Celtiberians adopted two variants of the Levantine Iberian script for writing their language. This shows the cultural importance of the Iberian culture at that time, as this adoption was made when the Romans were already in Spain: this is specially valid for the Botorrita variant, created in the second half of the II century b.C. rejecting also the previous Celtiberian writing variant.

    The main differences between the Celtiberian variants are the signs used for the nasals /n/ and /m/. The Luzaga variant uses Iberian n for /m/, and Iberian accented m (m') for /n/, whereas the Botorrita variant uses Iberian n for /n/, and Iberian m for /m/ (remember that Iberian m is NOT /m/).

    As the Iberian semisyllabary isn't suitable for a Celtic language, specially as they adopted a non Catalan variant (that is, a variant that didn't denote the voiced opposition between occlusives) and as Celtiberian language had groups of 'muta cum liquida', there are many problems in converting the transcription of the inscriptions in a phonemic representation. Worth noting is the fact that whereas Iberian accented s (s') denotes a normal sibilant and the group /ks/, the Iberian s denotes a probably sibilant Celtiberian phoneme from etymological /d/ in most medial positions and in final position (as Untermann, I don't believe in the other Villar's proposal that s also denotes a voiced sibilant, but it is worth noting De Bernardo's idea that s marks a palatalized dental, although I can't accept this as the only origin). Consequently the Celtiberian sign coming from Iberian accented s is simply transcribed s (without a tilde), whereas the one coming from Iberian s is transcribed as z or as d with slight variants of 'd' in order to show its fricative value.