Allow to assign more than one tier to a sequence.

[LinguList]

Our current tier-annotation works in such a way that a sound class or a number of sounds indicated by [] brackets are given a default tier of '' (empty string), which is the Segments tiere. In fact, the same behaviour can be reached when writing @[a b c] and parsing it.

What is missing by now is to allow to specify multiple tiers for just one position.

In the self-context, this is not a direct problem, as we can adjust the regex first and later analyze the whole regex into parts.

For the other aspects, we would need to adjust the parser, since it would need to produce the output:

>>> get_context("@initial[p t k]@[b] [p t k] _")
[ 
  [ 
    [ "p", "t", "k"],
    [ "b"]
  ],
  [
    ["p", "t", "k"]
  ]
], 
[ [ "initial", ""], [""] ]

[LinguList]

So we have a nested list here, that represents ALL tiers in a given position.

[LinguList]

The Segments_Left_2 position are two tiers, the Segments_1_Left position is only one.

[LinguList]

Not sure if this is needed. Tier combination is also increasing complexity, of course.