Saturday, 27 September 2008
This is one of those blogposts which I dithered about writing. Not to put too fine a point on it, a few years ago I became very interested in one of J. R. R. Tolkien’s invented languages, which I still find by far the most gripping aspect of his legendarium. (Fantasy! Elves! Oh Christ - readers who would rather I said something about Early Byzantine art, or offered a Freudian reading of A Midsummer Night’s Dream, or wittered about Kant, had just better switch off now.)
Being the kind of child who always wanted to know what things mean and where they come from, and thus having a very male fondness for structure, pattern and systematisation, Tolkien’s uncannily realistic artificial languages captured my imagination. (Considering my life for a second – strong friendships, a sense of proportion, and a commitment to humanistic education – I reassure myself that I’m probably not anywhere on the Asperger’s spectrum, despite my rampant obsessionalism.*) So let’s press on.
Tolkien, who was a philologist of astounding skill, made up languages throughout his life. (As his friend C. S. Lewis said, ‘it was as if he had been inside language’.) By the time he had written The Lord of the Rings, two of these languages had reached a vast level of order and complexity, easily enough for him to compose texts in them. As he repeatedly explained to incredulous correspondents, the languages were the reason for the stories, not the other way round. The languages came first.
The two most complex of his languages are called ‘Quenya’ and ‘Sindarin’, and form part of a family of languages descending from ancestral language, ‘Primitive Quendian’, or what we might call Proto-Elvish. Though Tolkien devised Dwarfish and Mannish languages too, it was the languages of his immortal, beautiful elves which clearly most preoccupied him. Quenya and Sindarin are thus related, but distantly, having developed among two very widely separated groups of Elves, according to the backstory of the mythos. Tolkien carefully elaborated their vocabulary, tracing the two languages’ descent from the proto-language from a system of ‘roots’. He did this – and this is the point – with such skill and verisimilitude that his languages are analysable with the tools of historical linguistics.
Both languages are designed – as Tolkien specifically said – to be ‘distinctively European in style and structure (not in detail.)’ Quenya resembles Finnish in phonological and grammatical ‘style’, whereas Sindarin resembles Welsh. Ah, there, you see – Welsh. My absolute linguistic first love. Tolkien knew (medieval) Welsh very well, purchasing John Morris-Jones’s A Welsh Grammar when he was an undergraduate at Exeter College, Oxford. Welsh ‘pierced my linguistic heart’, as he wrote, and I had exactly the same experience at a similar age. Sindarin, by far my favourite of the two main Elvish languages, is, in part, a wonderfully complex, lifelong homage to the Welsh language.
The most important thing to get across about Sindarin is the phonaesthetic element behind its creation. It, like Quenya, is designed to be supremely beautiful to the ear. Tolkien found Welsh extraordinarily lovely, a judgement which I share wholeheartedly, and his artificial language has the ‘feel’ of Welsh, as well as its own distinctive quality. (It’s noticeable that the sound –ll-, the famous raspy Welsh double –l, is very much rarer in Sindarin than Welsh. He also specified that the vowel –i- had a slightly longer value than in Welsh, always being more like English –ee-.) The overall effect is of an exquisite fusion of Welsh and Italian.
But Sindarin isn’t cod-Welsh or any kind of knock-off. Almost no words are the same in the two languages, though because the phonologies are almost identical there are numerous words that look identical, but which have different meanings. (For example, nîdh, ‘honeycomb’ (S), is pronounced identically to W Nudd, the father of the mythological character Gwynn ap Nudd. A few are reminiscent: o means ‘from’ in both languages; a means ‘and’ in both; S nand, ‘valley’, resembles W nant of the same meaning. But these are uncommon.) The level of design is astonishingly intricate. Tolkien seems to have worked, as it were, from both ends simultaneously. He made a list of roots, usually of the form consonant-vowel-consonant, which he added to continually. These roots could then be elaborated into words using a number of suffixes, prefixes, and so on, just as historical linguists analyse words in Indo-European, but in reverse. These roots reflect the very ancient words coined by the first elves, but are not quite themselves words. Tolkien then applied a massively complex sequence of sound-changes to these ‘proto-words’, tracing their imagined development through Old Sindarin and Middle Sindarin, until he was satisfied that he had the ‘real’ Sindarin word.
For example, the primitive Elvish root √G-LIB, ‘slippery substance’ gave Sindarin glaew, ‘salve’ (from OS glaibe) and glûdh, ‘soap’ (from OS glūdā, then in turn from gljūdā, older gliudā, back to *glibdā). The latter shows a primitive noun-suffix –dā added to the root; the former doesn’t, giving two different looking words of different meanings. This is exactly how ‘real’ languages work. For a sample of selected vocabulary, go here, which is a slightly out of date discussion of the language.
By his death, Tolkien had elaborated more than twelve thousand words. (Twelve thousand!)
Sometimes he seems to have worked the other way round, ‘hearing’ a fully-formed word and than having to chase it backwards to get to its ancestral root. The whole process is about as far from ‘just making words up’ as can be imagined, and Sindarin has to be called, not an ‘artificial’ language, but a simulated one. In the amazing array of words he devised, Tolkien came up with terms for ‘sudden move’ (rinc), ‘stunted’ (naug), ‘the sound of bells’ (nellad), ‘kingfisher’ (heledirn, literally ‘fish-watcher’) and ‘virginity’ (gweneth).
Indeed, Sindarin and Elvish in general is particularly full of words which somehow feel appropriate, that taste in the mind like the thing they denote. (Go here for a taster of the lexicon of both languages.) I especially like gorgoroth, 'extreme horror', and rhîw, 'winter'. (Reminiscent of Welsh rhew, 'frost'.) Glawar is a wonderful word for 'sunlight', and naneth is just right for 'mother'. (Presumably elf-children called their mothers nana, 'mum'.) Pethron strikes my ear as spot on for 'narrator', and mithren is perfect for 'grey'. Gannada- ,'to play a harp' is lovely, but not as lovely as tinúviel, 'nightingale', literally 'daughter of twilight'. Bizarrely, toloth sounds so right to me for 'the number eight' that I've caught myself accidentally saying it instead of wyth when speaking Welsh. Other favourites are alagos, 'storm of wind', daedelu, 'canopy', and the exquisite gwilwileth, 'butterfly'.
Tolkien closely echoed the Welsh system of consonant mutations, fitting his language out with soft mutation, nasal mutation, stop mutation, liquid mutation and (perhaps) a mixed mutation, each with its intricate ‘historical’ justification. He devised prepositions, several hundred verbs, adjectives, adverbs, exclamations and conjunctions. Forms in his papers are carefully, donnishly marked as ‘poetic’, ‘archaic’, ‘dialectical’, and so on. There are irregular verbs and analogical forms. There are lexical borrowings and a handful of mysterious, unanalysable words.
So can one actually speak it?
Here lies the rub. It seems never to have occurred to Tolkien that other people might take an interest in his private hobby (he went to far as to call it a ‘vice’) after his death. Thus he never put the grammar of any of his languages into a final form, and continued to vacillate and tinker with quite important aspects of the grammar and vocabulary right into the last weeks of his life. He also seems to have written surprisingly little in it - the entire known corpus of texts in Sindarin would fit on two pages of A4. It’s thus important to state that, in a sense, not even Tolkien could speak Sindarin, because it was never finished.
This hasn’t stopped people, and a young linguist named David Salo has published A Gateway to Sindarin, which is a charmingly bizarre, and very learned, grammar of the language, complete with historical phonology, very much like Morris-Jones’ Welsh Grammar which the young Tolkien read with such delight. (A well-informed review is available here.) Salo was responsible for the Elvish dialogues in the Lord of the Rings films, which added so much to the atmosphere of Peter Jackson’s trilogy. People regularly attempt to compose poetry and prose in Sindarin, and I’ve tried my hand at it myself, inspired by that wonderful line in Ulysses, 'the heaventree of stars hung with humid nightblue fruit':
A menel-‘aladh elenath
‘lingannen sí an adanath
na if luin a fuinui,
dannar lemmaid velthin vîdhui,
uin ylf chaer chelegremmin,
taethol nerdh im in elin.
O heaven-tree of stars
(hung now for men
with fruit blue and night-dark)
golden dewy voices are falling
from your distant ice-enwoven branches,
threading knots between the stars.
Ain't it twee.
However, there are massive problems with this kind of thing. The first is vocabulary. Tolkien’s papers, including most of his writings on his languages, remain unpublished, and only emerge into the light of day at a tormentingly slow rate. Very often, one lacks a perfectly ordinary word, such as ‘strange’ or ‘beginning’, or even 'before' or 'after'. We have, for example, almost no numbers above ten. Often, historical linguistics can ride to the rescue, and one can use Tolkien’s own system of roots, and equivalent words in his other Elvish languages, to come up with a logical Sindarin equivalent. For example, we know the Quenya word for ‘therefore’ was etta; the Sindarin one should be *eth, but that is completely unattested. We know that the word for 'leg' in Quenya was telco; it should be telch in Sindarin, but there is no reason why a resurrected Tolkien might not look at our use of telch as 'leg' in disgust and say: ‘Telch, indeed the cognate of Quenya telco, was used only for the downstroke in a letter when writing with a pen, the 'leg' of the character, as it were. The word for a bodily leg was tudh.’ (Or rhasal. Or elmath. Or whatever he liked.) In other words, though we can reconstruct entirely logical new words based on the extant roots and cognates, we are instantly thereby writing 'Neo-Sindarin' or even 'pseudo-Sindarin', because the only person who could tell us whether such a word were valid or not has been dead for thirty years.
Further, the grammar has huge gaps in it. We have hardly any of the verb 'to be'. Until this year, no one knew what the 2nd singular person verbal ending was. People had hypothesised –ch, -g, or –l (-ch was used in the films). Then a new tranche of Tolkien’s papers was edited and made available, and it turned out that there were two: a polite form, ending –g, and an informal form, ending in –dh. The ending –ch did in fact exist; but it turned out that it was the 2nd person dual formal form, ‘you two…’ The verbal system was both much more complicated and rather different from what we had thought, on the basis of the pitifully small number of examples which were attested before last year. The Sindarin verb, it was revealed, had dual as well as singular and plural inflections, and there were at least four ways in which the past tense could be formed, depending on the verbal class, and whether the verb in question was transitive or intransitive. Also in this new bundle of material, we learned of a whole slew of hitherto unknown words, including a verb ‘to have’, at long last. (It was sevin.) The lack of a word for ‘if’ is particularly annoying.
Further, we remain very short on pronouns, and could do with more explicit material about the forms of adverbs and the comparison of adjectives, as, at present, we don’t know how to do ‘good, better, best’. In this area, a form occuring in LOTR led us up the garden-path: Elrond refers to Tom Bombadil - Tolkien's most irritating creation - as Iarwain ben-adar, 'oldest and fatherless'. We know that pen-adar must mean 'without father', and Iarwain contains a reduced form of the adjective iaur, 'old'. But the -wain was mysterious, and given Elrond's gloss, many took it to be a superlative suffix equivalent to English -est. So people started creating forms like celegwain, 'swiftest', from celeg, 'swift'. The trouble is, -wain looks very odd, and it was suggested by some bright spark that it might in fact be the adjective (g)wain, 'new, young'; hence Iarwain would literally be 'Old-young', 'One who is ancient yet has not aged', hence Elrond's not-very-accurate translation, 'oldest'. An unpublished letter of Tolkien's from 1968 showed this surmise to be absolutely correct, and the useful superlative suffix -wain had to be consigned to the bin. Quenya suggests that there would be one intensive form made by prefixing an intensifier an- to an adjective; used attributively this would be the comparative, and used with a following genitive it would be the superlative; e.g., adan geleg, 'a swift man', adan angeleg, 'a swifter man', angeleg in edain, 'the swiftest of the men, the fastest man'. This is vaguely similar to the way the Gaelic languages do the comparison of adjectives, incidentally. But even so, judging by the Indo-European tongues, the intensified forms of adjectives like 'good', 'bad', 'big', little' etc. are unlikely to be regular: compare Latin bonus, melior, optimus, Welsh da, gwell, gorau, 'good, better, best'.
As in all communities of tiny specialisms, there are nasty spats. As Gore Vidal said of student politics, the arguments are so vicious only because the stakes are so small. It’s much resented that the Tolkien estate will release the linguistic papers only to the small group who publish the journal Parma Eldalamberon, who seem to look down rather on people who attempt to compose in Tolkien’s languages. (Often with some logic.) Their leisurely pace of editing is a notorious frustration, especially as the material they possess can, as shown above, utterly transform our understanding of the language(s) overnight. Around the time when David Salo was working on his grammar, the editors of Parma Eldalamberon chose to release, for example, a delightful bit of lore about the nicknames Elf-children used for the fingers and thumb. Charming, of course, but at the same time, they were sitting on a much fuller account of the verb system than any in the public domain, which they chose not to edit for another couple of years.
So. A harmless hobby for a harmless drudge; and I tend to think of it as a homeopathic dose of philology - a drop of made-up grammar steels the mental immune system to cope with the grammar and phonology of Middle Welsh and Old Irish. It’s like sudoku for the linguist – totally non-functional, but terribly involving if you have that kind of mind. What I like about it is the idea that the hidden texture of language itself - grammar, syntax, phonology - can be an art-medium, quite abstracted from anything actually written in that language. 'Philological Beauty' - the Greeks should have had a muse for it. What is remarkable is that one only has to lay down a fairly small number of 'rules' before the inherent forces of linguistics start generating ever more beautiful and ever more complex matter. So it becomes like discovering, rather than inventing, a language. Something rather similar occurs with people who try to revive Cornish; they have to invent new words, for concepts and objects which didn't exist in 1776 when the last speaker expired, but they also have to supply words which must have existed but which haven't happened to have been preserved in the corpus. From Welsh rhath and Breton razh we know with certainty that the Cornish word for 'rat' was rath - but it doesn't survive anywhere. Similarly, it's very likely that 'to praise' was *daetha in Sindarin, because that would be the philologically-justified cognate of Quenya laita, both from Proto-Elvish *daitā and assuming that the stem is Sindarin was augmented with a common verbal suffix -ta.
And, of course, as Tolkien intended, Sindarin can often be hauntingly beautiful. In this, its creator admirably succeeded in his avowed aim. For a little piece of evidence for this assertion, here's a clip from the soundtrack to The Fellowship of the Ring; it’s a hymn in honour of Elbereth, the ‘goddess’ of the stars in Tolkien’s mythology. In the novel, Tolkien gave an English version – ‘Snow white! Snow white! O Lady Clear!’ – which David Salo translated into Sindarin, miraculously keeping the metre and rhyme scheme of Tolkien’s original.
* All my books – about 800 in a small flat – have to go on the shelves horizontally. I like looking at the calming beige of the ends of the books rather than all those different-coloured, clashing spines. Of course, I can never find anything.