Language as a Geometry in Wittgenstein"s Tractatus

In TLP 4.011, while admitting that propositions expressed
by the phonetic notation, or the alphabet, just like the
written notes of a piece of music, do not seem at first sight
to be pictures of what they represent, the Tractatus insists
that those "sign-languages" (that is, the phonetic notation
and the written musical notes) prove to be pictures of what
they represent (that is, our speech and the piece of music,
respectively) "even in the ordinary sense". (TLP 4.016 also
says that "alphabetic script developed out of [hieroglyphic
script] without losing what was essential to depiction".) So,
contrary to the view of some commentators (e.g. Pears
1987, 115-121), instead of making an analogy here, the
Tractatus holds that a proposition is a picture literally. How
can a proposition be a picture literally?