Secrets of the Words You Know
Blog post 24: PIE Stop Consonants *d
Now we turn to the dentals. They are, like the labials, a generally well-behaved group, in the sense that most PIE dental sounds remain as dentals of one kind or another in most languages. However there are some important exceptions, involving a switch to labials and also from ‘d’ to ‘l’, which we will highlight.
We start with PIE voiced *d
(i) *d at the start of a word
If you are in doubt whether a particular language is Indo-European or not, a good initial test is to discover the words for ‘two’ and ‘ten’. If both begin with ‘d’ you can be pretty sure that it is. This is because both numerals began with *d in PIE and because initial *d has proved a stable sound. Borrowings from other Indo-European languages into English which demonstrate this are numerous, from ‘duopoly’ (from Greek) to ‘decimate’ (from Latin). The latter term, now used in a general sense of ‘to cut down massively’, comes from an originally more precise punishment used by the Roman army, whereby one soldier in 10 (chosen by lot) would be killed following instances of indiscipline, cowardice, mutiny etc.
The Germanic languages, thanks to the Grimm’s Law changes, are the main exceptions, all but one showing the expected change to voiceless ‘t’ as in ‘two’ and ‘ten’. As we saw in Blog post 9 the outlier is German itself, where Proto Germanic *t has evolved in initial position to ‘tsv’, spelt ‘zw’, as a result of the Old High German sound shift.
So why do so many borrowed words about duality, from ‘binoculars’ to ‘bisexual’, start with ‘bi’, with a shift from a dental to a labial? It is because of a sound change in Latin. The Latin for ‘two’ is ‘duo’, with the ‘u’ sounded as a vowel, and the expected ‘d’ unchanged. But in related words where the inherited ‘u’ is sounded as its matching consonant (‘w’) the resulting initial cluster *dw evolves to ‘b’, no doubt because the lost labial ‘w’ has ‘coloured’ the originally dental ‘d’.
Latin shows another rather similar change where *d is immediately followed by a ‘y’. A good example comes from the PIE root *dye(u), which is the origin of words about the Gods such as ‘deity’ or ‘divine’ (from Latin). It is also the source of the familiar Indian term ‘deva’ for a god.
In some words from this root, Latin – where it also gives the normal word for ‘day’ – conserves a sequence ‘di’ before a vowel. We can see this in words we have borrowed such as ‘diurnal’ and ‘meridian’. However, when the syllabic ‘i’ in the Latin words mentioned above shifts to its matching consonant ‘y’ (typically spelt ‘j’ in Latin), the ‘d’ is lost, giving for example ‘Jupiter’ (literally ‘day father’).
This would in Classical times have been pronounced with an initial sound like English ‘y’, which then, as we saw under *y, would have evolved to a standard French voiced fricative still spelt with a ‘j’, but sounding like the ‘s’ of ‘leisure’ and often notated as ‘zh’. However, most words beginning with a French ‘j’ were taken into English with a Norman pronunciation as the affricate ‘dge’. This sound was not present in Old English in initial position, but was relatively common in non-initial position as in words like ‘bridge’. So the initial sound of Jupiter has evolved from ‘dy’ to Classical Latin ‘y’, French ‘zh’, but English ‘dge’, re-introducing the ‘d’ sound that would have existed in pre-Latin – by Jove!
In Greek, the sequence ‘dy’ also gave rise to an unusual development, seen in the name of Jupiter’s equivalent as ‘top God’, Zeus. The pronunciation of the first letter (Greek ‘zeta’) in Classical times seems to have been ‘zd’ or ‘dz’ before evolving to a sound like an English ‘z’.
What about English inherited terms from this root?
If we believed meaning to be more stable than sounds, one obvious candidate would be the word ‘day’ itself, but you know from Technical Note 2 that such a presumption would be wrong, and from Blog post 20 that such a word should (indeed must) start with a ‘t’ in accordance with Grimm’s Law. So, can you think of say a Germanic god whose name begins with a ‘T’ and all of us use on a regular basis? The answer will be in the next email to Subscribers.
While we are focused on Latin idiosyncrasies from *d, there are two others in common words. The English words ‘tongue’ and ‘tear’ (as in ‘shedding a tear’) both go back, as you would now expect, to PIE words beginning with *d, which would normally be conserved in Latin. But in both cases we see instead in Latin an unexpected ‘l’.
That’s why when we study ‘tongues’, we speak of ‘linguistics’ and (via French) ‘language’ rather than the expected ‘*dinguistics’ and ‘*danguage’. Although ‘d’ and ‘l’ are articulated quite similarly, as you can see by trying them in succession, the most likely reason seems to be a form of semantic analogy (Technical Note 3, Box 2), caused by the words for ‘lick’ that we considered under *l. Something similar has happened with this root in Armenian and Lithuanian.
It’s also why we may be accused, when crying, of being ‘lachrymose’ rather than ‘:dacrymose’, as ‘tear’ and forms in languages such as Greek would suggest. In this case, the use of ‘l’ may possibly be from another Italic dialect. A similar change is seen in the contrast between Latin ‘Ulysses’ for Homeric and Classical Greek ‘Odysseus’, though here the variant with ‘l’ is already found in some dialects of Greek.
Somewhat similarly, Germanic and Latin share the simplification of an initial *dl to a simple ‘l’. A notable example is the PIE root *del- meaning ‘long’, and visible in the Greek ‘dolichocephalic’ for ‘long-headed’, from the o-grade of the root. In Germanic and Latin, words formed from the zero grade *dl- lose initial *d, giving very similar words for ‘long’ in both sets of languages.
Moving East, the root *derk-, meaning ‘to see’, not inherited by English, is at the origin of some perhaps unexpected words in Greek and the Indian subcontinent. The Greeks, who used a derivative of this root as a prime word for ‘to see’, are thought to have named the mythical ‘dragon’ either from a presumed ability to freeze its prey with a glance or because some snakes have transparent eyelids and so appear to have their eyes permanently open. The same root forms the second element of the Indian National Television service ‘Doordarshan’ (literally ‘far-sight’, just like the Greek/Latin invented blend ‘Tele-vision’). And if you ‘take a dekko’ at something, you are using another Hindi imperative in ‘-o’ (as in ‘shampoo’), this time from the Hindi ‘dekho’, ‘Look!’ from the same root, with loss of ‘r’.
We can learn something about how the original speech community oriented themselves by observing that the PIE *dek̂s-, the source of our borrowings from Latin like ‘dexterous’ (from the Latin for ‘right’ as opposed to ‘left’), is also the origin of the Indian term ‘Deccan’ for the peninsula formed by Southern India. If the ‘South’ is on your right, you are of course facing the East, where the sun rises, and this seems to have been the basis for the orientation system of the first Indo-Europeans. It’s a nice example of historical linguistics giving us information on how people thought and behaved 6,000 years ago which we would not know from any other source.
Finally, let’s take the PIE root for ‘tree’: *deru- or, with metathesis, *dreu-. The initial *d is well-preserved in most Indo-European languages. English has acquired words from at least three of them:
- From Greek we have ‘dryad’ (a wood-nymph), and words with the element ‘dendro’ such as ‘dendrochronology’ or the ‘rose-tree’ ‘rhododendron’, which are formed from a reduplicated form of the root. (The ‘n’ in the reduplicated root seems to be another example of dissimilation (Technical Note 3, Box 7), to avoid repeating the ‘r’ in the reduplicated version, so *derdro-> dendro-.)
- The Deodar tree, a cedar native to the Himalayas, means ‘divine tree’ in Hindi, the first element going back to Sanskrit ‘deva’ which we noted above;
- Trees have qualities of durability and reliability (indeed, that may well have been the initial sense of the root). Under *w we noted that the two elements of ‘druid’, of Celtic origin, were first tree-like steadfastness and strength, and second knowledge or wisdom. A druid is therefore a ‘true seer’. (Modern Irish and Gaelic still use the word draoi (genitive druadh) to mean a magician.)
Inherited words of course will start with ‘t’. Apart from the obvious ‘tree’ and also ‘tar’ (originally for a gum derived from a tree) the root is also reflected in wooden objects, for example ‘tray’ and, with a *k-suffix, ‘trough’ and the Old Norse borrowing ‘trug’.
English, like the Celtic languages, has a set of ‘tree-related’ words that refer to tree-like qualities, such as ‘true’, ‘trust’, and ‘truth’ and related ‘troth’. We also have, with an *m-suffix, ‘trim’ – an adjective, noun and verb, all ultimately derived from a sense of solidity and reliability, but having diverged quite considerably among themselves in meaning over time.
So here is the sound-code for English initial ‘t’:
English initial ‘t’ in an inherited word = PIE *d = d in Latin (but *dl>l and very occasionally *d>l as in ‘language’ etc ) = d in Greek, Indo-Iranian, Celtic, Balto-Slavonic = t in other Germanic languages except German ‘zw’ pronounced ‘tsv’. *dl>l in Germanic, as in Latin.
No other PIE sound yields an initial English ‘t’.
(ii) *d elsewhere
*d is also quite stable in intermediate or final position, as the following examples show:
- *Hed- (‘to eat’), giving English inherited ‘eat’ and borrowings from Latin like ‘edible’. The present participle of this verb is the source of:
- ‘dental’ and similar ‘teeth’ words from Latin, the tooth being ‘the biter’;
- words borrowed or (often) made up from Greek with the element ‘-odon’ (the first ‘o’ once again signalling that the original root did indeed start with a laryngeal). A ‘mast-odon’ is literally a ‘breast-tooth’, because of nipple-like projections on its molars;
- the inherited words ‘tooth’ (we noted in Blog post 9 that the loss of the ‘n’ seen in both ‘dental’ and ‘-odon’ is a typical development in English before the fricative ‘th’) and, with an ‘-sk-’ suffix, ‘tusk’.
- *ned- (‘to tie’), giving English inherited ‘net’, as we saw under *n, and Latin borrowings like ‘node’ and ‘nodule’.
However, in Latin *d is lost in the combination *dw between vowels. This is why our Latin borrowing ‘suave’ shows no trace of the original dental which is found in the PIE stem *sweHdu-, mentioned under *s, and which is conserved in inherited ‘sweet’ and Greek-derived ‘hedonist’ (with the usual change in Greek of initial *s to ‘h’).
(iii) The secrets of Time and Tide
As an example of how various the meanings of a PIE root can become in the daughter languages, it’s worth pausing on the root *deH(i)-. Its most basic meaning seems to be ‘divide’, but that leads to its descendants gaining the meaning of whatever it is that is being divided.
So with a suffixed *m it is the origin of our borrowings from Greek like ‘democracy’ (’people power’) or ‘demagogue’ (‘people leader’), to say nothing of ‘pandemic’ (‘all the people’).
You can readily divide time as well as people, so it is also the origin of our inherited word ‘time’. With a dental suffix it also gives us ‘tide’, whose usage in English has largely narrowed to the daily ‘seasons’ of the sea, but is the origin of the standard German word for ‘time’, ‘Zeit’, with the expected initial written ‘z’ for English ‘t’.
Lastly, from the root with an *-i suffix, we have ‘demon’, from the Greek word ‘daemon’ for ‘divine spirit’ (originally ‘one who provides or divides’) that has developed a more pejorative sense on its journey to English via Latin.
(iv) What’s an inherited English word from another PIE root beginning with *d?
Our starting point is a PIE root *deik-, with a variant *deiĝ-, meaning ‘to show’ but also in the sense of ‘to make a solemn pronouncement’. The version with a final *k is at the origin of the general Latin term for ‘to say’, which has given us all manner of borrowings including the more solemn pronouncements like ‘verdict’, ‘edict’ or ‘abdicate’ as well as ‘dictate’, and, via French, ‘condition’ and ‘ditty’ and from Italian ‘ditto’ (’the said’ term in a list, to avoid repetition).
You task is to find an English inherited word from the version ending in *ĝ. There are at least two possible answers: one is a verb about an important form of communication of which we have all been the beneficiaries, and the other a noun which ‘shows’ something. Bear in mind that the final *ĝ (which will give Proto Germanic *k) may be palatalized or it may not, as you do not know the phonetic environment at the time of fronting (see Blog post 7 if you need to refresh your memory of the fronting of Proto Germanic *k in English). However, I can tell you that one of the two words shows fronting and the other does not.
As usual, Subscribers will be sent the answer in the next posting.