EN
The aim of my paper is the discussion of the three hypothesis concerning Proto-Indo-European consonant system from the typological point of view: (1) Glottalic hypothesis – with excessively complicated consonant model: glottalized voiceless stops – voiceless stops (aspirated / non-aspirated) – voiced stops (aspirated / non-aspirated) and not clear distribution of its facultative variants (aspirated / non-aspirated). (2) Simplified model of the guttural consonants (two or even one series) in the presence of traditional three series (velars – labiovelars – palatovelars) owing to its recessive character. (3) The laryngeal theory, now generally accepted – its contemporary version with three consonantal H, which in certain position underwent vocalization. It seems more probable that it was a sonant ə/ə̭ (parallel to i / i̭, u / ṷ), which as sonant ə had changed in Indo-Iranian into i, whereas in other languages into short a. While as glide ə it had been transformed into laryngeal attested in Hittite as h̬ (h̬h̬), whereas in other Indo-European languages it disappeared as linear phone.