Secrets of the Words You Know

  • Home
  • About
  • Blogs
  • Technical Notes
  • Comments
  • Contact


Secrets of the Words You Know

Blog 28 PIE Gutturals 2 *ĝh and *gh and the secrets of the legal law

Now to the PIE Voiced Aspirates:

(i) Palatal *ĝh and Velar*gh at the start of a word

For these aspirated PIE palatals and velars, we can expect the Grimm’s Law outcome in Proto Germanic to be the voiced but un-aspirated *g.

As you may recall from Blog 8, the English outcomes will however depend on whether the Proto-Germanic *g was followed at the time of fronting/palatalization by a front vowel or not. The Proto Germanic *g will give ‘y’ or ‘i’ in the first case and be preserved as ‘g’ in other cases.

Here is where we can reconnect with the secrets of the words ‘yellow’ and ‘gold’ that we mentioned as long ago as Blog 3. Both these words, inherited by English from Proto Germanic, descend from the PIE root *ĝhel-, which seems to have had the original meaning of ‘to shine’, thus also providing the source for bright colours (notably yellow) and shiny or yellow/green substances.

Many other English words from the same root have kept the Proto Germanic *g, as in these words it was not followed by a front vowel at the time of palatalization. One example is ‘gall’, in the sense of our greenish/yellow bile stored in the gall-bladder. But, as we saw earlier, in Blog posts 4 and 10, the root is even more widely visible in English in words derived from a zero-grade *gl- with various suffixes. These include ‘glass’ ‘glint’, ‘glisten’ and ‘glow’ but also ‘glad’ ‘glee’ and ‘gloat’, showing more distant derivatives of a core meaning around ‘shining’.

In Greek, the colour sense is green rather than yellow, as in ‘chlorophyll’. And the sense of ‘bile’ is evident in words like ‘melancholy’, literally ‘black bile’ and ‘cholera’. The sound ‘ch’ – originally a voiceless aspirate, but now in Greek a fricative, despite its pronunciation as a ‘k’ in English – is what we would expect in Greek from PIE *gh.

In Celtic the reflexes of the same root mean the ‘green’ – or perhaps better ‘grey-green’– of nature in place names such as the ‘grey/green hollow’ of Glasgow and the ‘grey/green wood’, Glascote. (In modern Welsh the word is used for ‘blue’ rather than ‘grey/green’.) These show that Celtic merges voiced aspirates into the plain voiced stops.

Turning to the ‘satəm’ languages, the fact that the root begins with the palatal *ĝh- rather than the plain velar *gh- is shown by the first letter of the Polish currency the ‘zloty’. (Issuers of currency often want their currency to be seen as ‘worth its weight in gold’ – usually when it isn’t. The former Dutch guilder was another example.) The Polish ‘z’ is as expected. As we have already seen, Slavic is one of the many language groups that merges PIE voiced aspirates with the plain voiced sounds, and as a ‘satəm’ language converts palatals into the relevant fricative, here the voiced ‘z’.

In Sanskrit, ‘hari’, also from this root, shows ’h’ as the Indic evolution of *ĝh, and as we seen earlier it is also another example of an original PIE *l becoming an ‘r’ in Sanskrit. It originally meant ‘yellow’ as in the word for turmeric, ‘haridrā’. However, it was then thought to be related to a very similar word meaning ‘to take away, remove or destroy’, which was then applied to gods who took away sins or ended the entanglements of reincarnation, and thus morphed into a word for the divine, as we have just seen with ‘Harijan’.

Much less obvious is ‘arsenic’, a word borrowed via a string of languages but going back to Middle Persian ‘zarnik’, meaning ‘golden’ in reference to its colour (note the characteristic shifts of PIE *gh to a ‘z’ and also *l to ‘r’ in Iranian). One reason why the modern word seems so unrelated to the Persian original is that the Greeks applied a ‘folk-etymology’ (Technical Note 3, Box 3) to this unfamiliar word: ‘arsen’ in Greek has the sense of ‘virile’.

There are no obvious borrowings from Latin from this root. We can however show that the usual evolution of *ĝh and*gh in Latin is ‘h’. A couple of examples from roots with *ĝh:

  • English borrowings from Latin ‘horrid’ and ‘horrible’ from a root *ĝhers-, meaning apparently ‘bristle’, and conserved in English as ‘gorse’. 
  • A root *ĝhorH-, meaning ‘entrails’ or ‘guts’, gives us a neat trio of inherited ‘yarn’ (with fronting), Latin ‘hernia’ and Greek-derived ‘chord’ (in the sense of a string rather than a musical chord) – and also ‘cord’, the loss of the original Greek ‘ch’ in the spelling of this word reflecting French spelling conventions.

However, Latin is not fully consistent. When *gh is followed by *u the evolution in Latin is to ‘f’. This can be seen from words derived ultimately from a PIE root *gheu-, which has to do with pouring liquid, including as a religious libation. English ‘gut’ is one such word, as are our borrowings from Old Norse ‘gust’ and ‘geyser’, but there are also copious borrowings from Latin, some of these where pouring is central, such as ‘infuse’, but more often where the meaning has drifted far away from this core, such as ‘confuse’ and ‘refuse’.

Let’s pause briefly on another example of Sanskrit ‘h’ for *ĝh. This can be seen in a word ultimately derived from the main PIE word for ‘winter’ *ĝheim-: the snowy mountain chain of the Himalayas (in Sanskrit the word has changed meaning from ‘winter’ to ‘snow’). It’s not a root visible in mainstream English inherited words, though there is a dialect term ‘gimmer’, for a lamb that has seen its first winter, that comes from the Old Norse equivalent (as the ‘g’ rather than ‘y’ shows).

So far all the roots have begun with the palatal *ĝh; let’s now add a few starting with the velar *gh. We can expect exactly the same reflections of this sound in English, Latin and Greek, but a harder sound in the ‘satəm’ languages.

Here are a couple of examples.

Let’s think first about gardens.

What do we practice in them? – horticulture! Could we be looking at another pairing of a ‘basic’, inherited English word and a more ‘scientific’ Latin borrowing? Well, nearly. It turns out that the more direct English descendant of the root lying behind ‘horticulture’ (the first element of which is indeed based on the Latin for ‘garden’) is actually ‘yard’, in the sense of an enclosed area, showing fronting of the Proto Germanic ‘g’. English has taken ‘garden’ from a late Latin word meaning ‘enclosed’, which was itself borrowed from a Germanic source from the same root, which conserved the Proto Germanic ‘g’

The root underlying all this is a PIE root *gher- with a suffix -dh-, which meant roughly the same as its English descendant to ‘gird’ or to ‘enclose’, and a related noun *ghordho- (this is a typical alternation of ‘e’ and ‘o’-grades between verb stems and noun derivatives). The latter appears in English not just as ‘yard’ but also its doublet ‘garth’ deriving from Old Norse. ‘Asgard’ is the Norse equivalent of the mythical home of the Greek Gods, Mount Olympus, and means literally ‘The Garden of the Gods’. (We will come back to the first syllable in a minute).

The same root would seem to link also to Slavic ‘grad’ for ‘city’ as in ‘Volgagrad’ (note the metathesis of the ‘r’ and the ‘a’) or the more typically Russian version we saw earlier in ‘Novgorod’ under *n. These words show the expected retention of the hard *g from *gh before the back vowels ‘a’ and ‘o’.

It is less certain whether Greek-derived words like ‘chorus’, ‘choir’ (via Latin) and possibly also ‘carol’ (via French) belong here, the chorus in Greek tragedy singing and dancing in the enclosed space of the ‘orchestra’ of an outdoor theatre, or whether they are from a different root.

A second example is a root *ghos-ti-, which provides a set of terms about ‘strangers’[1], and which we mentioned in Technical Note 2 and Blog post 6 as an example of very different senses developing from an original meaning.

Again, we see alongside ‘guest’ (from Old Norse, hence the hard ‘g’) ’host’, ‘hospitality’ and ‘hostility’ showing Latin ‘h’. ‘Xenophobia’, a borrowing from Greek, goes back to a zero-grade root which brings the *gh and the *s together.

Two very common English words, both as noted earlier borrowed from Old Norse, start with ’g’ from velar *gh: ‘give’ and ‘get’.

  • ‘Give’ derives from a root *ghabh-, perhaps with an original sense around ‘hold’ but yielding words meaning ‘take’ rather than ‘give’ in many other Indo-European groups. In Latin, it is the main word for ‘have’, as in the legal phrase ‘habeas corpus’ (mentioned in Technical Note 2 when we explained that the English word ‘have’ could not be related to this Latin verb), and it lies behind many borrowed words from ‘habit’ and ‘exhibit’ to far less obvious ones, including ‘able’, originally with a sense of ‘easily handled’. (The ‘h’ is lost in this form.) Even less evident, and hidden behind a prefix ‘de’, is a set of words about ‘owing’ from ‘debt’ to ‘due’, where the ‘b’ and ‘u’ are the last reflections of the labial in the root.
  • *ghed-, the source of ‘get’ (and indeed ‘forget’, where ‘for’ is used as a form of negation) seems to have a basic meaning around ‘seize’, and is apparent in more borrowings from Latin like ‘apprehend’, from a form of the root with an infixed ‘n’, so *ghend-. As the English use of this word shows, one can ‘seize’ something either literally or figuratively by grasping a meaning and so learning. An ‘apprentice’ (taken from the French descendant of the same Latin word) is therefore ‘one who learns’. The verb is unusual in Latin in that is found only with the prefix ‘pre-’, and often a second prefix as in ‘comprehend’ or ‘reprehend’.

As a final example, there is a PIE root *ghaiso-, which appears to have had the original meaning of a stick, but particularly a spear or pointed object. Its English descendants also show the change from *s to r between vowels, though the following vowel has often subsequently been lost. These are evident in the word ‘gore’ (as in ‘to spear’ someone) and in gar-lic, literally the ‘pointed-leek’. Proper names are often a relic of an otherwise obscure root, and this one shows up in the ‘spear-wielding’ Gerald, the ‘spear-hard’ Gerard and ‘God’s spear’ Oscar, the ‘Os-‘ element being, like the ‘As-‘ of Asgard (above) from a root about God or spirit, evident also in Oswald (‘God’s power’) and far away in Persia, the supreme deity of Zoroastrianism, the ‘Supreme Spirit of Wisdom’, Ahura-Mazda.

The sound code for English initial ‘g’ can be set out as:

English initial g in an inherited word = PIE*ĝh or *gh = Latin h (but f before *u) = Celtic g = g in other Germanic languages. PIE *ĝh = Balto-Slavic and Iranian z = Indic h. PIE *gh = g in Balto-Slavic and Indo-Iranian before original back vowels, palatalised before original front vowels.

English initial y has same sound-code if from PIE*ĝh or *gh; but see Blog post 17 (y) if from PIE *y

(ii) Secrets of the Legal Law

In the musical ‘Camelot’ the climate is determined by ‘the legal laws’. That may seem fanciful, but here’s the question for us: are ‘law’ and ‘legal’, clearly close in meaning, from the same root or not?

‘Law’ is generally agreed to be among our many borrowings from Old Norse: ‘lagu’, with a Proto-Germanic origin of *lagam, meaning roughly ‘put or lay’. The Indo-European root is *legh-, whose core meaning is along the lines of the inherited verbs ‘lay’ and ‘lie’, the ‘y’-sound in English reflecting an earlier ‘g’, as we saw in words like ‘day’, ‘say’ and ‘way’ in Blog post 8. So ‘law’ is something ‘laid down’.  Related words in English include ‘lair’ (somewhere where you ‘lie low’, ‘low’ being another word from the same family) – as well as the borrowing ‘lager’ from the German for the ‘storehouse’ where such beers were kept, and ‘laager’ from Dutch and Afrikaans for ‘defensive encampment’, both of which conserve the Proto-Germanic *g. From the French, and ultimately from the Latin, for ‘bed’, which goes back to the same PIE root with the common verbal suffix *t, we also get ‘litter’ in at least three senses: a vehicle to carry someone in a couch, the offspring of an animal, and mess (going back to when beds were just of straw).

Most analyses of ‘legal’, a borrowing from Latin, assume that it goes back to a rather similar root *leg-, but with an unaspirated final consonant. (A *gh between vowels should give a Latin ‘h’ rather than a ‘g’, as you can see from our borrowing ‘vehicle’, which is a doublet of our Dutch borrowing ‘wagon’, showing a Proto Germanic ‘g’ which necessarily goes back to a PIE *gh.) This root has a different meaning, along the lines of ‘gather’ or ‘collect’ (the latter itself a term derived from the root, and borrowed from Latin). It has few direct descendants in English, though these include ‘leech’ in its now obsolete usage as a medical doctor, going back perhaps to a sense that a doctor is someone who can tell you something important. But it has also given rise to words for ‘read’ (Latin) and ‘speak’ (Greek) in the sense of ‘putting words together’, which in turn provide a vast number of English words borrowed from both languages. For example, Latin provides us with many words ending in ‘-lect’ such as ‘dialect’, ‘elect’, ‘select’, and ‘neglect’, and reading-related like ‘lectern ’or ‘legible’; and Greek an even larger number of words ending in ‘-logue’ and ‘-logy’, ‘logic’ and its derivatives, ‘dyslexia’, and so on.

So, in ‘law’ and ‘legal’ we have two words with similar meanings in modern English, which are both borrowed, one from Old Norse and the other from Latin, and whose ultimate PIE roots are in fact different – phonetically similar but not particularly close in terms of meaning.

(iii) What’s an inherited English word from another PIE root beginning with *gh?

The root *ĝhor- has the sense of ‘strong desire’, visible in the borrowing from Latin ‘exhort’. Greek-derived terms are in words which have the rather different sense of ‘grace’ like ‘charisma’ and ‘Eucharist’ (literally ‘Thanksgiving’, and indeed the basis of the Modern Greek for ‘Thank you’). What’s an inherited English verb from this root with the sense of ‘to desire’ (perhaps particularly something lost)? Start by considering what the consonants of the root would have yielded in Proto Germanic. Hint: the English verb has a suffix, so three consonants in all.


[1] But not ghosts! ‘Ghost’ goes back to a separate root *ĝheis- with a sense of ‘terrify’ or ‘scare’, with cognate words in Indo-Iranian languages, not however borrowed into English