Connecting degree and polarity: An artificial language learning study