Abstract:
It has recently been suggested that formal features such as phi-features are merged at C (Boeckx 2003, Chomsky 2005, Kornfilt 2003, Miyagawa 2005). The idea is that only phase heads host phi-features, so v is also a possible site for formal features, but not, for example, T (Chomsky 2005, Miyagawa 2005). The phi-features at C are inherited by T in the familiar cases of subject-verb agreement (Chomsky 2005). What if a phi-feature doesn’t get inherited by T, but stays at C? Given its uninterpretable nature, it has to find a goal so it can be valued. One possibility I will explore is that it finds its “goal” in discourse. Miok Pak (2006) argues convincingly that in Korean, the promissive, the exhortative, and the imperative markings that occur in the region of C are not force markers as previously analyzed, but they are forms of agreement analyzable by the features [+/- first person, ±second person]. That is, these forms of agreement in Korean find their goal in the immediate discourse context. In this paper I will look at the politeness marker in Japanese, -desu/-masu, and show that it is a form of person agreement. The basic facts are those reported in my 1987 LI squib on wh-question and the -desu/-masu form of the predicate. What is striking is that in Basque, there is a special second-person agreement called “allocutive” which has a distribution that parallels -desu/-masu. The allocutive agreement, which has the same form as the regular second person agreement, occurs separately from the regular subject-verb agreement and it marks politeness. |