Four types that (to me) should be "the same" (although they are different types of course) needs to be initialized differently, why?

Why.swift

struct SA<T> {
    let v: T
    init(_ v: T) { self.v = v }
}

typealias SB = SA<(Int, Int)>
typealias SC = SA<W>
typealias W = (Int, Int)

struct SD {
    let v: (Int, Int)
    init(_ v: (Int, Int)) { self.v = v }
}

let a = SA(1, 2)   // Works with ZERO OR MORE parens around 1, 2 (not counting arg list parens).
let b = SB(1, 2)   // Works ONLY WITH ZERO parens.
let c = SC((1, 2)) // Works with with ONE OR MORE parens.
let d = SD((1, 2)) // Works with with ONE OR MORE parens.

// To me, the types of a, b, c and d are as similar as different types can be.
// Yet they are very different in the way they can be initialized, why?
// How come SB and SC must be initialized differently for example?
//
// (Xcode 7.1 beta 3)