The classical older fuþark as used in Scandinavia is well understood, and its 24 characters has a simple relationship* to the 22 phonemes of the (reconstructed) classical Proto-Norse language. The order of the runes is stable, with the exception of two consecutive pairs that vary in their internal order: 13/14 and 23/24. There is relatively little graphical variation; the second column gives the typical shapes.
*) The exact phonetic value of ʀ has been debated; it has its origin in Germanic /z/ and at the end of the Viking age it merges with /r/, the symbols chosen for the transliteration and the phoneme here is not intended as a strong claim to its pronunciation. The value of rune 13 is uncertain, but must have been allophonic like rune 22; the assumption sometimes made that it represents Proto-Germanic /æ/ is entirely specious. The three other marked allophones do not have signs of their own. Note that length is phonemically distinctive in both vowels and consonants throughout the period, but is never indicated in the runic orthography.
| Older fuþark | Late Scandinavian older fuþark | Earliest younger fuþark | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | ᚠ | f | /f/ | ᚠ | f | /f/ | ᚠ | f | /f/ | [v] | ||||
| 2 | ᚢ | u | /u/ | ᚢ | u | /u/ | /ũ/ /y/ /ỹ/ | ᚢ | u | /u/ | [w] | /ũ/ /y/ /ỹ/ /o/ /õ/ /ø/ /ø̃/ | ||
| 3 | ᚦ | þ | /θ/ | ᚦ | þ | /θ/ | ᚦ | þ | /θ/ | [ð] | ||||
| 4 | ᚨ | a | /a/ | ᚨ | a | /ã/ | /æ̃/ | ᚨ | ą | /ã/ | /æ̃/ /ɔ̃/ | |||
| 5 | ᚱ | r | /r/ | ᚱ | r | /r/ | ᚱ | r | /r/ | |||||
| 6 | ᚲ | k | /k/ | ᚴ | k | /k/ | ᚴ | k | /k/ | /g/ [ɣ] | ||||
| 7 | ᚷ | g | /ɣ/ | [g] | ᚷ | g | /ɣ/ | [g] | ||||||
| 8 | ᚹ | w | /w/ | ᚹ | w | /w/ | ||||||||
| 9 | ᚺ | h | /x/ | ᚺ | h | /h/ | ᚺ | h | /h/ | |||||
| 10 | ᚾ | n | /n/ | ᚾ | n | /n/ | [ŋ] | ᚾ | n | /n/ | [ŋ] | |||
| 11 | ᛁ | i | /i/ | ᛁ | i | /i/ | [j] | /ĩ/ | ᛁ | i | /i/ | [j] | /ĩ/ /e/ /ẽ/ | |
| 12 | ᛃ | j | /j/ | ᚼ | ᴀ | /a/ | /æ/ | ᚼ | a | /a/ | /æ/ /ɔ/ | |||
| 13 | ᛇ | ï | – | ? | ||||||||||
| 14 | ᛈ | p | /p/ | |||||||||||
| 15 | ᛉ | ʀ | /ɹ/ | ᛦ | ʀ | /ɹ/ | ᛦ | ʀ | /ɹ/ | |||||
| 16 | ᛊ | s | /s/ | ᛋ | s | /s/ | ᛋ | s | /s/ | |||||
| 17 | ᛏ | t | /t/ | ᛏ | t | /t/ | ᛏ | t | /t/ | /d/ | ||||
| 18 | ᛒ | b | /β/ | [b] | ᛒ | b | /β/ | [b] | /p/ | ᛒ | b | /b/ | [β] | /p/ |
| 19 | ᛖ | e | /e/ | ᛖ | e | /e/ | /ẽ/ | |||||||
| 20 | ᛗ | m | /m/ | ᛗ | m | /m/ | ᛗ | m | /m/ | |||||
| 21 | ᛚ | l | /l/ | ᛚ | l | /l/ | ᛚ | l | /l/ | |||||
| 22 | ᛜ | ŋ | – | [ŋ] | ||||||||||
| 23 | ᛞ | d | /ð/ | [d] | ᛞ | d | /ð/ | [d] | ||||||
| 24 | ᛟ | o | /o/ | ᛟ | o | /o/ | /õ/ /ø/ /ø̃/ | |||||||
As the language in Scandinavia goes through the changes separating the early Old Norse of the Viking age from Proto-Norse, several new allophones develop, and to a large degree becomes phonemised. A few rune-shapes change, primarily giving full line height to forms that was previously shorter. The Anglo-Saxon runes are created at a point when several of these changes have occurred, and have resulted in a situation where some runes cover more than one phoneme. While this trend only increases in Scandinavia, the creator of the Anglo-Saxon runes reverts the situation by introducing four new runes.
The second block of the table shows the situation when fronting has created four new vowel phonemes, and nasalisation has doubled the resulting nine vowel phonemes to eighteen. Phonemic opposition between the semivowels /j/ and /w/ and the vowels /i/ and /u/ gradually disappears. The first of these is reanalysed as an allophone of /i/, freeing its rune to be used for two of the four phonemes deriving from earlier /a/. The conventional transliteration reserves a for rune 4 which is now only used for the nasal vowels, introducing ᴀ for the rune 12, now used for the oral counterparts of these. Some inscriptions show similar tendency of using rune 8 for unfronted /u/ in addition to the semivowel, reserving rune 2 for /y/ (both regardless of nasality).
Positional voicing and devoicing of consonants in the same period does not create new phonemes, only a new allophone of /f/ and some blurring of the lines between pairs of voiced and unvoiced phonemes. Dropping a separate rune for the very rare phoneme /p/ in favour of using the /b/-rune might be due to this development in combination with the precedent for using the same rune for more than one phoneme established by the explosive growth of vowel phonemes.
In the table, a primary phoneme is given for each rune, followed first by any allophones of this, and finally all other phonemes covered by the same rune. The primary phoneme is the one featuing in the name of the rune, and that this was indeed considered the primary value is shown clearly both by the distribuiton of sound values between the old /a/- and /j/-runes as well as late developments within the younger runes.
This gradual development sets the stage for the final leap of a development where a writing system with close to a one-to-one correspondence between phonemes and graphemes against all logic reacts to a 50% increase in the number of phonemes by a 33% reduction in the number of graphemes. Three of the changes in the third block of the table could perhaps be ascribed to the terminal part of the preceeding phase: the reanalysis of [w] as an allophone of /u/ with the resulting loss of rune 8, and the phonemisation of /ɔ/ and /ɔ̃/.
The actual transition consists of the loss of two pairs of runes, with the transfer of their values to other runes, some of which already had more than one value. All of these changes seem to have occurred simultaneously, as a deliberate revision that quickly spread across all of Scandinavia.
The easiest to see a motivation for was the loss of a dedicated /d/-rune. This had for a long time had two allophones, [d] and [ð], with complementary distributions. The linguistic development with positional voicing and devoicing meant that [d] now alternated with [t] in the inflectional paradigm of many words, and that [ð] also became a positional allophone of /θ/. This could have been handled simply by accepting this alternation as part of the inflectional system; but the rapidity of these changes meant that regional and intergenerational differences were rife. Taking the earlier transfer of the phoneme /p/ to the rune with the primary value /b/ as a template, the Gordian knot was cut by declaring voicedness to be disregarded in writing, and distributing the plosive allophone [d] to a secondary but still phonemic value of t and the fricative [ð] to þ where it became allophonic as /θ/ had aquired voicing in the same positions as /d/ had been fricative.
This decision carried with it the implication that at least the plosive allophone of /g/ now would have to be written with the same rune as /k/. As earlier /x/ had developed into /h/ and become lost in positions where it otherwise would have become voiced, there was no other rune to express the allophone [ɣ], so this had to be written k as well. The third voiced plosive with a fricative allophone /b/ could of course not be written p since that rune had already been discarded if favour of b itself, and its fricative allophone [β] was at this time distinct from the newly developed voiced allophone [v] of /f/ by the former being bilabial and the latter labiodental.
The other pair of runes to be lost represented mid (or close-mid) vowels, their values being taken over by the runes for the corresponding close vowels. Previously these had contrasted in all positions, but by this time, they were merging in unstressed positions. As with voicedness in consonants as mentioned above, regional and intergenerational differences as mentioned above created great confusion in how to write these vowels, despite the clear distinction between them in stressed positions. The custom of using one rune for more than one phoneme was now so well established that this was seen as acceptable, despite one rune ending up with eight phonemic values (twelve when counting length).
Some minor details have been glossed over in the table. Classical Old Norse /y/ most frequently derive from fronting of /u/, but in some cases from rounding of /i/. These two variants had not yet merged at this point, so only the former is denoted /y/ here, and was normally written u. The latter was first an allophone of /i/, and to the degree it was phonemised before the merger with the former, it became a secondary phoneme of i.
Already at a quite early date, there appeared some exceptions to the system. Among the most important is the sporadic use of digraphs for some secondary phonemes: ai or ia for /æ/ or /e/, au for /ɔ/, /o/ or /ø/ (both with ą for a when nasal) and iu for /y/ (possibly limited to the variant arising from rounding of /i/). As this collided with the spellings of diphtongs, it was no big improvement and never became widely used.
Tor Gjerde