Groucho Marxism

In a recent blog post I argued, based on data collated by Lars Steensland in his 1973 PhD thesis, that his proposed word-initial delabialization before *o occurred separately in the western Indo-European languages rather than in Proto-Indo-European (see that post for more details). This supports the hypothesis that PIE only two types of velar consonant in this position: plain velars and labiovelars. In this blog post I will investigate whether the same can be said for other word-initial positions. Let us start with the position word-initially after *s. Steensland argued that only plain velars occur here, except before *i where only palatovelars occur. In a recent (2014) article, Robert Woodhouse criticized this view, claiming that only plain velars and labiovelars occur here. Either view supports the two-velar hypothesis.

Steensland provides two examples of words with palatovelars occurring word-initially before *H₂ (or *a in his reconstructions): *g’hH₂ensiH₂ ‘goose’ and *k’H₂eueH₂ ‘owl’. As noted by Steensland, there is a clear case of ‘gutturalwechsel’ (velar interchange) in the root from which the first word is derived, *g’hH₂ens ~ *g’heH₂ns. There is also is a clear case of gutturalwechsel in the root from which the second word is derived, *k’H₂eu- ~ *k’eH₂u- ~ *k’H₂u- ~ *k’uH₂-. These can be explained by positing that these roots originally contained a plain velar which was blocked from palatalizing before *H₂. This, along with the fact that there are no other examples in Steensland’s data of palatovelars occurring before *H₂, supports the hypothesis that only plain velars and labiovelars occurred in this position in PIE.

There are four examples of words with a plain velar word-initially before *l:*gloH₁imos ‘moist’, *gluH₁nos ‘knee’, *gloiH₁uos ‘sticky’, and *ghloumos ‘joke’.  The first two only have Albanian reflexes among the satem languages, whereas none of Steensland’s 17 examples of words with a palatovelar have Albanian reflexes, suggesting that palatalization was blocked before *l in Albanian. The latter two only have Balto-Slavic reflexes among the satem languages; but palatalization was not blocked before *l in Balto-Slavic as 11 of Steensland’s 17 examples of words with a palatovelar before *l have Balto-Slavic reflexes. However, in only two of these 11 words is the *l followed by *o, and in both cases palatalization could have been introduced analogically, suggesting that palatalization was blocked before *lo in Balto-Slavic.

Steensland demonstrates that labiovelars do not occur word-initially before *u in PIE, but both palatovelars and plain velars do occur in this position. This is consistent with the two-velar hypothesis if we assume that plain velars were palatalized in this position in the satem languages and labiovelars delabialized in this position in the centum languages. Here we have one instance where it seems plausible to posit a parallel development occurring independently across several languages, as delabialization before *u is a very natural sound change that has been observed occurring many times in different languages. Indeed, there is independent evidence for this sound change occurring in the different centum languages long after the breakup of PIE.

There are two examples of words with a plain velar occurring before *n: *gnet ‘knead’ and *ghnid ‘nit’. As these only have reflexes in Germanic among the centum languages, we need only posit that labiovelars were delabialized before *n in Germanic to explain the plain velars in these words. In fact this is an example of a broader delabialization before consonants in the western languages – Germanic, Italic, and Celtic – which was first noted by Antoine Meillet back in 1894, and for which there is significant independent evidence. It is difficult determine the distribution of velars before *m or before syllabic resonants due to lack of data. Suffice it to say that there is no solid evidence against the two-velar hypothesis in these positions either.

We thus arrive at the following picture. Proto-Indo-European had two types of velar in word-initial position: plain velars and labiovelars. In the satem languages, labiovelars were delabialized and plain velars were palatalized apart from after *s, and before *r, *s, and *H₂; palatalization was also blocked before *l in Albanian and *lo in Balto-Slavic. In the centum languages, labiovelars were delabialized before *u, and labiovelars were also delabialized before *o and before consonants in the western languages. This raises the question whether the two-velar hypothesis holds in word-medial position too; this is a more difficult question to answer, as there are more environments word-medially that need to be considered. I will therefore postpone attempting to answer this question to a future blog post.