ISO/IEC JTC1/SC22/WG20 N809R 2001-01-09 Internationalization International Organization for Standardization Organisation internationale de normalisation Œеждународнаяорганизацияпостандартизации

Doc Type: Working Group Document Title: Ordering the script Source: Michael Everson Status: Expert Contribution Date: 2001-01-09

On 2000-12-24 Olle Järnefors published on behalf of the ISORUNES Project in Sweden a proposal for ordering the in the Common Tailorable Template (CTT) of ISO/IEC 14651. In my view this ordering is unsuitable for the CTT for a number of reasons.

Runic ordering in ISO/IEC 10646. The Runes are encoded at U+16A0–U+16FF, in a unified set of characters encompassing the four major traditions of Runic use: Germanic, Anglo-Frisian, Danish, and Swedish-Norwegian, and Medieval. The Runes are arranged in the code table agreed by ISO/IEC JTC1/SC2/WG2 in an order based on the the traditional positions of the Runes in abecedaries, namely, the fuþark order. This order is known from hundreds of primary sources which list the Runes in sequence, often with no other text. Most scholarly texts refer to the fuþark in one way or another. Nearly all secondary texts, whether popular introductions to the Runes or New-Age esoterica, give primacy to the traditional fuþark sequence.

Runic names in ISO/IEC 10646. The names given to the Runes in the UCS may be a bit clumsy, but they are intended to serve the needs of scholars and amateurs alike; not everyone is familiar with Runic practices, and not everyone is conversant with the traditional names in Germanic, English, and Scandinavian usage. So the names concatenate those three together with the scholarly transliteration letter.

Runic and . It should be noted that the ISORUNES position has always been to treat the Runes as though they were just glyph variants of the Latin script: their original proposals gave the Runes sequenced in the code table according to their Latin , and gave only the Latin transliteration equivalents as the character names. SC2/WG2 had not treated any other historical script in this way, and chose not to set a precedent with the Runes. It is suggested here that SC22/WG20 do likewise.

Fuþark ordering. On page 4 below, I print first the binary order of the Runes in the UCS (omitting the three punctuation characters), and a proposed fuþark ordering for the CTT. On page 5, the structure of this ordering is given showing primary and secondary weighting, and showing how the four traditions are reflected in their traditional order. Primacy is given to the Germanic Runes, with secondary unifications where applicable in the order Anglo-Frisian, Danish, Swedish-Norwegian, and Medieval. The Calendar Runes are decomposed into their source ligatures for ordering. The only point I consider contentious here is the possible ordering of RUNIC LETTER Q as a variant of RUNIC LETTER WUNJO WYNN (despite the character’ origin in LATIN SMALL LETTER Q) and RUNIC LETTER X as a variant of RUNIC LETTER SIGEL LONG-BRANCH-SOL S (which it clearly is).

OrderingtheRunicscript/MichaelEverson Page1 Latin ordering. On page 6 below, I print first the ISORUNES proposal of 2000-12-24 (which follows the same conventions for primary and secondary weighting by tradition but which is based on the order of the characters according to their Latin transliteration), and a revision of this based on the actual order of Latin characters in the CTT. On page 7, the structure of this ordering is given showing primary and secondary weighting along with the Latin transliteration characters themselves. It is important to note that for at least some of the Runes, more than one transliteration is possible. While RUNIC LETTER CWEORTH is transliterated here, it is very often expanded to the digraph in English contexts. Another example is RUNIC LETTER ALGIZ EOLHX, which can be transliterated , , or .

Note that a universally-acceptable ordering according to Latin transliteration is not really feasible. Consider the Latin tailoring required to ensure that æ, ø, and þ are ordered as used in English, Danish, Swedish, Norwegian, and Icelandic contexts.

Sample orderings. First, I give here a Runic song printed by Ole Worm in Danica Litteratura Antiquissima in 1636 from an original Norwegian manuscript subsequently lost; text normalized into Icelandic spelling by . . Gordon. Note the fuþark order in vertical presentation (instead of the usual linear abecedary).

(fé) veldr frænda rógi; fœðisk úlfr í skógi. ¢ (úr) af illu járni; opt hleypr hreinn á hjarni. ¦ (þurs) veldr kvenna kvillu; kátr verðr fár af illu. ­ (óss) es flestra ferða fǫr, en skálpr er sverða. ± (reið) kveða hrossum versta; Reginn sló sverðit bezta. ´ (kaun) es beygja barna; bǫl gørir mann fǫlvan. ¼ (hagall) es kaldastr korna; Kristr skóp heim inn forna. ¾ (nauð) gørir hneppa kosti; nøktan kelr í frosti. Á (ís) kǫllum brú breiða; blindan þarf at leiða. Å (ár) es gumna góði; getk at ǫrr vas Fróði. Ë (sól) es landa ljómi; lútik helgum dómi. Ï (Týr) es einhendr Ása; opt verðr smiðr at blása. Ò (bjarkan)’s laufgrœnstr líma; Loki bar flærðar tíma. Ø (maðr) es moldar auki; mikil es greip á hauki. Ú (lǫgr)’s, es fellr ór fjalli, foss; en gull eru hnossir. ã (ýr) es vetrgrœnstr viða; vant’s, es brennr, at svíða.

On pages 8 and 9 I present a list of Runic letternames ordered according to the two approaches. The fuþark ordering is, in my view, superior because it unifies dotted Runes with other base characters, which in fact gives a more accurate result. From the final 1996-03-13 ISORUNES proposal to WG2 (N1330):

The dotted runes do not appear to have been counted as independent runes. They only seldom appear in inscriptions representing the fuþark, and then as a rule just a few of them appear at the end.

Note that using Latin transliteration order, a name like Þórgunn, which could be written ¦¢±´¢¾ þurkun or ¦¢±µ¢¾ þurgun, would be found in different parts of an ordered list. This could be finessed by unifying the dotted consonants with their base form, but then this rather sabotages the intended principle of following Latin transliteration.

OrderingtheRunicscript/MichaelEverson Page2 It seems clear to me that following the traditional fuþark ordering for the CTT, WG20 would be applying the same principles to Runic as are applied to other scripts. If for scholarly purposes a glossary of Runic terms were required, and Latin alphabetic order preferred, it would be far simpler to do as I have done on page 7, namely to begin each line in the glossary with the Latin transliteration itself, in which case the local Latin- tailored tables could apply. This reduces the burden on scholars, because they will be using English, Danish, Swedish, Norwegian, or Icelandic systems software which will order Latin as they prefer – and which will be provided by their software vendors as a matter of course. Requiring scholars in different countries to tailor Runic as well for local requirements (which the major software vendors are unlikely to do) will add unnecessary expense to their endeavours.

OrderingtheRunicscript/MichaelEverson Page3 Binaryorder. Fuþarkorder.Thisisshowninexpanded formonthefollowingpage. µ ~ ß ¶ Õ Û ¡ ¶ Ë à ¡ · É Ü ¢ · Ì á ¢ ¹ ~ Ý £ ¸ Í â ¤ é Ë Þ ¤ ¹ Î ã ¥ º ê ß ¥ º Ï ä ¦ » Ì ª ¦ » Ð å § ¼ Í «

§ ¼ Ñ æ ð (¦¦) ½ Î £ ¨ ½ Ò ç ¨ ¾ Ï à © ¾ Ó è © ¿ Ð ã ª ¿ Ô é ¬ À Ñ ¸ « À Õ ê ­ Á Ò ä ¬ Á Ö î ® Â Ó á ­ Â × ï ¯ Ã Ô â ® Ã Ø ð ° Ä Ö å ¯ Ä Ù ± Å × æ ° Å Ú ² Æ Ø ç

± Æ Û ³ î (ÆÚ) Ù è

² Ç Ü ´ Ç ï (ØØ) ³ È Ý µ È Ú ´ É Þ

OrderingtheRunicscript/MichaelEverson Page4 Fuþarkorder.Conventionsusedhere:Eachlanguagegroup'slettersareordered traditionallywithasuperscriptnumbershowingtheranking;inthefollowinghierarchy: =Germanic,E=Anglo-Frisian,=Danish,S=Swedish-Norwegian,I=Icelandic,= Medieval.DottedRunesandothervariantsfollownext,markedhereonlybysuperscript D andV.TheCalendarRuneshavebeenorderedasligatures.

G1 E1 D1 S1 ¶ MD Õ MV Û MD

¡ MD · G7 E7 É G15 E15 Ü G22

¢ G2 E2 D2 S2 ¹ G8 E8 ~ G16 Ý E22

¤ MD é M Ë E16 D11 Þ G23E23

¥ MV º G9 ê M ß G24 E24

¦ G3 E3 D3 S3 » E9 Ì S11 ª E25

§ MD ¼ D7 Í MD « E26

ð (¦¦) ½ S7 Î MV £ E27

¨ G4 ¾ G10 E10 D8 Ï G17 E17 D12 à E28

© E4 ¿ S8 Ð S12 ã E29

¬ D4 À MD Ñ MD ¸ E30

­ S4 Á G11 E11 D9 S9 Ò G18 E18 D13 ä E31

® MV Â MD Ó S13 á E32

¯ M Ã G12 Ô MD â E33

° M Ä E12 Ö G19 E19 å E34

± G5 E5 D5 S5 Å D10 × G20 E20 æ D16

² G6 Æ S10 Ø D14 ç S16

³ E6 î (ÆÚ) Ù S14 è I

´ D6 S6 Ç G13 E13 ï (ØØ)

µ MD È G14 E14 Ú G21 E21 D15 S15

OrderingtheRunicscript/MichaelEverson Page5 ISORUNESorderasspecified2000-12-24 Latinorder.Thisisshowninexpanded formonthefollowingpage.

¨ a1 » Ü ŋ24 Ï t36 ¨ a1 µ g {} ï (mm) ~ s25

ª a ¼ h Ý ŋ Ð s ª a ¸ ǧ ¾ n17 Ë s

Å a  h ¶ ŋ ¦ þ37 Å a º h10 ¿ Ì s

Æ a Á i13 ß {œ}25 ¢ u38 Æ a » h À N18 å(st)

« æ2 Ç ï{˜}14 © o26 ¡ v39 ¬ ą{o} ¼ h Ü ŋ19 Ï t26

Ò b3 á (io)15 ¬ o{ą}27 ¹ w40 ­ ą{o} Â h Ý ŋ Ð

Ó Ã j16 ­ o{ą} ¥ w î (al) Á i11 ¶ ŋ ¢ u27

³ c4 Ä ® o28 ê x41 « æ2 Ç ï{˜} ß o{œ}20 ¡ v28

Í ² k17 È p29 £ y42 Ò b3 á (io) © o ¹ w29

Þ d5 ã Ô p ¤ y Ó b à j12 ® o ¥ w

Ñ d ´ k Õ p è y ³ c4 Ä j ° ǫ ê x30

6 18 13 31 § ð ä ǩ â q{cw}30 É {,x}43 Í c ² k ¯ ø £ y

Ö e7 Ú l19 é q31 Î z44 Þ d5 ã k È p21 ¤ y

½ e Û L20 ± r32 ¯ ø45 Ñ d ´ k Ô p è y

8 21 33 46 6 à (ea) × m æ R ° ǫ § ð ä ǩ Õ p É z{R,x}32

9 47 7 14 f Ø m ç R î 17 Ö e Ú â q{cw}22 Î z

· g10 Ù m ~ s34 ï 1848 ½ e Û L15 é q ¦ þ33

µ g ¾ n22 Ë s ð 1949 à (ea) × m16 ± r23 ð (þþ)

¸ g {}11 ¿ n Ì s f8 Ø m æ R24

º h12 À N23 å (st)35 · g9 Ù m ç R

OrderingtheRunicscript/MichaelEverson Page6 Latinorder,basedonusualtransliterationsandfollowingthesameorderasgivenfor LatinintheDefaultTailorableTemplate.Notethatthemainproblemwithtranslitera- tionsisthatmanycharactershavemorethanonetransliterationaccordingtodifferent traditions.

¨ a µ g{} ï (mm) ~ s

ª a ¸ ǧ ¾ n Ë s

Å a º h ¿ n Ì s

Æ a » h À N å (st)

¬ ą{o} ¼ h Ü ŋ Ï t

­ ą{o} Â h Ý ŋ Ð t

î (al) Á i ¶ ŋ ¢ u

« æ Ç ï{˜} ß o{œ} ¡ v

Ò b á (io) © o ¹ w

Ó b à j ® o ¥ w

³ c Ä j ° ǫ ê x

Í c ² k ¯ ø £ y

Þ d ã k È p ¤ y

Ñ d ´ k Ô p è y

§ ð ä ǩ Õ p É z{R,x}

Ö e Ú l â q{cw} Î z

½ e Û L é q ¦ þ

à (ea) × m ± r ð (þþ)

f Ø m æ R

· g Ù m ç R

OrderingtheRunicscript/MichaelEverson Page7 Runicnamesinfuþarkorder.

Ö fe ÄÖ± ger Þ«¸ dæǧ Ö©» feoh ÃÖ±¨¾ jeran ߦ¨Ú¨¾ oþalan Öº¢ fehu ű ar ߦÖÚ oþel ¢± ur Ʊ ar ª³ ac ¢±¢É uruz ǹ¨É ïwaz «Ë³ æsc ¦¢±Á~¨É þurisaz ÈÖ©±¦ peorþ £± yr ¦¢±Ë þurs ÈÖ±¦ß perþo à± ear ¦¢±Ì þurs ËÅÚ sol ãªÚ³ kalc ¦©±¾ þorn ÌÆÚ sol ¸ª± ǧar ¨¾~¢É ansuz ËÁ¸ÖÚ siǧel äàÚ³ ǩealc ¨Ú·ÁÉ algiz ~ß¹ÁÚß sowilo á± ior ±¨ÁÞß raido ÏÁ± tir âÖ©±¦ qeorþ ±ªÞ rad ÐÁ± tir 媾 stan ±ÖÁÞ reid ÏÁ±¨É tiraz æ± yR ´Å¢¾ kaun Ï£± tyr ç± yR ´Æ¢¿ kaun ÒÁű´Å¾ biarkan è± yR ²¨¢¾¨ kauna ÓÁƱ´Æ¿ biarkan ³Ö¾ cen ÒÖ©±³ beorc ·ÖÒß gebo ÒÖ±²¨¾¨¾ berkanan ·£ ¢ gyfu Ö©» eoh ¹¢¾Ãß wunjo Ö©ÚÉ eolx ¹£¾¾ wynn Ö» eh º¨·Ú¨É haglaz Öº¹¨É ehwaz ¼ÅµÅÚÚ hagall ת¾ man ½ÆµÆÚÚ hagall ר¾¾¨É mannaz »«¸Ú hæǧl ØÅÞ± madr ¾Å¢Þ naud ÙÆÞ± madr ¿Æ¢Þ naud Ú¨¢²¨É laukaz ¾¨¢ÞÁÉ naudiz Úŵ± lagr ¾£Þ nyd ÚƵ± lagr ÁË is Úª·¢ lagu Á~¨É isaz ÁÝ iŋ ÁËË iss Áܹ¨É iŋwaz ÁÌÌ iss Þ¨·¨É dagaz OrderingtheRunicscript/MichaelEverson Page8 TransliteratedRunicnamesinLatinorder. ac ª³ iss ÁËË tir ÏÁ± algiz ¨Ú·ÁÉ iss ÁÌÌ tir ÐÁ± ansuz ¨¾~¢É ïwaz ǹ¨É tiraz ÏÁ±¨É ar ű jeran ÃÖ±¨¾ tyr Ï£± ar Ʊ kalc ãªÚ³ ur ¢± æsc «Ë³ kaun ´Å¢¾ uruz ¢±¢É beorc ÒÖ©±³ kaun ´Æ¢¿ wunjo ¹¢¾Ãß berkanan ÒÖ±²¨¾¨¾ kauna ²¨¢¾¨ wynn ¹£¾¾ biarkan ÒÁű´Å¾ ǩealc äàÚ³ yr £± biarkan ÓÁƱ´Æ¿ lagr Úŵ± yR æ± cen ³Ö¾ lagr ÚƵ± yR ç± dagaz Þ¨·¨É lagu Úª·¢ yR è± dæǧ Þ«¸ laukaz Ú¨¢²¨É þorn ¦©±¾ ear à± madr ØÅÞ± þurisaz ¦¢±Á~¨É eh Ö» madr ÙÆÞ± þurs ¦¢±Ë ehwaz Öº¹¨É man ת¾ þurs ¦¢±Ì eoh Ö©» mannaz ר¾¾¨É eolx Ö©ÚÉ naud ¾Å¢Þ fe Ö naud ¿Æ¢Þ fehu Öº¢ naudiz ¾¨¢ÞÁÉ feoh Ö©» nyd ¾£Þ ǧar ¸ª± oþalan ߦ¨Ú¨¾ gebo ·ÖÒß oþel ߦÖÚ ger ÄÖ± peorþ ÈÖ©±¦ gyfu ·£ ¢ perþo ÈÖ±¦ß hagall ¼ÅµÅÚÚ qeorþ âÖ©±¦ hagall ½ÆµÆÚÚ rad ±ªÞ haglaz º¨·Ú¨É raido ±¨ÁÞß hæǧl »«¸Ú reid ±ÖÁÞ iŋ ÁÝ siǧel ËÁ¸ÖÚ iŋwaz Áܹ¨É sol ËÅÚ ior á± sol ÌÆÚ is ÁË sowilo ~ß¹ÁÚß isaz Á~¨É stan 媾 OrderingtheRunicscript/MichaelEverson Page9