SHETLANDIC ORTHOGRAPHY - PROCESSES AND PRINCIPLES
The following is a description of the processes and principles involved in creating an experimental orthography for the Shetlandic language as it might exist in a parallel universe.
The Shetlandic orthography is here presented as a fait accompli, that is, as if it were the real orthography of the language - which is real. This is largely to avoid the current situation in which dialect spellings tend to emphasise dialect variants rather than the underlying unity of the phonological system, giving an impression of almost random variations from standard English. By using an orthography which starts with the language itself, rather than presenting it as alterations to something else, this approach treats it as most other languages (though not English) are treated in writing. As well as illustrating how the phonology can be represented in writing, this serves as a practical description of that phonology, which is otherwise obscured by the constant intrusion of English norms.
This is in contrast to the present situation, where academic linguistics typically treats Shetlandic as a dialect of Scots, and presents it in terms - such as phonetic script, and the Wells word lists based on the categories of standard English as spoken in southern England - which make its characteristics opaque to anyone who is not trained in these methods. In this approach, the orthography and phonology reinforce each other, just as they do for literate speakers and learners of most languages. In this way, the orthography becomes an interface between the written and spoken tongue.
However, in inventing the orthography, decisions had to be made. This section explains the process which led to the orthography.
The following discussion uses some technical linguistic vocabulary, but avoids phonetic script, which is not widely understood. Instead, it uses approximate English pronunciations in inverted commas - e.g. ‘oo’ for the sound in boot, and ‘j’ for the sound in jam, rather than the equivalent phonetic [u] and [dʒ]. The underlying vowels (diaphonemes) of the Shetlandic sound system are given in capitals as they are spelt in the Sjætlan orthography - e.g. Á and Ǽ rather than /ɑ:/ and /e:/. This also avoids pinning them down phonetically, since they are often pronounced differently in different Shetlandic dialects. Words in the Sjætlan orthography are in bold type, e.g. bad, bad, hæm, home, with the English meanings following. Words from English and other languages used as examples - e.g. boot and jam above - are in italics.
Some technical terms are listed at the end of the article.
DEFINING THE LANGUAGE
Before you can start to create an orthography for a language, you have to define what that language is.
Shetlandic is here conceived of as a language in its own right, in contradistinction to the narratives which represent it as the sad remnants of Norn, or a dialect of English or of Scots, which is itself conceived of as a continuum of dialects. As such, all the words used by Shetlanders in traditional speech are part of that language, whatever their etymological source, and the same basic spelling system is used for words of Norn, Scots and English origin. Exceptions are necessary for some words of Latin, Greek and French origin, especially technical terms.
On the other hand, words that are not part of traditional speech and do not fit into the Shetlandic grammatical system - such as much and which, or literary Scotticisms such as leid and owerset - are not regarded as being Shetlandic, although they can still be spelt with the same orthography. The same is true of English words with close Shetlandic cognates, such as home (Sj. hæm) roof (Sj. røf) and take (Sj. tak). English-derived variants such as hǫm for hæm and bǫn for bæn can obviously also be spelt with the same system.
This is in contrast to the narrative which regards Shetland dialect as consisting only of those words which are peculiar to Shetland. So on forums, it is common to hear that any word of Scots origin - such as glig, intelligent - is not a ‘Shetland’ but a ‘Scottish’ word. This is part of the prevalent dialect narrative, and has the effect of reducing what is conceived of as Shetland dialect to those words which are not shared with any other language. Obviously, languages do not follow this exclusionary approach, or none of the Nordic or Romance languages would consist of very much, as they all share most of their vocabulary with other languages in the same family. Here, words shared with English like man and net, Scots ones like haem, home, and tráchel, struggle, Norn ones like sjalder, oyster catcher, and sjarg, nag, French ones like fasj, bother (via Scots) and byf, beef (via English and Scots), Latin ones like spyrít, spirit, and Greek ones like baptís, baptise, are all regarded as Shetlandic words, and combined in a single orthographic system.
On the other hand, it would seem pointless to include, in a definition of Shetlandic, words for which there are traditional Shetlandic equivalents, such as palm (of the hand, Sj. løf) or stream (Sj. streip, born). That, however, is a wider issue, and does not directly affect orthography.
ORTHOGRAPHIC PRINCIPLES
Starting with the Phonology.
Dialect spellings typically start with English spelling and represent a particular Shetland dialect by varying from that. The result is that the characteristics represented are not necessarily those intrinsic to Shetlandic, but those which are perceived as being different enough from English to be represented in writing. The letter Ö is the only obvious exception, and even there, there is no way to distinguish the long vowel in e.g. brǿl, bellow, from the short one in e.g. støl, stool. This is presumably because, vowel length not being an important consideration in English spelling, this characteristic of Shetlandic is simply not recognised, with the result that the spelling does not reinforce it. Other essential characteristics - like the sound of the Æ vowel in words such as pæt, peat, læk, like, ræt, rate - do not have a consistent spelling, with inconsistent spellings such as paet, lik and rate - presumably perceived as minimum alterations to English - being typical. Also, long and short vowels are often spelt with the same graphemes, as in rate above, which has a short vowel, and hate, which has a different, long vowel. This approach has been adopted in Shetland as part of the dialect narrative, inculcated by the Shetland media and intelligentsia, which repudiates any attempt to raise the status of Shetlandic by giving it a name other than ‘dialect,’ or representing it as an entity in its own right.
While it would have been possible to have represented the characteristics of Shetland phonology using English-like spelling conventions - such as UI for the short and OE for the long ‘ö’ sound, or consistently using AE for the vowel in words like ‘paet’ and ‘laek’ and A-E or AI in ones like ‘bait’ and ‘hate’ - this would have required the acceptance of a series of spelling conventions. Since it is precisely this sort of systemic approach which is repudiated under the dialect narrative, and English-like spellings are perceived as belonging to that narrative, this orthography uses an approach which avoids confusion with English spellings as far as is practical.
Diaphonology.
Treating the language as a whole means that the spelling shows distinctions that are made in most Shetland dialects - and therefore in Shetlandic as a whole - and doesn’t show differences in pronunciation between dialects. So for example, it shows the difference between fat, fat, and fát, fault, which are pronounced differently from each other whatever dialect you speak, but not the fact that fát is pronounced ‘faat’ in some dialects and ‘faut’ in others. It is assumed that the letter Á has these dialect variations. This leads to the principle of diaphonology, meaning that the same written letters are pronounced differently in different areas, and the same spelling can be used for all dialects.
Etymology.
Language features which come from different sources are also spelt under the same system as far as possible. So both the long vowel in býr, beer, from Scots and English, and that in nýb, nod off, from Norn, are spelt using the same letter. Occasionally these coincide, as in stýr, steer, govern, cp. both Scots and English steer, and Icelandic stýra. Exceptions to this are largely words of Greek, Latin, and sometimes French origin, such as osteopath, quadrilateral, exfoliate and audiens (here re-Latinised from French audience.) More common Latinate words, such as spyrít, spirit, and næsjon, nation, are conformed to the default system. All of these words are regarded as Shetlandic words, just as, in English, they are regarded as English words irrespective of their source.
Homophones.
Since words are spelt phonologically, words that sound the same - homophones - are usually spelt the same, e.g. stýr, steer, and stýr, stir. Likewise rig, spine, rig, cultivated field, and rig as in oil rig. These don’t confuse us in speech, so there is no reason why they should do so in writing.
Frequent Words.
Very frequent words such as pronouns and prepositions sometimes do not conform to the usual phonetic rules. They are usually spelt according to their stressed pronunciations, although some words also have unstressed forms - e.g. maj, unstressed mi, my. The acute is usually omitted on final vowels in such words, e.g. dæ, they, sjø, she, even though their stressed pronunciations are long.
Morphology.
Some spellings are grammatical or morphological rather than phonological. So the final -an and íng for the present participle and verbal noun are not pronounced that way in all dialects. Endings such as -tion, in words such as nation, may be conformed to the system, giving e.g. næsjon, nation.
Diacritics
The Acute. The main diacritic is the acute - e.g. á - which mainly shows that a vowel is long, as in the following list of words containing short and long vowels.
bak - back
bák - balk, beam
hæt - heat
hǽt - hate
bot - living room
bót - boat
nyp - swede (veg)
nýb - nap
lup - loop
úb - whimper
gød - good
rǿð - talk nonsense
The remaining vowel phonemes - E, I and Ǫ - have no long forms.
The acute is the only diacritic which is part of the actual spelling, in that it can distinguish one underlying vowel, or diaphoneme, from another. Other diacritics are only occasionally useful, e.g. in word lists. They are not regarded as part of the actual spelling, as the sounds they mark are automatic to native speakers.
The Grave. The grave e.g. - à - can be used to emphasise a vowel. This may be because of unusual stress - as in miràkel, ‘maim,’ or the lengthening of open vowels which occurs especially before voiceless fricatives, as in grèth, urine, and grǫ̀f, rough.
The Trema. The trema (Umlaut) or soft sign - e.g. ä - can be used to show the differing pronunciations of vowels which occur mostly before voiced consonants, in e.g. bäd, bad, bëd, bed, hæ̈d, head, rïn, run, rǫ̈d, road. These pronunciations are mostly automatic to native speakers, and vary a lot between dialects, so this sign is not usually necessary.
The soft Æ̈ vowel merges with E in Central Mainland type dialects, and with Y in North Isles type dialects, having a distinct pronunciation only in certain areas. So hæd, head, which has a distinct pronunciation in e.g. Burra and South Dunrossness, is pronounced ‘hed/heyd’ in the Central Mainland, rhyming with e.g. bed, bed, and ‘heed’ in the North Isles, rhyming with ryd, read.
Exceptional Usages
The letters E and I have no contrastive long forms, and the acute may be used to show that they are pronounced as ‘ee,’ mostly when I is unstressed, or in Latin and Greek words familiar from English, like rhésus, sérum, abílity and spyrít.
The letters U and Y do not have soft equivalents, and the trema may be used to show the English pronunciations of Y and U in words which are not otherwise conformed to the Shetlandic orthography, as in Greek nÿmph and Latin calümny.
THE PROCESS
Phoneme-Grapheme Allocation.
The first step is to assign a letter to each of the underlying phonemes. This could have been done in various ways, but because of the historical connections between Shetlandic and the Nordic languages, and the fact that Old Norse spelling conveniently represents many of the characteristics of Shetlandic phonology - especially vowel length - I chose Old Icelandic as the basis.
This should be distinguished from other Nordic-like spellings which are primarily ideological rather than phonological. Before the Shetland Dictionary, many dialect writers used some Nordic like spellings, but these were usually based vaguely on Norwegian, were not consistent, and did not represent the phonology of Shetlandic accurately. This is the only Shetlandic spelling system which starts from the phonology of the language itself. Because of its Norse origins it has a certain resemblance to Jakobsen’s spellings in his dictionary of Shetland Norn. But he was interested only in Norn, not in the living language, and his spellings do not consistently represent essential features such as vowel length, for which he relied on phonetic notation.
In the case of vowels, a single letter, with or without an acute diacritic, is assigned to each vowel. Old Icelandic has more than enough letters to do this, but assigning them according to the Old or Modern Icelandic pronunciations does not always produce the most practical results. Some spellings - such as A, Á, Ø and Ǿ below - are obvious. Others require a reshuffling of letters away from the rather bizarre English conventions towards those used in Nordic - and, indeed, most other - languages. Vowel digraphs - e.g. AI, EI - are used only for double vowel sounds, or diphthongs.
Consonants are at first sight easier to identify than vowels, and most - such as P, T, K, B, D, G - can be assigned easily. But there are sounds and characteristics of Shetlandic which require consonant combinations - e.g. the ‘sh’ and ‘th’ sounds - and also some consonant sounds which vary between dialects - especially the palatalised K and G which become ‘ky’ and ‘gy’ sounds in some dialects, and ‘ty’ and ‘dy’ sounds in others. Like the vowels, these are generally spelt according to broad Nordic conventions, with adaptations to the Shetlandic sound system.
Exclusions
The next step is to decide which aspects of both English and Old Icelandic spelling are not necessary.
Silent Letters. There is generally no need for silent letters. In particular, the following E, as in English mate as opposed to mat, and hate as opposed to hat, is potentially confusing, particularly as mate and hate do not rhyme in Shetlandic. These are replaced by single and unambiguous vowel spellings - mæt, mat, hǽt and hat. The only silent letters used are W before R in words like wreit, write, wrek, wreck, and some that may be assimilated or omitted in certain dialects, such as the K in e.g. knap, speak English, H in words like feht, fight, tǫht, thought, or the -EN in ones like lipen, expect, which becomes ‘lipm’ in some accents.
Redundant Letters. There is no need for the letters Q and X, except in Latinate technical terms such as quadrilateral and exfoliate. X is merely K followed by S. Words which begin with QU in English - such as queen, quick - merge with words beginning with WH such as what and where. These follow a traditional pattern where they are all pronounced like ‘wh’ in some districts and like ‘qu’ in others. (This is also the case in modern Icelandic, where similar words are written with hv but pronounced ‘kv’ in some accents.) In the Shetlandic orthography, these words are all spelt with HW, giving hwyn, queen, hwik, quick, hwit, what, hwár, where.
(For my own use, I have configured a switchable Shetlandic keyboard in which the Q key produces the letter Ǫ, the X key the letter Ø, and the # key the letter Æ. The right and left square bracket keys produce the acute diacritic and the letter ð, the \ key the grave diacritic, and the ¬ key the soft sign (trema.) This makes typing the Shetlandic orthography as easy as with any other language which uses the Nordic characters.)
Double Consonants. In Old Norse, double consonant spellings showed that the consonants had long pronunciations. In the modern continental Scandinavian languages - Danish, Swedish and Norwegian - they also show that the preceding vowel is short, so that diacritics which show vowel length are no longer required in these languages. The underlying criterion is the syllable, which can either be Short Vowel + Long Consonant or Long Vowel + Short Consonant. Modern Icelandic and Faroese follow a similar pattern, in that the diacritics, which used to show vowel length, now show different vowel qualities, and any of them can be long or short depending on the following consonant.
Shetlandic has a similar syllable structure to the Scandinavian languages, and like those, it is not necessary to show both vowel length and consonant length, because a short vowel will always be followed by a long consonant and vice versa. (There are some exceptions to this in the spelling of the actual Scandinavian languages.) It might be possible, therefore, to use a similar method to the Scandinavian languages, giving e.g. bakk, back, but bak, beam, and hætt, heat, but hæt, hate. However, unless you are already familiar with this method, it is much less direct and intuitive than indicating the vowel length on the actual vowel, as bak, bák and hæt, hǽt. And because it is vowel length and quality that is phonologically important, and there is no need to show both vowel and consonant length, double written consonants are unnecessary.
In English, double written consonants alter the pronunciation of preceding vowels in the opposite way to silent following E, as in the examples hat, hate, hatter and hater. Since following E is not used in the Shetlandic orthography, and the vowel in e.g. hate and hater is spelt differently - Ǽ - from that in hat and hatter - A - double consonants are not necessary for this purpose either.
Exceptions
Since the intention is to produce a realistic orthography rather than a transcription system, there are certain classes of words which are more practically treated as exceptions.
Frequent words. Small words used very frequently in speech - especially pronouns and prepositions - tend to have unstressed or anomalous pronunciations. Only a few of these are distinctive enough to warrant a different spelling - e.g. mi and di as unstressed maj (my) and daj (your - singular familiar). Others are spelt without the long sign on the final vowel - e.g. dæ, they, as opposed to dǽ, day - which better accomodates both stressed and unstressed pronunciations.
Borrowed words. Words of Latin, Greek and French origin often have spellings and pronunciations which conflict with the rest of the orthography. Generally, more frequently used ones such as nation are conformed to the system - so næsjon - while technical terms either retain their English spelling, or modify it in certain respects - e.g. audiens, a re-Latinisation of French-derived audience.
Proper names. Proper names in the Scandinavian languages are often spelt differently from the rest of the language system - e.g. Carlsen rather than Karlsen, Sundqvist rather than Sundkvist - and the Sjætlan orthography adopts an approach where traditional spellings may be retained (e.g. George, Geordie) conformed (Cjórgj, Cjórdi) or modified (Ghórgj, Ghórdi).
VOWELS
Shetlandic has 15 vowels - phonemes - which need to be distinguished from each other, and six diphthongs. The vowels are the 12 underlying vowels of Scots - the Aitken or Abercrombie vowels - plus a further three which are phonologically long in Shetlandic but not in Scots, which has no phonological vowel length. The six diphthongs are also of Scots origin - also given Aitken numbers - but Mainland Scots dialects have lost some of these, and most have also lost some of the twelve vowel phonemes. So Shetlandic has a more extensive phoneme inventory than Mainland Scots, consisting of sounds which have been lost in most Scots dialects, plus further distinctions which heark back to Norn.
Two particular features of Shetlandic vowel phonology and phonetics which have to be taken account of are vowel length and soft mutation.
Vowel Length
Vowel length is intrinsic to Shetlandic phonology, and is sometimes the only difference between words. While it would have been possible to represent this by English-type conventions, this would have required mutual recognition by both writer and reader, and the dialect narrative militates against this. In the Sjætlan system, long vowels are indicated explicitly by the acute diacritic, e.g. bák, beam, hǽt, hate, rǿð, speak nonsense. (Note that in some dialects the main difference may be in quality rather than length, e.g. in the North Isles where Á is pronounced as ‘aw,’ or Whalsay and Fair Isle where Ǽ is pronounced as ‘uy.’) Final vowels are long by default, and the acute can be omitted on frequent words such as fu, how, distinguishing them from lexical words such as fú, full.
Shetlandic long vowels are phonological - that is, they can carry meaning, and are regarded as different letters from their short counterparts. So fat, fat, is distinct in both meaning and pronunciation from fát, fault, and hæt, heat, is similarly distinct from hǽt, hate. The long vowels are sometimes derived from Scots (usually Scots rather than English) - e.g. fát from Scots faut rather than directly from English fault - and sometimes from Norn, e.g. the long vowel sounds in nýb, nod off (Old Norse hnípa), rúg, heap (hrúga). Some long vowels come from lengthening in Norn before single consonants, as in the modern Nordic languages, e.g. klág, cackle, from Old Norse klaka. (Note that long Norn vowels do not occur before e.g. P, T and K, because these typically changed to B, D and G in Norn, as in hnípa to nýb.) Some long vowels come from the Scottish Vowel Length rule, which however is not fully predictable in Shetlandic. The Shetlandic orthography treats these all alike, because they all form a single coherent Shetlandic vowel system.
Soft Mutation
Soft mutation is a change in vowel quality which occurs mainly before voiced consonants. So the vowels in bad, hæd, led, lid, rǫd and kod (bad, head, laid, lid, road, cud) are usually pronounced differently from those in bat, sæt, set, pit, rǫt, kot (bat, seat, set, put, rot, cut.) As these pronunciations (technically, allophones of the same phoneme) are automatic to native speakers and vary a lot between dialects, they do not need to be shown in the spelling, but can be shown - e.g. in word lists - by the trema (‘Umlaut’) or soft sign, e.g. bäd, hæ̈d, lëd, lïd, rǫ̈d and köd. Those on long vowels may be shown by replacing the acute diacritic by the Hungarian double acute, e.g. kla̋g (cluck, etc) a̋bir (keen, sharp) for words otherwise spelt klág and áber.
A and Á
The letter A naturally represents the sound in words like ‘bak,’ back, and ‘hap,’ shawl, and the long sound in words like ‘baak/bauk,’ beam, ‘faat/faut,’ fault, can then be represented by Á, with the acute diacritic, giving bak, back, hap, shawl, klag, to stick, bák, beam, fát, fault, klág, cackle.
Ø and Ǿ
Ø is the Danish, Norwegian and Faroese sound and spelling, as in gød, good, røt, root, støl, stool. The long sound again has the acute diacritic, as in lǿr, pollack, rǿð, speak nonsense, brǿl, bellow.
E, Æ and Ǽ
In Old Icelandic, the Æ letter was always long, and in modern Icelandic, it represents a diphthong. In Norwegian and Danish, it represents a more open vowel than E. But in Shetlandic, it is an obvious choice for the vowel often spelt AE in dialect spellings, as in ‘paet’, peat, which is a closer vowel than E, giving e.g. pæt, peat, hæt, heat, with the acute showing the long vowel, as in hǽt, hate, bǽt, bait. (Old Norse Æ usually became this sound in Shetlandic.) This leaves the E vowel for the more open sound in kep, catch, gret, wept, including words with identical English cognates such as set and pet.
U, Ú, O, Ó and Ǫ
If U is used for the ‘oo’ sound - as in most languages, including Nordic ones - and O for the Shetlandic pronunciation of words like English ‘luck’ and ‘rug’ - which is how that sound would naturally be spelt in Nordic languages, e.g. by Jakobsen - that leaves the open ‘o’ sound in words like ‘pot’ and ‘rod,’ road. The Old Norse letter Ǫ - not used in any extant Nordic language - fills the gap. The acute sign is used to show long O and U vowels. This gives e.g. rum, room, strup, spout, stúr, dust, rúg, heap, lok, luck/entice, brok, rubbish, óg, crawl, vór, spring, pǫt, pot, rǫd, road.
I, Y, Í and Ý
In Old Norse, Y and I represented different vowel sounds. The Y sound survives as a separate sound in the continental Scandinavian languages, but in the otherwise more conservative Icelandic and Faroese, this sound has merged with the sound of I, so that Y and I sound the same, and Ý and Í sound the same. To represent the three relevant sounds - the ‘i,’ ‘ee’ and long ‘ee’ sounds - of Shetlandic, the letters are assigned as follows:
I represents the ‘i’ sound, as in pit, put, fin, find
Y represents the short ‘ee’ sound, as in ryd, read, slyp, sleep
Ý represents the long ‘ee’ sound, as in nýr, near/kidney, nýb, droop, nod off
Unstressed I is pronounced ‘ee’ at the end of a word, as in muti, tiny, pǫki, paper bag, and normally elsewhere, e.g. difend, defend, ǫfiser, officer, spyrit, spirit. Because I has no long form, this can be shown by the acute, e.g., dífend, ǫfíser, spyrít.
Unstressed Vowels
As in English and Scots, any vowel may be reduced to the ‘shewa’ sound - the sound of ‘e’ in English ‘the’ - in unstressed positions. In theory, any vowel in an unstressed syllable may be pronounced in this way. However, the Shetlandic orthography adopts some conventions for representing these unstressed sounds, based partly on morphology, and partly on analogy with morphological endings.
A is used
Mostly initially and finally, e.g. da, the, afór, before, ahint, behind, fjilska, merriment, wista, wilderness.
The present participle ending - as in finan, finding, hǫkan, digging - is spelt -AN. This has the ‘-in’ pronunciation in many accents, but is pronounced with a clear A sound in some accents - e.g. Fair Isle - where it is distinguished from the verbal noun. The verbal and other similar nouns, which are also pronounced ‘-in’ in many accents, are pronounced ‘eeng’ or ‘een’ in others, and are written -ing (or íng), e.g. mórning, morning, ývning, evening.
E is used in most other circumstances, including
Comparatives and superlatives - pyrier, smaller, wyter, wetter, and pyriest, wytest.
Before N in the past participle of strong verbs - rǫten, rotten, soken, sunken - and in many analogous words - lipen, expect, strenten, strengthen.
In the plurals of nouns which require an extra vowel - e.g. pys-es, pieces, hǫrs-es, horses, las-es, girls - and in the 2nd and 3rd singular of verbs which require an extra vowel, e.g. wasj-es, washes, pres-es, presses.
Note that this helps to distinguish pronunciation from the plurals of words ending in -I which are pronounced differently, e.g. keisji-s, baskets, pǫki-s, paper bags
Before T in the past tenses and past participles of weak verbs ending in ‘p, t, k, b, d’ or ‘g’ sounds, e.g. nýbet, nodded off, lypet, parboiled. Also in adjectives of similar construction, e.g. plypset, whinging, slúmet, underhand.
Before S in the possessives of nouns which require an extra vowel, e.g. las-es, girl’s, greis-es, pig’s.
Before R in occupational descriptions, e.g. sǽler, sailor, and analogous words like ofíser, officer.
Since any vowel may be reduced to shewa when unstressed, there are many words - including proper names - where none of these apply, e.g. Magnus, Sinklar, Sinclair, nǽval, naval.
Diphthongs and Digraphs.
Diphthongs, or double vowel sounds, are spelt logically using two letters, or digraphs. At the ends of words and derivatives of such words, I is replaced by J. Note, however, that EJ has a different sound from EI.
AI / AJ - the sound of I in ‘five’ - airn, iron, faiv, five, faier, fire, traj, try, trajs, tries,
EI - the sound of I in Scottish English ‘fine’ - eil, isle, fein, fine, teil, tile
EJ - similar to the sound of ‘a’ in RP (posh English) ‘lake’ - ej, always, hej, hay
ØI / ØJ - Ø followed by I - øil, oil, grøint, grunt as made by a pig, bøj, boy
ǪI / ǪJ - like ‘oi’ in English ‘noise’ - nǫis, noise, rǫid, unruly, fǫj, festival
OW - like ‘ow’ in English ‘how’ - ows, bale out, sjow, chew, bow, buoy.
Note that the above are phonological diphthongs - those which are recognised as separate phonemes in the Shetlandic vowel system as a whole, and as such have a distinct spelling. Other vowels may be pronounced as diphthongs in certain positions - e.g. the hard pronunciation of Ǽ is often followed by a shewa sound in words like hǽr, hair, and bǽt, bait, and in Whalsay is pronounced as an ‘uy’ diphthong. Soft Ë is pronounced as ‘ey’ in many dialects in words like bed, bed and neb, beak. But these are regional variations - in Cunningsburgh, bed and neb are pronounced as single sounds, or monophthongs, and in Yell, words like bǽt are pronounced with a long ‘eh’ sound. As such, these phonemes are not spelt as diphthongs in the orthography.
I+Vowel digraphs like IU are explained below under Palatalisation.
CONSONANTS
Most standalone consonants can be used and pronounced much as in English. However, there are certain characteristics of Shetlandic consonant sounds which have to be taken into consideration in designing an orthography.
Devoicing
There are several sounds which are voiced in English and Scots which tend to be devoiced, or partially devoiced, in Shetlandic.
The English ‘j’ sound at the beginning of words is typically devoiced to the equivalent of the English ‘ch’ sound. This is spelt CJ by default, and may be spelt JH or GH in proper names, e.g. cjǫb - job, Jhǫn - John, Ghórdi - Geordie (George)
There is a general tendency also for words pronounced ‘ch-’ in English to be pronounced ‘sh-’ in Shetlandic, but this is not universal. So in my accent, sjǽr, chair, sjǫk, choke, but cjek, check, cjýs, cheese. Again these vary according to district, and it is not clear whether anomalies like cjǽn vs sjǽn, chain, are owing to regional variation or more recent Anglicisation. A combination spelling, scj-, can be used where the the pronunciation varies, e.g. scjængj, change.
V and Z tend to be devoiced to F and S in certain accents. In my accent, kliv, hoof, sounds like English cliff. This seems to correlate with the presence of short vowels in words like hyven, heaven, and dyvel, devil, which have long vowels before V in Scots, according to the Scottish Vowel Length Rule. However, since the vowel is marked as short by the absence of the long sign, and since the V seems to be voiced in some accents, it is retained in the orthography.
Dark and Clear L
Unlike Scots and Scottish Standard English, Shetlandic has both a ‘dark’ and ‘clear’ L sound. Generally speaking, the dark L follows back vowels such as Ǫ, O and U, and the clear L follows front vowels such as E, I and Y. However, the soft allophones of some back vowels are front vowels, and this creates certain regional differences in pronunciation. For example, the word hil, hill, is pronounced in some areas with a hard vowel + dark L, and in others with a soft vowel + clear L. However, it is the same word, regional pronunciations are automatic, and there is therefore no need to differentiate dark and clear L in writing.
Palatalisation
Palatalisation is when consonants are followed by a ‘y’ sound, often before a front vowel such as I or E, which can then become another sound. In English, this is how words like church came to have ‘ch’ sounds rather than the original ‘k’ sounds which are preserved in Scots and Shetlandic kirk. This also occurs in the Nordic languages, which is why K is often pronounced with a sound similar to ‘ch’ in Swedish, Norwegian and Faroese.
In Shetlandic, this varies according to dialect. A word like *kemp, compete, may be pronounced ‘kemp,’ ‘kyemp’ or (approximately) ‘tyemp’ in different regions. The same is true of words beginning with G. Most other consonants may be followed by a ‘y’ sound, although this does not lead to such obvious sound changes. For example, the word I know as lúmi, an oilslick, is recorded by Jakobsen as ljumi. On the other hand, the cognate of English duck may be pronounced ‘dyook’ or ‘jook.’
Another, ostensibly similar phenomenon is the variation in pronunciation in words like book (‘book’ or ‘byook’) took (‘took’ or ‘tyook’) etc. This, however, is the Scots vowel sometimes spelt EU, as in beuk, which is what the original UI vowel - normally Ø in Shetlandic - turned into before K and CH. The variation between ‘oo’ and ‘yoo’ pronunciations is probably owing to the influence of English, which only has an ‘oo’ sound in cognate words.
Effectively, there are two stages of variation - in whether the initial consonant of a word is pronounced with a following ‘y’ sound (‘kemp’ or ‘kyemp’, ‘book’ or ‘byook’) and in whether that then turns into another sound (‘kyemp’ or ‘tyemp’, ‘dyook’ or ‘jook.’). As this is to some extent a lexical consideration, the orthography adopts a heuristic approach using the following rules of thumb.
Consonant + J spellings are treated as consonants, and are generally used where the pronunciation is felt to be characteristic or established, e.g. gjó, steep inlet, mjáchtles, weak.
I + Vowel spellings are treated as vowels, and are typically used for the etymological UI/EU vowel before K and CH - e.g. tiuk, took, liuch, laughed. They may be used as an equivalent to Consonant+ + J spellings for other words where the pronunciation is felt to be more variable, e.g. kiemp, liúmi.
Single Consonants
B, D, M, N, P, T, and W are pronounced more or less as in English. The Shetland pronunciation of R does not affect the spelling.
Q and X
Q and X are not used, except in technical Latinate words like quadrilateral and exfoliate.
C, K and S
C is strictly speaking unnecessary, being mostly replaced by K and S in reasonably common words, e.g. sement, cement, kat, cat. As in Scandinavian languages, it may be used in Latinate words, especially when technical terms, such as calümny, cerebral. It occurs in the combinations CH and CJ, discussed below.
K followed by a front vowel or J is pronounced as a palatal consonant - approximately ‘ty’ - in Whalsay, as often in the Nordic languages.
G, J, and Y
J has the English ‘y’ sound, as in the Nordic and other Germanic languages, e.g. jon, yon, that, ju, you. It combines with other consonants as shown below. The English ‘j’ sound has other spellings. The letter Y is used only as a vowel.
G, in English, has both a ‘g’ and a ‘j’ sound. In the Sjætlan orthography, it is mostly reserved for the ‘g’ sound, with the ’j’ sound shown by other spellings. When followed by a front vowel or J, it is pronounced as a palatal consonant - approximately ‘dy’ - in Whalsay, as often in the Nordic languages.
F, V, S and Z
In Scots and Scottish English, certain vowels are always long before V and Z - e.g. the ‘ee’ sound in grieve, a farm manager, as opposed to grief. This is partly true in Shetlandic, but is not entirely predictable.
V and Z tend to become ‘f’ and ‘s’ in at least some accents, so that in my accent, kliv, hoof, sounds like English cliff, and dyvel, devil, sounds like ‘deefle’, with a short vowel. In these cases, the original V is retained in the spelling, as it is easy for native speakers to pronounce the words with their own accents.
Z and S are further complicated by the fact that, in both English and Scots, a final ‘z’ sound is sometimes spelt as S, e.g. cheese. Since vowel length is always indicated in the Sjætlan orthography, and Z tends to be devoiced to ‘s’ anyway, the Sjætlan orthography mostly uses final S rather than Z, e.g. cjýs, cheese, frýs, freeze, rǿs, praise. Z is, however, used finally where it shows that the preceding vowel is soft, e.g. cjaz (cjäz), jazz, as opposed to pas, pass.
Extra Consonants
Ð ð. In Icelandic, this is the voiced sound of English ‘th’ in ‘the.’ In Shetlandic, this sound occurs in certain districts - mostly in the South of Shetland - where other dialects have a ‘d’ sound. The letter ð is used to show the varying dialect pronunciations of such words. So mǽð, maggot, bǿð, booth, mýð, landmark, pronounced with a voiced ‘th’ sound in some areas, and with a ‘d’ sound in others.
Consonant Combinations and Digraphs
With H
CH - the sound in Scots ‘loch,’ e.g. lǫch, lake, brǫch, pictish fort.
GH - may be used for the sound of English ‘ch’ in cheese in a few proper names, e.g. Ghórgj, George.
HT occurs in words like neiht, night, feht, fight, tǫht, thought. It is pronounced in some accents as a sound similar to ‘ch’ in loch, but produced in the roof of the mouth rather than the throat. In other accents, it is omitted. It often alters the sound of the preceding written vowel.
Words like tǫht - thought - may be pronounced approximately as ‘tocht’, ‘towht’ or ‘towt.’
Words like feht - fight - like ‘fecht’, ‘feyht’ or ‘feyt’.
Words like neiht - night - like ‘nycht’ or ‘nite.’
PH for the ‘f’ sound in words mostly of Greek origin, e.g. phénol.
JH - may be used for the sound of English ‘ch’ in cheese in a few proper names, mostly derived from Latin or Hebrew - Jhǫn, John, Jhésus, Jesus, Jhøn - June (month).
TH - the sound in English ‘thin.’ This occurs mostly at the ends of words, with most English words beginning with ‘th’ traditionally pronounced ‘t’ - as tin, thin, tik, thick, but bǽth, both, ruth, rowlock base in a traditional boat.
With J
Most consonants can have J after them - e.g. bjǫk, retch, fjárm, fawn, ingratiate, hjuk, hook, ljóg, hollow in the landscape, mjáchtles, weak, njág, gnaw, nag, pjág, toil, tjǫch, tough. These are usually pronounced more or less as written, and where they do represent other sounds (such as, technically, a palatal fricative in hjuk or a palatal lateral in ljóg) the pronunciation is automatic, and mostly passes unnoticed in native speech. Other combinations are used to represent more salient pronunciations.
CJ - this is the sound of English ‘ch’ in cheese - cjýs, cheese, witcj, witch, brǫcj, brooch.
DJ - at the beginning of a word, this is pronounced like ‘dy’ or like English ‘j’, depending on dialect, e.g. djuk, duck, as ‘dyook’ or ‘jook’. Elsewhere it has the sound of English ‘j’, e.g. bodjet, budget, modjek, midge, wædj, wedge.
GJ - at the beginning of a word, as ‘gy’ - gjó, cleft in a cliff, gjimer, young ewe. Elsewhere it has the sound of English ‘j’, e.g. pǽgj, page, gǽgj, gauge, though it may be devoiced to a ‘ch’ sound in words like mǽrígj, marriage. In Whalsay, at the beginning of a word, it is usually pronounced approximately ‘dy.’
KJ is pronounced as ‘ky,’ or approximately ‘ty’ in Whalsay - kjoker, to revive, kjób, bribe.
SJ - the sound of English ‘sh’ - sjal, shell, sjalder, oyster-catcher, sjø, she, wasj, wash.
TECHNICAL TERMS
Allophones - different pronunciations of the same underlying vowel, or phoneme, either in different dialects - like Á pronounced 'ah' or 'aw' in different districts - or in different environments, like E pronounced 'eh' in e.g. set, lep, sek, and 'ey' in e.g. bed, hen, neb.
Diacritic - a mark, or ‘accent,’ used to distinguish one letter from another, e.g. the acute diacritic above á in bák, a beam.
Diaphonology - the underlying phonology, or sound system, of a language as a whole.
Digraph - a spelling consisting of two letters, e.g. OO, EE, AI, TH
Diphthong - a two-part vowel sound, e.g. the sounds in English ‘five’ and ‘down.’
Grapheme - a way of spelling a phoneme. So in English, B, T and SH are graphemes, each representing a particular sound.
Homographs - words which are written the same but are different words, e.g. English row (line of things) and row (move a boat by means of oars). They may also have different sounds, like row (line, row a boat) and row (a noisy argument), or not, like ruler (measuring device or potentate).
Homophones - words which sound the same but have different meanings, e.g. English cite and sight.
Morphology - the form of words. Usually applied to word endings such as plural -S.
Norn - the Nordic language spoken in Shetland from the Viking era, until it was replaced by Shetlandic, probably in the late 1700s.
Orthography - a conventional spelling system for a language.
Phoneme - see phonology below
Phonetics - the exact pronunciation of underlying phonemes. This cannot be represented accurately in writing except by phonetic script as used by linguists.
Phonology - the sounds which can carry meaning. For example, the sound difference between fat, fat, and fát, fault, is a phonological difference - it distinguishes one word from another. The difference between a Central Mainland and North Isles pronunciation of fát as ‘faat’ or ‘faut’ is merely a phonetic difference - it is the same word pronounced with a different accent. The underlying sounds of a language are called ‘phonemes.’ It is these sounds that are normally represented in the spelling of a language.
RP - Received Pronunciation or ‘BBC’ English, the standard spoken English of England.
SSE - Scottish Standard English, as spoken - with certain local variations - by Shetlanders speaking English.
Scots - the traditional language of Lowland Scotland, ostensibly recognised as a language, but in practice treated as a continuum of dialects with no official status or written form.