Whistled speech is a form of modified speech where some frequencies of vowels and consonants are augmented and transposed to whistling, modifying the timbre and the construction of each phoneme. These transformations cause only some elements of the signal to be intelligible for naive listeners, which, according to previous studies, includes vowel recognition. Here, we analyze naive listeners’ capacities for whistled consonant categorization for four consonants:/p/,//,/t/and/s/by presenting the findings of two behavioral experiments. Though both experiments measure whistled consonant categorization, we used modified frequencieslowered with a phase vocoder-of the whistled stimuli in the second experiment to better identify the relative nature of pitch cues employed in this process. Results show that participants obtained approximately 50% of correct responses (when chance is at 25%). These findings show specific consonant preferences for “s” and “t” over “” and “p”, specifically when stimuli is unmodified. Previous research on whistled consonants systems has often opposed “s” and “t” to “” and “p”, due to their strong pitch modulations. The preference for these two consonants underlines the importance of these cues in phoneme processing.