Blog 27 *ĝ and *g, and the Secrets of Knowledge
In Blog post 11, I set out the general view that Proto Indo European had three series of ‘guttural’ consonants:
- Palatals, probably articulated against the hard palate, designated by the addition of a ‘hat’ as in *ĝ
- Velars, probably articulated further back, against the soft palate, designated with a simple letter, as in *g
- Labio-Velars, which combine a velar sound with a simultaneous articulation of the lips, designated with a superscript ‘w’, as in *gʷ
I also explained that the so-called ‘centum’ language groups, such as Germanic, Celtic, Italic and Greek, merged the palatal and velar series, whereas the ‘satəm’ language groups, such as Indo-Iranian and Balto-Slavonic, merged the velar and the labio-velar series.
For ease of reference, in this set of Blog posts, I shall follow the tradition set by the ‘centum’ language groups, including of course Germanic, from which English is derived, and deal with each of the palatals and velars (voiced, voiced aspirate and voiceless) together. Within each of these Blog posts (27-29), I will discuss the palatals first and the velars second.
I will then treat the labio-velars separately in three further Blog posts (30-32).
We start with Voiced *ĝ (palatal)and *g (velar)
(i) *ĝ or *g at the start of a word
In initial position we can expect, under Grimm’s Law, a PIE *ĝ or *g to give a ‘k’ in English, except that before a front vowel fronting (‘palatization’ in technical parlance) will have taken place. In that case the sound will be ‘ch’, as we illustrated with ‘chill’ and other examples in Blog post 7.
As examples from PIE roots starting with the palatal *ĝ, we may quote the following:
- From the root *ĝr̥Hn-, meaning ‘grain’, we have the inherited English word ‘corn’. The word ‘kernel’ is also from the same root, and the inherited ‘k’ is conserved because at the time of palatization the following vowel in this case was a ‘u’, which would not cause the ‘k’ to change to ‘ch’. English has multiple borrowings from Latin, usually via French, most notably the word ‘grain’, but also ‘garner’, ‘granule’, ‘granary’, ‘granite’ (the ‘grained’ rock) and ‘grenade’. All these conserve the initial hard consonant of the PIE root.
- The PIE root *ĝonu- is the source of words for ‘knee’ (inherited) in ten different Indo-European language groups. As we have seen, English words beginning with ‘kn’ are good examples of how English spelling conserves the history of sounds now lost in pronunciation. English has borrowed words derived from the same root most obviously in Latin borrowing ‘genuflect’, but also in Greek borrowings like ‘Pentagon’ or Trigonometry’, Greek having developed from its version of the same root a word meaning ‘corner’ or ‘angle’.
- From a root *ĝeus-, whose core meaning seems to be ‘to taste’, we have inherited ‘choose’ (at the time of palatization the *e of the root still immediately followed the ‘k’ resulting from Grimm’s Law), which can be compared to ‘goȗt’ as in the well-known French phrase ‘chacun à son goȗt’ (‘everyone to his own taste’) and Italian ‘gusto’. The original *ĝ is also conserved in Celtic, as for example in the names ‘Fergus’, which we mentioned under *w, and ‘Angus’, where the second element has evolved to the sense of ability or strength, so ‘man-strength’ and ‘solo-strength’ respectively (the first syllable of ‘An-gus’ goes back to the PIE root from which English has derived the word ‘one’).
Now to the velar *g.
Let’s take PIE *gel-, which has a core meaning of ‘cold’. We came across the Proto-Germanic form of this root, *kal-, in Blog post 7, from which ‘cold’ and fronted ‘chill’ are derived. The root is easily visible in the little-used Latin borrowing ‘gelid’ but also in other more routine borrowings from Latin like ‘glacier’ (a good example of zero grade) or from French like ‘congeal’, ‘gelatine’ and ‘jelly’.
We have not mentioned any examples above from satəm languages, though we will give some Indic examples below. Borrowed words from these languages are rather scarce from these sounds.
As you will see in the sound-codes for all the gutturals, including the labio-velars, the evolution of these sounds is more complex than the sounds we have discussed to this point. Here, to make the point, is the sound-code for initial ‘k’, including written but silent ‘k’ as in ‘knee’, for hard ‘c’, and for ‘ch’ in an English inherited word:
k/hard c/ch = PIE*ĝ or *g = Greek, Celtic g = Latin g (but g lost in initial cluster gn – see below). For the main satəm languages (see Blog 11), PIE *ĝ yields Slavic z and Indic j (pronounced ‘zh’ – see below), while PIE *g gives g before original back vowels and various palatalized forms before original front vowels.
In addition, however, as we shall see shortly, PIE*gʷ can also yield English initial hard c in a few cases in an inherited word: see the sound-code for ‘qu’ in Blog post 30 (to come).
(ii) *ĝ or *g elsewhere
Both *ĝ and *g are quite well maintained in medial and final position, as we can see from various words descending from a root *Herĝ– (‘white’) such as Latin-derived ‘argent’ (‘silver’) and so also ‘Argentina’ (the ‘land of silver’), the Greek city name ‘Argos’ and – with the expected evolution of palatal *ĝ > Sanskrit ‘zh’ – the hero of the Sanskrit Mahabharata epic, ‘Arjuna’ (the white or shining one). We have already seen that Sanskrit ‘yoga’ maintains the original velar *g before a back vowel.
An important word with a *ĝ in its root is the PIE for ‘I’, *Heĝo(m), easily visible in the familiar Latin word ‘ego’. In English, we might have expected a version like German ‘ich’, in which the PGmc *k from PIE *ĝ would appear in palatalized form. You will occasionally find this in an old text, as in the madrigal line ‘Heigh-ho, chill love no more’, where ‘chill’ stands for ‘ich will’. But the consonant was lost, dialect by dialect, with a typical lengthening of the vowel (which then itself developed from a long [ī] to the present [ai] as part of the Great Vowel Shift, discussed in a later Blog post).
In a few words PIE *ĝ gives Sanskrit ‘h’ rather than ‘zh’, as we saw under *m with ‘Maha(rajah) as opposed to Greek ‘mega’ and other words that seem to reflect an original PIE *ĝ. This appears to be due to a laryngeal immediately following the *ĝ in the original root (so *meĝH-).
In Latin, inherited *ĝ/*g is lost in some phonetic contexts:
- before ‘y’, spelt in Latin ‘j’ or ‘i’. As an example, the Latin suffix that marks a comparative adjective begins with written ‘i’, so in a root ending ‘g’, the ‘g’ will then disappear. We saw this in the Latin word that provides our borrowing ‘major’ from the PIE root *meĝ-, which we discussed under *m in Blog post 12;
- before ‘m’. That’s why, as we saw under *bh, the Latin doublet to our borrowing from Greek ‘phlegm’ is ‘flame’, without the ‘g’ preceding the ‘m’.
We noted under *bh that Grimm’s Law does not apply to labials or velars before an *s or *t. In the case of velars, all velar stop consonants (including labio-velars) appear in Proto-Germanic as *h in such contexts. The *h is usually visible in English as written ‘gh’ but is no longer pronounced. You can see this in alternations in English like ‘think’/’thought’, ‘bring’/’brought’ (the ‘n’ in both words is lost before the fricative ‘gh’, as we saw with ‘goose’, ‘tooth’ and other examples in Blog post 9). Similarly the PIE stem *Hreĝto- (itself from the root behind the words for ‘ruler’ that we looked at under *r) yields English ‘right’.
(iii) Secrets of knowledge
The English inherited word to ‘know’ is our way in to a broad word-family based on that core meaning, and all descending from a PIE root *ĝneH-, with an o-coloured laryngeal – still in fact visible in the ‘o’ of ‘to know’.
Greek, Latin and French are our main sources of borrowed words from this root, though related words in satəm languages confirm that the initial sound was indeed the palatal *ĝ rather than velar *g.
Greek conserves the initial *ĝ as ‘g’, usually followed immediately by ‘n’, as for example in initial position in ‘Gnostic’ (early Christian sects claiming knowledge beyond the Gospel or the Church) or ‘gnomic’, or in intermediate position ‘agnostic’, ‘diagnosis’ or ‘physiognomy’.
Latin also usually conserves the *ĝ as ‘g’, but not when it is immediately before ‘n’ at the beginning of a word. So from *ĝneH– we have words like ‘ignorant’, ‘cognizance’ or ‘recognition’. But ‘ignoble’ is matched by ‘noble’ (a nobleman being someone who is ‘known’) without an initial ‘g’.
Other common English words borrowed from Latin derivatives of this root with loss of initial ‘g’ include ‘notice’, ‘notorious’ and ‘narrate’, all of which are about gaining, experiencing or sharing knowledge.
Borrowings from French, with the characteristic assimilation of ‘gn’ to ‘nn’, include ‘reconnaissance’ (contrast ‘recognize’, which retains the Latin ‘gn’) and ‘connoisseur’.
But perhaps the less obvious descendants of this root are a number of inherited words other than ‘know’. These include some in limited use:
- the rare and fossilized word ‘kith’, meaning ‘family’ or much more distant ‘relations’: in other words ‘the people you know’
- the noun ‘ken’ (as in the phrase ‘beyond our ken’);
- ‘uncouth’ (’not well known’, and so ‘strange’, ‘crude’). The rarely heard ‘couth’, means ‘famous’ or ‘excellent’ as in ‘Cuthbert’, literally ‘famous-bright’.
At the other extreme they include the ubiquitous ‘can’. Scottish and Northern English dialect ‘ken’ as a verb preserves the full sense of ‘to know’ which has evolved in ‘can’ to the sense of ‘be able’. ‘Canny’ and ‘cunning’ are other examples of how the idea of knowledge may be either something to be admired or perhaps in the second case feared.
(iv) What’s an inherited English word from another PIE root beginning with *ĝ or *g?
We will select a rather similar-looking root *ĝenH- , though in this case the original laryngeal was ‘e-coloured’. It means ‘to give birth’, or ‘to beget’.
It’s particularly well conserved in borrowings from Greek like ‘genetics’, ‘Genesis’ and ‘homogenous’ or from Latin like ‘genus’, ‘generic’, ‘genius’, ‘gender’ etc.
In zero grade, the *ĝ would have been immediately followed by the *n. As we have just seen, Latin, while keeping ‘g’ in intermediate position (as in ‘pregnant’: literally ‘pre-birthing’), loses it before ‘n’ in initial position, just as with the previous root. This is shown in ‘birth’-related borrowings beginning in ‘n’ such as ‘nativity’, ‘native’, ‘national’ and so on.
English has as usual benefitted also from borrowings from Latin via French, such as ‘gentle’ (originally with a sense of ‘belonging to the same group’, then ‘well-born’, then ‘well-mannered’), and the later borrowing of the same word as ‘genteel’ and also – more surprisingly – ‘jaunty’.
In Sanskrit, we see the palatalization we would expect in a ‘satəm’ language, which in the case of *ĝ produces the fricative ‘zh’. The root *ĝenH- therefore underlies the word Janata, familiar from, for example, the Bharatiya Janata Party, which means ‘Indian People’s Party’ or from the term ‘Harijan’, literally ‘God’s People’, a term popularised by Mahatma Gandhi with reference to the Dalit communities of India, but now seldom used.
In line with Grimm’s Law, we can expect Germanic languages to have related words starting with a ‘k’ sound, with the next consonant being ‘n’. So where are the English equivalents? They turn out to be rather few in comparison to the vast number of words derived from this root which English has borrowed from Greek, Latin and French. See if you can find any (answer in next email to Subscribers). The answer mentions three nouns, two of them sounding a bit old-fashioned, and one much more common adjective.