Indo-European Phonetics
Palatalalized & Labiovelar Stops

Chart of phonemic correspondences
IE *k' *g' *g'h *kw *gw *gwh
Celtic k g k kw, p (5) b g
Italic k g g, h (2) qu, p (4) gu, v, b (4) f, v
Venetic k g h kw    
Hellenic k g kh p, t, k (3) b, d, g (3) ph, kh, th (3)
Albanian s (1) z (1) z (1) k, s g, z g, z
Thracian s (1) z (1) z (1) k, kh g, k g
Illyrian s z z      
Armenian s c dz kh k g
Phrygian s, k z, g g, k k b g
Germanic h k g / [g] (7) hw kw g / [w] (7)
Slavic s z z k g g
Baltic š z' z' k g g
Indic ç j h (6) k c' g gh
Iranian s z z k c' g g
Anatolian k g g kw kw kw
Tocharian k k k k, kw k, kw k, kw

Transcription of symbols:

š - in different languages [sh] sound
ç - in Indic languages a soft [sh] sound
z' - a Balto-Slavic sound pronounced like in French jour
c' - in different languages a [ch] sound


1. IE *k', *g', *g'h originally reflected in Balkan languages as spirants [þ] and [ð], later in Albanian turning into s, z.
2. IE *g'h > Latin h or g (depending on their location in the word) and Osco-Umbrian *kh > h.
3. IE *kw, *gw, *gwh could have three different reflexes in Greek:
        > t, d, th before e, é, i (IE *kwis > Greek tis)
        > k, g, kh before u (IE *wl.kwos > Greek lupos)
        > p, b, ph before a, o (IE *sekw- > Greek hep-)
    However, this distribution was shaped in Classical Greek dialects, like Attic and Doric, while in Mycenaean Greek *kw remained, and in Aeolian it was transformed into p (which allows some linguits to speak about its close ties to the Osco-Umbrian branch of Italic tongues).
4. IE *kw and *gw are represented in two different ways in two Italic subgroups: qu, gu in Latin (*kwis > quis), and p, b in Osco-Umbrian (*kwis > pis).
5. Within the Celtic group, there are in fact two branches called p-Celtic and q-Celtic depending on the reflection of IE *kw: Brythonic and Lepontic languages are considered to be p-Celtic, while Goidelic and Celtiberian are q-Celtic; Gaulish had different dialects reflecting it both ways (IE *ekwo- > eqos, epos)
6. IE *g'h > Common Indo-Iranian *j'h > Indic h, Iranian z.
7.  It seems that Germanic reflexes of Indo-European voiced stops had spirant allophones which are evident in Gothic in the intervocal position.


1. IE *k'
*k'm.tom 'hundred'
*swek'ur- 'husband's father'

2. IE *g'
*arg'- 'shining, white'
*reg'- 'to direct, to rule'

3. IE *g'h
*weg'h- 'to carry, to drive'
*g'heim- 'winter'

4. IE *kw
*ekwo- 'horse'
*nokw- 'dark, night'

5. IE *gw
*gwen- 'woman'
*gwou- 'bull'

6. IE *gwh
*sneigwh- 'snow'
*gwhen- 'to kill'

Linguistic comment

One of the most complicated and arguable questions of the Indo-European studies is the question of dialectal groups which appeared right after the Proto-Indo-European community was broken. It is evident that two or three major dialectal groups were formed, and they existed for some time before in their turn split into separate language groups. In the nineteenth century linguists made clear distinction between Indo-European dialects basing on the reflection of palatalized consonants in various IE tongues. This is how 'centum' and 'satem' languages were defined: 'centum' ones (Celtic, Germanic, Italic, Greek, Tocharian, Venetic, Anatolian) demonstrated k, g instead of Indo-European *k', *g', while 'satem' languages (Baltic, Slavic, Armenian, Balkan, Indic, Iranian) turned palatals into sibilants: s, z or [sh], [zh].

However, it seems today that 'satem' languages were quite unstable in this feature, and sometimes this transformation or palatals does not exactly occur. Such unstable languages include ancient Balkan tongues (Illyrian, Thracian, Phrygian, Albanian) and the Balto-Slavic languages. It is even more obvious that this distinction between 'centum' and 'satem' did not reflect any Proto-dialects of the Indo-European language. The very status of the palatalized consonants is not quite clear at all: there is no historical language of the family which preserved any of them, and this makes the reconstruction dubious sometimes.

Labiovelar consonants were more viable, as far as we can see: some traces of their pronunciation is still seen in Italic languages and in Irish Ogham inscriptions (where the sound [kw] was marked by the special symbol Q like in maqi 'of a son'). In ancient Germanic languages, Gothic first of all, the labiovelar phonemes q and hw were also considered single sounds, written as single letters. A lot of other languages of the Indo-European family simply lost the labial component of this sound, turning it into k. Another interesting process clearly seen in Celtic, Osco-Umbrian and Aeolian Greek where *kw > p. This trend is usually subject to careful studies of possible close contacts of p-Celtic, Osco-Umbrian and Aeolian dialects which could have place in the Balkan region in the second millennium BC. It was even more interesting to discover the same p in Hittite as well in a few pronominal forms (pippid 'something, someone', directly cognate to Latin quisquid 'something').

Today there is no living language in the Indo-European family where original palatalized or labiovelar sounds are still used.