Docs / Syntax / Minimalist / Feature / Interpretability

Interpretable and uninterpretable features

import Linglib.Syntax.Minimalist.Feature.Interpretability · namespace Minimalist

This file defines interpretability and the dual-feature system of Adamson & Anagnostopoulou 2025, together with the operations of that system on coordinations.

A feature is interpretable if it is read at LF and uninterpretable otherwise (Chomsky 1995). In a dual-feature system (Smith 2015) a nominal bears two sets of features of one kind: its interpretable features (iFs), which Transfer sends to LF, and its uninterpretable features (uFs), which Transfer sends to PF. A nominal with no uFs sends its iFs to PF as well, by the redundancy rule.

In a coordination, one feature set of each conjunct percolates to &P, either its iFs or its uFs. Conversion intersects the percolated iF sets, so that &P bears the iFs that every conjunct bears. Percolated uF sets are not intersected. PF realizes each of them, and realization succeeds only if every set receives the same exponent.

Main definitions

Implementation notes

A bundle is a function on Interpretability, so bundles are ordered and combined pointwise. The features of a nominal built from two layers are the join ⊔ of the layers' bundles.

References

Declarations

L66–67
L69–71
L96–97

The exponent that every set in l receives, if l is nonempty and there is one. PF realizes each percolated set separately (p. 25 of Adamson & Anagnostopoulou 2025).

def DualFeature.realizeAllL106–110
{F : Type u_1} {E : Type u_2} [DecidableEq E] (realize : Finset F → Option E) :
List (Finset F) → Option E

Used by DualFeature.realizeAll_eq_some_iff, DualFeature.realizeAll_singleton

Used in AdamsonAnagnostopoulou2025

L112–122
L124–126