Landmark discoveries in Indo-European (IE) phonology include so-called "Laws" such as "Grimm's Law." Translations of such major German landmark articles recall this period of intellectual history and are available online in Lehmann's Reader; and Allan Bomhard's Reconstructing Proto-Indo-European Phonology sketches assumptions that now underlie comparative-historical reconstruction of sounds.
Indo-European phonological issues involve constraints on the shape of reconstructed Proto-Indo-European (PIE) roots, the nature of the vowel system, syllable structure, and suprasegmental phonemes such as pitch and stress. A list of PIE phonemes may be illustrated by roots in which consonant or resonant phonemes begin or end the root; roots may be given in the e-grade form, not accounting for ablaut variants.
Still controversial is the assumption that the earliest PIE obstruent consonants contained a set of glottalic stop variants (cf. Bomhard's essay). These account, among other things, for constraints on the shape of PIE roots and for the fact that no PIE root seems to begin with /b/. Our dual listings of obstruents, with traditional PIE phonemes vs. glottalic stops, facilitates translation between traditional reconstructions of roots (e.g. *ped- 'foot') and glottalic reconstructions (e.g. *pet'- 'foot').
It is assumed, according to the Benvenistian theory of the PIE root, that the basic PIE vowel was /e/. At the earliest stage of PIE there were possibly three "laryngeal" consonants with properties of resonants; these later disappeared, leaving in their wake additional vowels /a/ and /o/. In yet later PIE, roots might begin with these vowels, although in earlier PIE they would have begun with a corresponding laryngeal. So early PIE *Hent- 'forehead; before' corresponds to later PIE *Hant- and yet-later PIE *ant-.