Secrets of the Words You Know
Blog 10: Key Features of Proto Indo European
The previous Blog asked you to write a ‘sound-rule’ that would describe words where English has lost a medial nasal sound which is maintained in German.
Answer: English loses an inherited nasal before a fricative.
There are no written records of Proto-Indo-European itself. Scholars have been able to infer a good deal about the structure of its sounds and grammar from daughter languages, particularly those with ancient texts available to study. Certainty is not always feasible, and several elements remain the subject of debate among specialists.
An initial consensus on the sound system – which is our focus – began to emerge during the nineteenth century, initially giving particular weight to the evidence from Sanskrit, and then balancing this more carefully against the evidence from other known Indo-European languages. The later addition of evidence from language groups now extinct (notably Anatolian languages such as Hittite) has also affected the generally accepted view of the sound system.
Here, we will start with two features of the reconstructed language that are particularly important for understanding the sound-codes that show the relationships between English words that are inherited and borrowed from other descendants of PIE. These are, firstly, the PIE system of roots and suffixes, and secondly its system of what is called ‘vowel-gradation’.
Roots and Suffixes
As we saw from ‘footpedal’ and other words in the first two blogs, the first two consonants of English words seem to be a useful guide to such relationships. Why might this be?
The main reason is that PIE words seem to have been constructed from single-syllable ‘roots’, which carried the core meaning of words formed from them (as in ‘ped-al’ from a root ‘ped-’, which gives the word its core meaning around ‘foot’).
A very high proportion of such roots are of the form consonant-vowel-consonant (CVC for short), so that two consonants define the root, and their descendants are normally visible in all words derived from such roots across the various daughter languages of PIE. However, consonant clusters are also common at the beginning, as we saw with ‘spew’ and ‘sputum’ in Technical Note 2, or at the end of a root. This is one reason why focusing on the first two consonants can only be a ‘broad-brush’ approach: we will come to another reason shortly.
There are various rules in PIE about which consonants can be combined in roots: for example roots could not, it seems, start and finish with a voiced consonant.
Roots are usually fixed, but here are a couple of variants which we shall come across as words are built from them.
- Why is that ‘tactile’ dance, the ‘Tango’, spelt with an ‘n’ between the ‘t’ and the velar consonant (‘c/g’)? It’s because in PIE the sound *n was not infrequently ‘infixed’ within a root. One common use is as a marker of the present tense. Our borrowing ‘tactile’ is an adjective drawn from the past participle of the Latin verb meaning ‘to touch’, showing a root of the shape ‘tag-’ (the *g changes to ‘c’ before the unvoiced ‘t’ of ‘tactile’). The present tense of this verb in Latin has the infixed *n. Indeed, the Latin for ‘I touch’ is ‘tango’.
The same pattern is also visible in the inherited English pairing of present tense ‘stand’ versus past tense ‘stood’.
- Occasionally, roots are ‘reduplicated’. This means that the root is prefixed by the first consonant plus *e. For example, the Latin word that gives us ‘memory’ goes back to a PIE root *mer, which is also found in the reduplicated form *me-mor-.
In order to form words, roots are typically followed in PIE by one or more ‘suffixes’. Some of these have well-defined meanings (for example, the suffix *-ter or -*tor to show someone who carries out an action, as in Latin-derived ‘agitator’). Or they may signal a noun or adjective related to the root meaning, as in the ‘-al’ of pedal and other borrowings from Latin like ‘verb-al’ and ‘norm-al’. It is relatively easy to identify these, just as the English suffix ‘-ly’ creates adverbs from adjectives (quick/quick-ly, for example).
But in some cases, PIE roots seem to be extended by just an extra consonant whose own ‘meaning’ seems obscure. Take for example….
……………The Secrets of your Arm
‘Arm’: an obvious enough word, you might think. But it is also a great example of how Indo-European languages have used a variety of suffixes to build a wide set of meanings on a single root.
‘Arm’ is a word English has inherited from Proto-Germanic. At its heart is a root *ar (we will see in the next blog that this may have originally started with a breathy consonant, which has generally disappeared). The ‘m’ is therefore a suffix, whose actual meaning is not known.
The same ‘m-suffix’ exists also in Latin, where the word ‘arma’ is the origin of our use of the word ‘arms’ to mean weapons, and lies behind familiar words like French ‘Gendarme’ and Spanish ‘Armada’. A similar suffix can be seen in our borrowing from Greek ‘harmony’, which is ultimately derived from a Greek word meaning ‘joint’ or more specifically ‘shoulder’.
Various suffixes all with a *t element give us Latin-derived terms around skill like ‘art’, ‘artisan’, and ‘inert’, with a characteristic Latin vowel change that we will be discussing later, and also ‘articulate’ from a word again meaning a joint. Greek uses a different dental suffix for its main word for a joint, from which we derive the medical term for inflammation of the joints, ‘arthritis’.
All these terms suggest an origin in a sense of ‘to fit together’, which also appears in a less literal sense in our word ‘aristocracy’, from a Greek word meaning ‘rule of the best’. (The ‘-isto-’ element is a typical Greek way of forming of the superlative of an adjective, so the sense would originally have been ‘the most fitting’.) It is even possible that with yet another suffix, this time a *y, the root could also be the origin of the Indo-Iranian word ‘Aryan’, used in Sanskrit to designate the ‘noble’ Indo-European speakers and still seen in the state name ‘Iran’.
Aryan aristocracy, armies, art, harmony and arthritis – that’s quite an armful! But although suffixes distinguish such words from one another, we have no real ‘meaning’ for any of them in this example, whatever may have been the case thousands of years ago.
Vowel Gradation
Proto Indo European had a system of ‘vowel gradation’. This means that vowels routinely changed within a word in order to signal its use – without changing its basic meaning. The relics of this system are still evident in English, for example ‘I sing’ versus ‘I sang’ and ‘well-sung’, and in the difference between the verb ‘sing’ and the noun ‘song’.
In Proto Indo European, a typical form of vowel gradation involved an alternation between a short *e, a short *o and no vowel at all. These are known as ‘e-grade’, ‘o-grade’ and ‘zero-grade’ respectively. As the system gradually broke down, one may find that one daughter language has conserved a form with a single one of these, resulting in an apparent mismatch with a related word in another daughter language. For example, from the PIE root for ‘foot’, *ped-, Latin has based itself on the e-grade forms, visible in borrowings like ‘pedestrian’ or ‘quadruped’; and Greek on the o-grade forms.
So, Question 1: Can you think of any words in English where the element ‘pod’ features with a meaning related to ‘foot’, and which you think might come from Greek (often a source of technical or medical terms)?
Answer, as usual, in the next blog…
The form of word with no vowel at all (‘zero-grade’) needs a bit more explanation.
Where the consonants on either side of the ‘missing’ vowel are both stops or involve an *s, the result is a consonant cluster and the loss of a syllable.
We have already seen an example in Blog 4, where the root behind ‘yellow’ and ‘gold’ (*ĝhel- in its basic ‘e-grade’ form) was also the source of a large number of English words about light and brightness such as ‘gleam’, ‘glitter’, ‘glisten’ and ‘glass’. These all derive from the zero-grade form of the root, *ĝhl-.
Similarly, the word ‘nest’ is made up of a now obsolete preposition *ni-, meaning ‘down’ and ‘st’, which is historically a zero grade of the PIE root *sed- from which ‘sit’ is derived: a ne-st is literally a ‘sit down’ place.
There’s a similar history to some less usual consonant clusters in borrowed words. We have already had a look (Blog 2) at ‘pterodactyl’, a modern scientific term constructed, as often, from words in Ancient Greek. Literally, it means ‘wing finger’. The Greek word for ‘I fly’ is ‘pet-omai’, derived from the PIE e-grade root *pet- (as is ‘feather’, consistently with the ‘f/p’ relationship we saw in Blogs 1 and 2), and the Greek word for wing (or indeed feather) reflects zero grade ‘pt-eron’.
However, when the consonant immediately following the *e in such forms is a sonorant consonant (‘m’, ‘n’, ‘l’, ‘r’, ‘w’, or ‘y’ – see Technical Note 1) and is followed by another consonant (or ends the word), in zero grade the sonorant will be sounded as a vowel, conserving the syllable.
As an example, take the PIE root *bheid-, the origin of English words about ‘biting’. The heart of the syllable is the *e. The following *i completes the diphthong, the tongue moving rather as it does when you articulate an English ‘y’. Its zero grade, with loss of the *e, would be *bhid-, where the *i now assumes the role of the vowel sound. This alternation is indeed still distantly reflected in the difference between ‘bite’ and ‘bit’, which go back ultimately to these two forms.
The same outcome arises when a root begins with a sonorant and ends with a stop consonant. An English reflection of this is in the relationship – another ‘secret’ that is easy to miss – between ‘water’ and ‘otter’. The first syllable of ‘water’ goes back to a PIE *wod-ōr, starting with an o-grade syllable from a root *wed-, from which English has also inherited ‘wet’. In zero-grade, the *o of *wod- disappears and the *w changes to its vowel equivalent, *u. This zero-grade gives us the first syllable of the PIE word for the otter, *ud-ros, which literally means ‘the watery (one)’, the *-ro- suffix being a typical way of deriving an adjective from a noun in PIE. The PIE *u has in turn developed through a sequence of changes into the present English sound.
In both these examples I have chosen semi-vowels (*i/y and *u/w) to illustrate how a zero grade is formed.
What about the other sonorants? You may be surprised that ‘m’, ‘n’, ‘l’ and ‘r’ can also be sounded as vowels, conventionally shown by adding a dot or circle under the letter (‘m̥’ etc). In fact, most of us sound them as vowels in particular circumstances, particularly when talking quickly or colloquially, for example:
- For syllablic ‘m̥’, consider how you pronounce ‘ ’em’ in a phrase like ‘I just love ’em’;
- For syllabic ‘n̥’, try saying ‘often’;
- For syllabic ‘l̥’, try ‘bottle’;
- For syllabic ‘r̥’, a sound pronounced very differently among different dialect groups, it’s a bit harder, but think ‘Brr!’ as an exclamation on a cold morning.
In PIE, these sonorants became syllable-forming vowels in the same positions as the semi-vowels. They are however quite unstable, and in the daughter languages of PIE, they have usually evolved into combinations of a vowel plus the sonorant in question, with the vowel often differing between language groups.
‘Usually’, but not in every case. In Greek and Indo-Iranian, both *m̥ and *n̥ appear as simply ‘a’. Time for:
Question 2: Can you think of a case where some words in English start with a vowel plus a nasal, while in other words (usually of Greek origin) a very similar function is performed by just an ‘a’? Answer in the next blog.
To conclude: vowel gradation reinforces what we concluded from our look at English and German: when looking for matches across Indo-European languages it is safer to focus on consonants than on vowels.
In the next blog, we will set out a standard view of the consonants of Poto-Indo-European, which will then be our guide to more secrets of the words you know.