Groucho Marxism

A language family is a group of languages related through descent from a common ancestor, referred to as the proto-language. The Indo-European languages form a particularly large language family comprising of languages native to most of Europe, the Iranian plateau, and the northern Indian subcontinent. The proto-language of the Indo-European languages is referred to as Proto-Indo-European, or PIE. This language is not directly attested but has been reconstructed by linguists based on the attested daughter languages. Languages of the Indo-European family are classified as either ‘centum’ or ‘satem’ according to how the velar consonants of PIE – sounds produced using back part of the tongue and the roof of the mouth, or velum – subsequently developed.

PIE is usually reconstructed with three types of velar consonants: palatovelars, pronounced using the front of the velum; plain velars, pronounced using the back of the velum; and labiovelars, pronounced with concomitant lip-rounding. This traditional model implies that in the centum languages – Anatolian, Celtic, Germanic, Greek, Italic, and Tocharian – the palatovelars merged with the plain velars, whereas in the satem languages – Albanian, Armenian, Balto-Slavic and Indo-Iranian – the plain velars merged with the labiovelars. However the existence of all three types of velar has long been a source of controversy, and many scholars, going all the way back to Antoine Meillet in 1894, have suspected that there were in fact only two types of velar consonant in PIE.

There are several arguments in favour of the two-velar hypothesis. First, the plain velars are rarer than the other two types, are almost entirely absent from affixes, and appear most often in certain phonological environments. This suggests they were in complementary distribution with the palatovelars or labiovelars (or both). Second, it is extremely rare cross-linguistically for palatovelars to move backwards in the mouth, with the opposite process, known as palatalization, being much more common. Yet the traditional reconstruction implies that this ‘de-palatalization’ occurred on seven separate occasions: in Anatolian, Tocharian, Greek, Germanic, Italic, and Celtic.

Third, in the satem languages, where the palatovelars supposedly remained intact, other palatalizations also occurred, implying that palatalization was a general trend. This suggests it is not necessary to reconstruct a separate palatovelar series for PIE. And fourth, alternations between palatovelars and plain velars and are common across different satem languages, with the same root appearing with a palatovelar in some languages and a plain velar in others, or even a palatovelar and a plain velar within the same language. These alternations, referred to as ‘gutturalwechsel’, are consistent with the analogical generalization of one velar type in an originally alternating paradigm but difficult to explain otherwise.

One position where these alternations are particularly common is after so-called ‘mobile *s’. This refers to the phenomenon whereby a PIE root appears to begin with an *s which is sometimes but not always present (the * just means we are dealing with a reconstructed form). There are many examples of roots with a mobile *s followed by a velar where the root yielded reflexes in the Satem languages beginning either with *s plus a plan velar, corresponding to the form with the *s, or with a palatovelar, corresponding to the form without the *s. This is consistent with the hypothesis that there were only two velar types in PIE, a plain velar and a labiovelar, with the ‘palatovelars’ being the result of palatalization of plain velars in the satem languages which was blocked after *s.

In his 1973 PhD thesis, the linguist Lars Steensland demonstrated a complimentary distribution between the three types of velar. He first showed that in word-initial position, only plain velars occur after *s, except before *i where only palatovelars occur, which backs up the argument set out in the paragraph above. Steensland then showed that in word-initial position not following *s, palatovelars occur everywhere apart from before *r, and *s; plain velars occur everywhere apart from before *e, *i, and what we would now call *H₁; and labiovelars only occur before *e, *i, *r, and what we would now call *H₁. Thus, in word-initial position, there isn’t a single environment where all three types of velar occur.

One explanation for this distribution runs as follows. PIE originally had two types of velar: a plain velar and a labiovelar. In the satem languages, plain velars became palatalized in word-initial position in all environments apart from (1) after *s, unless followed by *i; and (2) before *r and *s. In the centum languages, labiovelars were delabialized in word-initial position in all environments apart from before *e, *i, *r, and *H₁. The problem with this explanation is that it implies that the latter development occurred separately in the centum languages on four separate occasions. This seems rather implausible and removes one of the key arguments in favour of the two-velar reconstruction. (The satem languages were contiguous so could plausibly have undergone such parallel changes.)

Nonetheless, I believe Steensland was on the right lines with his approach – mainly because it involved looking at the actual data. We probably just need to modify his theory slightly to allow for different developments in the centum languages. I will take this up in a future blog post.