This article discusses language universality and language variation, and suggests that there is no feature
variation in initial syntax, featural variation arising by metamorphosis under transfer from syntax to PF-morphology. In
particular, it explores the Zero Hypothesis, stating that Universal Grammar, UG, only provides two building elements, Root Zero
and Edge Feature Zero, zero, as they are purely structural/formal elements with no semantic content in UG. Their potential content
is provided by the Concept Mine, a mind-internal but language-external department. UG and narrow syntax has access to the Concept
Mine, and this Syntax-Concept Access is unique to humans, a prerequisite for the evolution of language (
Section 1). A related idea (also in
Section 1) is coined the Generalized
Edge Feature Approach, GEFA. It states that Merge always involves at least one edge feature, which precludes symmetric structures
and enables Simplest Merge (no Pair-Merge, no Hilbert epsilon operator). The article advocates that there is no syntactic feature
selection (
Section 2), all syntactic features being universally accessible in the Concept
Mine, via Root Zero and Edge Feature Zero. In contrast, there is feature selection in PF (including morphology), yielding
variation (
Section 3), Gender being a clear example (
Section 4). However, there is a widely neglected syntax-to-PF-morphology metamorphosis (
Section 5), such that morphological features like [past] are distinct from albeit related to syntactic features like
Speech Time. Parameters operate on selected PF features, and not on purely syntactic features, so parameter setting is plausibly
closely tied to the syntax-to-PF-morphology metamorphosis (the concluding
Section 6). It is
suggested that parameters are on the externalization side of language, part of or related to the sensory-motor system,
facilitating motoric learning in language acquisition.