For the Arabic and Spanish data, two translation tiers are created.

The first is a phrase level translation equivalent, designed to give non-native investigators an insight into idiomatic linguistic constructions. The phrase-level boundaries are preserved to the extent possible, but in some cases, it is impracticable to provide a translation equivalent in this way.

The second translation tier is a rough word-level gloss, whose interval boundaries correspond exactly to those of the original orthography word tier.

The following is an example of some phrase intervals that had to be collapsed into a single interval to yield a coherent translation equivalent, while the word boundaries were preserved: