The aim of this pilot study was to identify the most common speech sound confusions of 5 Norwegian cochlear implanted post-lingually deafened adults. We played recorded nonwords, aCa, iCi and bVb, to our informants, asked them to repeat what they heard, recorded their repetitions and transcribed these phonetically. We arranged the collected data in confusion matrices to find the most common and most uncommon speech sound confusions. We found that the voiced and unvoiced consonants are seldom confused. We also found that there was a higher rate of consonant confusion for the iCi words than for the aCa words. The most frequent confusion was [engma] perceived as [n], [m] perceived as [n] and [upsilon] perceived as [n]. For the consonants, manner of articulation was rarely confused, but place of articulation was often confused. An exception from this was the confusion of [l] and [n], which differs only in manner of articulation. The latter is in accordance with reports we get from clinicians. We postulate that this is caused by the speech processing of the cochlear implant. We found less confusion of the vowels, which can be explained by the fact that vowels have much higher energy and longer duration than most of the consonants. The most frequent confusion was [a] perceived as [ash] and [u] perceived as [openo]. [e], [i] and [openo] were never confused with other vowels. (Contains 5 tables and 4 figures.)