Secrets of the Words You Know
Blog 20: Introduction to the PIE Initial Stop Consonants (2)
It is now time to introduce the German scholar Jacob Grimm. He was the elder of the ‘Brothers Grimm’, who are best known for their pioneering work in collecting German folk and ‘fairy’ tales and who also embarked together on the largest German Dictionary of the Nineteenth Century, which was left unfinished at their deaths.
Grimm wrote a major history of the German language. Like others before him, he observed the frequent changes visible between Germanic languages and other ‘Indo-Germanic’ languages, such as Sanskrit, Latin and Greek, of the kind that we looked at in the very first blog, where we noted the equivalence of English initial ‘f’ with Latin or Greek ‘p’.
Grimm’s great achievement, presented in the second edition of his ‘Deutsche Grammatik’ in 1822, was to see these changes as an organised system, whereby:
- PIE voiced aspirates became Germanic voiced stop consonants (*bh>b; *dh>d; *ĝh and *gh>g).
- PIE voiced stop consonants became Germanic voiceless stop consonants ([*b>p [very few examples]; *d>t; *ĝ and *g>k)
- PIE voiceless stop consonants became Germanic fricatives: (*p>f; *t>th; *k̂ and *k>h)
This set of changes is known to this day as ‘Grimm’s Law’. Only a very small number of historical linguists have a ‘law’ to their name, and Grimm was the first to achieve this distinction. As often, other linguists had observed at least part of the picture that Grimm set out, notably the Danish scholar Rasmus Rask, but Grimm was its clearest and most systematic proponent.
Grimm also pointed out the somewhat similar (but of course much later) evolution of stop consonants (the dental stop consonants in particular) in the Old High German Sound Shift, which we discussed in Blog post 9.
English has in general conserved the Grimm’s Law changes in word-initial position except for the palatization of velars before front vowels, discussed in Blog posts 7-8.
It is true that these patterns break down in some contexts. For example, the evolution of the PIE labio-velars has its own particularities. Different outcomes take place in some phonetic environments: as we saw in Technical Note 2, *p is conserved in the sequence *sp, and the same is true in Proto Germanic for other voiceless stops after *s. We shall see two other important examples later where the ‘normal’ pattern is overridden by particular phonetic contexts, though, as with the *p of *sp, these are not in initial position.
The wider picture for stops in initial position (and for many language groups also elsewhere) is as follows:
- Apart from Armenian, which has a set of changes with some similarity to Germanic, most Indo-European languages have largely conserved the PIE voiced and voiceless labial and dental stops. That’s what we have been seeing throughout in the regular preservation of initial *p in other languages as opposed to English initial ‘f’. (As we shall see, initial *p has a unique outcome in Celtic.)
- For the PIE voiced and voiceless palatal and velar stops[1]:
- In the ‘centum’ language groups (including Greek, Latin, Germanic and Celtic) the palatal and velar voiced and voiceless stops merge together. Fronting, typically only before front vowels as with *k>ch in English (‘chill’) or *k> ‘soft c’ in French (‘cent’), has taken place only in the last 1500 years. The evolution of labio-velar stops is quite various.
- In the ‘satǝm’ language groups (including the Indic, Iranian, Baltic and Slavonic languages), the palatals were fronted far earlier before all vowels (typically*k̂>’sh’ or ‘s’) , and the velars and labio-velars, which merge together, are fronted only before front vowels, and then only as far as affricates (eg *k or *kʷ> ‘ch’).
- When we turn to the voiced aspirated consonants, the Germanic group is less of an outlier, since Celtic, Balto-Slavic and Old Persian also lose the aspirated element of the inherited sound[2]. (The shift of *bh>b in these language groups is one way in which the apparent ‘gap’ of the near-absent PIE *b has been ‘filled’ over time.)
However in these other languages, unlike Germanic, the inherited unaspirated voiced consonant remains unchanged. So as an example, for such languages, in the dental series PIE *d and *dh fall together as ‘d’, while in Proto-Germanic they would give ‘t’ and ‘d’ respectively.
As we saw in the previous blog post, Latin and Greek – the languages who have supplied most of our borrowed words – also distinguish an inherited aspirated stop consonant from the comparable unaspirated voiced consonant, but in a different way from Germanic. Both keep the voiced stops unchanged. For the voiced aspirates, Latin uses either a fricative (‘f’) or the voiceless aspirate ‘h’; while Classical Greek used voiceless aspirates, which subsequently developed into the fricatives still evident in Modern Greek and typically written in English as ‘ph’, ‘th’ and ‘ch’.
Sanskrit and important modern Indian languages are the most conservative, since they have in general maintained the voiced aspirate itself.
In the next 12 postings we will discuss the PIE stop consonants in the following groups:
- Labials
- Dentals
- Gutturals (ie both palatals and velars together)
- Labio-velars
In each case, for reasons to be explained later, we will discuss these in the order Voiced; Voiced Aspirates; Voiceless.
[1] You may like to look back at Section 3 of Blog 11 on the Consonants of Proto-Indo-European to refresh your memory on these sounds.
[2] The same is true of Armenian, but in Armenian the voiced and voiceless stops also shift in a way that partially matches what happened in Proto-Germanic.