Arguments in favour of only two series of velars include:
Palatovelars appear to be generally allophones resulting from the neutralisation of the other two series in specific phonetic circumstances. Their dialectal articulation was probably constrained, either to an especial phonetic environment (such as the Romance evolution of Latin k before e and i), or to the analogy of alternating phonetic forms.
However, it is difficult to pinpoint exactly what the circumstances of the allophony are, although it is generally accepted that neutralisation occurred after *s and *u, and often before *r or *a; also apparently before *m and *n in some Baltic dialects. The original allophonic distinction was disturbed when labiovelars were merged with plain velars. This produced a new phonemic distinction between palatal and plain velars, with an unpredictable alternation between palatal and plain velars in related forms of some roots (those originally with plain velars) but not others (those originally with labiovelars). Subsequent analogical processes generalised either the plain or palatal consonant in all forms of a particular root. Those roots where the plain consonant was generalised are those traditionally reconstructed as having plain velars in the parent language, in contrast to palatovelars.
The reconstructed palatovelars and plain velars appear mostly in complementary distributions, what supports their explanation as allophones of the same phonemes. Meillet (1902) established the contexts in which there are only velars: before *a, *r, and after *s, *u; while Georgiev (1966) clarified that the palatalisation of velars had happened before *e, *i, *i̯, and before liquid or nasal or *u̯ + e, i, offering statistical data supporting his conclusions. The presence of palatalised velar before o is thus explained as analogical, appearing in roots in which (due to the ablaut) the velar phoneme is found before e and o, so the alternation *kje/*ko was levelled to *kje/*kjo.
There is residual evidence in the so-called satem languages of a former distinction between velar and labiovelar consonants:
· In Sanskrit and Balto-Slavic, in some environments, resonants become *iR after plain velars but *uR after labiovelars.
· In Armenian, *kw seems to be in some cases distinguishable from *k before front vowels.
· In Albanian, *kw and *gw have distinct outputs from *k and *g before front vowels.
This evidence shows that the labiovelar series was distinct from the plain velar series in Late Proto-Indo-European, and could not have been a secondary development in the centum languages. However, it says nothing about the palatovelar vs. plain velar series.
When this debate initially arose, the concept of a phoneme and its historical emergence was not clearly understood, and as a result it was often claimed (and sometimes is still claimed) that evidence of three-way velar distinction in the history of a particular Indo-European language indicates that this distinction must be reconstructed for the parent language. This is theoretically unsound, as it overlooks the possibility of a secondary origin for the distinction.
The palatovelar hypothesis would support an evolution *kj → *k of centum dialects, i.e. a move of palatovelars to back consonants, which is clearly against the general tendency of velars to move forward its articulation and palatalise in these environments. A trend of this kind is unparallelled and therefore typologically a priori unlikely (although not impossible), and needs further assumptions to be made.
The plain velar series is statistically rarer than the other two in a PIE lexicon reconstructed with three series; it appears in words entirely absent from affixes, and most of them are of a phonetic shape that could have inhibited palatalisation.
Common examples include:
· *i̯ug-óm ‘yoke’, cf. Hitt. iukan, Gk. zdugón, Skt. źugá-, Lat. iugum, OCS igo, Goth. juk.
· *ghosti- ‘guest, stranger’, cf. Lat. hostis, Goth. gasts, OCS gostĭ.
According to Clackson (2007), “The paradigm of the word for ‘yoke’ could have shown a palatalising environment only in the vocative *yug-e, which is unlikely ever to have been in common usage, and the word for ‘stranger’ *ghosti- only ever appears with the vocalism o.”
Alternations between plain velars and palatals are common in a number of roots across different satem languages, where the same root appears with a palatal in some languages but a plain velar in others.
This is consistent with the analogical generalisation of one or another consonant in an originally alternating paradigm, but difficult to explain otherwise:
· *ak-/ok- ‘sharp’, cf. Lith. akúotas, OCS ostrŭ, OInd. asrís, Arm. aseln, but Lith. asrùs.
· *akmon- ‘stone’, cf. Lith. akmuõ, OCS kamŭ, OInd. áśma, but Lith. âsmens.
· *keu- ‘shine’, cf. Lith. kiáune, Russ. kuna, OInd. svas, Arm. sukh.
· *bhleg- ‘shine’, cf. OInd. bhárgas, Lith. balgans, OCS blagŭ, but Ltv. blâzt.
· *gherdh- ‘enclose’, cf. OInd. gṛhá, Av. gərəda, Lith. gardas, OCS gradŭ, Lith. zardas, Ltv. zârdas.
· *su̯ekros ‘father-in-law’, cf. OCS svekrŭ, OInd. śvaśru.
· *peku- ‘stock animal’, cf. OLith. pẽkus, Skt. paśu-, Av. pasu-.
· *kleus- ‘hear’, cf. Skt. śrus, OCS slušatĭ, Lith. kláusiu.
It could be argued, as does Clackson (2007), that “such forms could be taken to reflect the fact that Baltic is geographically peripheral to the satem languages and consequently did not participate in the palatalisation to the same degree as other languages.”
There are different pairs of satemised and non-satemised velars found within the same language.
The old argument proposed by Brugmann (and later copied in many dictionaries) about “centum loans” is not tenable today. For more on this, see Szemerényi (1978) Mayrhofer (1952), or Bernabé (1971). Examples include:
· *selg- ‘throw’, cf. OInd. sṛjáti, sargas.
· *kau/keu- ‘shout’, cf. Lith. kaukti, OCS kui̯ati, Russ. sova, OInd. kauti, suka-.
· *kleu-, ‘hear’, Lith. klausýti, slove, OCS slovo; OInd. karnas, sruti, srósati, śrnóti, sravas.
· *leuk-, ‘light’, OInd. rokás, ruśant-.
The number and periods of satemisation trends reconstructed for the different branches are not coincident (see above §3.4.1. Indo-Iranian evolution and §4.13.1. Balto-Slavic evolution).
In most attested languages which present aspirates as a result of the so-called palatovelars, the palatalisation of other phonemes is also attested (e.g. palatalisation of labiovelars before e, i), which may indicate that there is an old trend to palatalise all possible sounds, of which the palatalisation of velars is the oldest attested result.
It is generally believed that satemisation could have started as a late dialectal ‘wave’, which eventually affected almost all PIE dialectal groups. The origin is probably to be found in velars followed by e, i, even though alternating forms like *gen/gon caused natural analogical corrections within each dialect, which obscures still more the original situation. Thus, non-satemised forms in so-called satem languages would be non-satemised remains of the original situation, just as Spanish has feliz and not ˟heliz, or fácil and not ˟hácil, or French facile and nature, and not ˟fêle or ˟nûre as one should expect from its phonetic evolution.
Contrasting with the idea of an areal centum-satem distinction is the existence of satem languages like Armenian, related to Greek, a centum one; or Balto-Slavic, a North-West Indo-European language; as well as the presence of Tocharian, a centum dialect, in Central Asia, a satem territory; and Albanian, a satem language in the Balkans, a centum territory.
The traditional explanation of a three-way dorsal split requires that all centum languages share a common innovation that eliminated the palatovelar series, due to the a priori unlikely move of palatovelars to back consonants (see above). Unlike for the satem languages, however, there is no evidence of any areal connection among the centum languages, and in fact there is evidence against such a connection – the centum languages are geographically non-contiguous.
Furthermore, if such an areal innovation happened, we would expect to see some dialect differences in its implementation (cf. the above differences between Balto-Slavic and Indo-Iranian), and residual evidence of a distinct palatalised series. However, neither type of evidence exists, suggesting that there was never a palatovelar series in the centum languages. Evidence does, however, exist for a distinct labiovelar series in the satem languages (see above).
External evidence shows a conspicuous absence of reconstructed palatovelars in Uralic loanwords of Late Proto-Indo-European origin (Holopainen 2018), with only later dialectal borrowings—of Indo-Iranian or Balto-Slavic origin—displaying clear phonetic correspondances to palatalised velars.
A system of two gutturals, velars and labiovelars, is a linguistic anomaly, isolated in the Indo-European occlusive subsystem—there are no parallel oppositions bw-b, pw-p, tw-t, dw-d, etc. Only one feature, their pronunciation with an accompanying rounding of the lips, helps distinguish them from each other. Such a system has been attested in some ancient Indo-European languages. A system of three gutturals—palatovelars, velars and labiovelars—with a threefold distinction isolated in the occlusive system, is still less likely.
In the two-dorsal system, labiovelars turn into velars before *-u, and there are some neutralisation positions which help identify labiovelars and velars. Also, in some contexts (e.g. before *-i, *-e) velars tend to move forward its articulation and eventually palatalise. Both trends led eventually to centum and satem dialectalisation.