Syntactic objects are constructed by Merge in its simplest form and are labeled by the Labeling Algorithm. This paper considers the labelability of heads and proposes a principled explanation by arguing that it is reducible to interpretability at the interfaces. This proposal removes unwanted stipulations on labeling assumed in Chomsky (2015) and deduces the labelability of T in rich agreement languages as well as that of T and V (i.e., R categorized by v*/v) when they are pair-merged to phase heads; the proposal also suggests that there are other unlabelable heads and it is shown that n is an unlabelable head. The paper reconsiders EPP, arguing that it is not attributable to labeling but to externalization. It is concluded that the discussion upholds the Strong Minimalist Thesis, the basic hypothesis in minimalism.
1. Introduction
2. Weakness of T as a Label
3. Labelable and Unlabelable Heads
4. Labelable T
5. The Labelability of Other Heads
6. Conclusion
References
(0)
(0)