three different example sentences for each word (3-sentence condition). After each learning session, priming tasks (lexical decisions with Uchinaaguchi words as primes and common words as targets) were performed. Priming effects for semantically related (SR) word pairs were found only in 1-sentence condition, which indicated that repeating same context could strengthen the links between old and new information in the semantic network. However, results of recognition tests, which followed the last learning session, showed inhibitory effects on RTs in both 1- and 3-sentence condition. The results implied that there were some effects of learning even in 3-sentence condition.