When humans speak or animals vocalize, they can produce sounds that are further combined into larger sequences. The flexibility of sound combinations into larger meaningful sequences is one of the hallmarks of human language. To some extent, this has also been found in other species, like chimpanzees and birds. The current study investigates the structure of sounds when speakers are asked to communicate the meaning of 20 selected concepts without using language. Our results show that the structure of sounds between pauses is frequently limited to 1–3 sounds. This structure is less complex than when humans use their native language. The acoustic distance between sounds depends largely on the concept apart from concepts referring to animals, which show a higher diversity of involved sounds. This exploratory analysis might provide evidence of how the structure of sound could have changed from simple to complex in evolution.