This study sought to answer the question “is private speech really private?” by assessing if participants spoke more to themselves when in the company of the experimenter or when they were alone. The similarity between adults and children, with regard to production rates, and content of private speech was also investigated, as was the potential curvilinear development of private speech within the child population. Finally, there was an analysis of the relationship between private speech and accuracy conducted. Participants were asked to complete two classification tasks which were disguised as a computer game. The aim of the game was to tidy or put toy characters in their appropriate box or house. The presence of the experimenter was manipulated by the taking of a fake phone call. No relationship between private speech and task accuracy was found, but as expected adults performed significantly better than children. The curvilinear development of private speech was also not supported. However, it appeared that private speech was affected by the presence of another, with participants speaking more to themselves when in the company of the experimenter than when they were alone. Moreover, the adults and children produced the same relative amounts of private speech, with content of these utterances being similar and predominantly task relevant. A proposed hypothesis for these findings is that individuals are able to detect potentially social situations, which then engages the speech centre in the brain, thus giving rise to increased amounts of private speech. Findings also have implications with regard to education, and it is suggested that teachers should not only be aware of the phenomenon but that they should encourage it so to assist children’s self directed learning. Conclusions are drawn which imply that private speech is not confined by age and that it is actually a paradoxical term.