This paper deals with the question concerning the reconstruction of the third series of stop consonants for Proto-Indo-European, the so-called “voiced aspirates”. The traditionally reconstructed system of stops assumes a triple opposition voiceless : voiced : voiced aspirates. Since the Seventies, as a matter of fact, such a system has been discussed, because it shows, from a cross-linguistic perspective, a marked asymmetry: cross-linguistically voiced aspirates hardly ever exist without the corresponding voiceless aspirates (§ 1-2). Within the “glottalic theory” an alternative reconstruction has been proposed: this theory assumes a typologically acceptable system of stops but, on the other hand, it raises several phonetic problems. Since some scholars have proposed that the 228 Proto-Indo-European voiced aspirates could have been breathy voiced, in this work the acoustic-articulatory features (§ 3), the cross-linguistic diffusion and the possible diachronic developments (§ 4) of the breathy voiced stops have been pointed out and analysed. In fact, the assumption of a Proto-Indo-European system of stop consonants voiceless : voiced : breathy voiced could solve the problem of typological plausibility that affects the traditional reconstruction, since if in a tripartite consonantic system the voiced aspirates are not fully voiced the system is not typologically marked (§ 5).