Old Norse phonology

Vowels
The vowel phonemes mostly come in pairs of long and short. The standardized orthography marks the long vowels with an acute accent. In medieval manuscripts, it is often unmarked but sometimes marked with an accent or through gemination. All phonemes have, more or less, the expected phonetic realization.

Old Norse had nasalized versions of all nine vowel places. These occurred as allophones of the vowels before nasal consonants and in places where a nasal had followed it in an older form of the word, before it was absorbed into a neighboring sound. If the nasal was absorbed by a stressed vowel, it would also lengthen the vowel. These nasalizations also occurred in the other Germanic languages, but were not retained long. They were noted in the First Grammatical Treatise, and otherwise might have remained unknown. The First Grammarian marked these with a dot above the letter. This notation did not catch on, and would soon be obsolete. Nasal and oral vowels probably merged around the 11th century in most of Old East Norse. :3 However, the distinction still holds in Dalecarlian dialects. :4 The dots in the following vowel table separate the oral from nasal phonemes.

Note: The low/low-mid vowels may be indicated differently:



Sometime around the 13th century, Ǫ merged with Ø or O in all dialects except Old Danish. In Icelandic, all Ǫ merged with Ø. This can be determined by their distinction within the 12th-century First Grammatical Treatise but not within the early 13th-century Prose Edda. The nasals, also noted in the First Grammatical Treatise, are assumed to have been lost by this time. See Old Icelandic for the Œ > Æ and Ę > E mergers.

Consonants
Old Norse has six plosive phonemes. Of these is rare word-initially and  and  are realized as voiced fricative allophones  between vowels, except in compound words (e.g. veðrabati), already in the Proto-Germanic language (e.g. *b  > [v] between vowels). The phoneme is realized as  after an n or another g and as  before  and. It is realized as a voiced velar fricative, by some accounts inside words, and by others between vowels (and otherwise as ). The Old East Norse /ʀ/ was an apical consonant whose position isn't precisely known, being reconstructed as a palatal sibilant :2. It descended from Proto-Germanic and eventually developed into, as it already had done in Old West Norse.

The consonant digraphs hl, hr, hn occurred word-initially. It is unclear whether they were sequences of two consonants (with the first element realised as or perhaps ), or as single voiceless sonorants,  and  respectively. In Old Norwegian, Old Danish and later Old Swedish the groups hl, hr, hn were reduced to plain l, r, n, suggesting that they were most likely realised as voiceless sonorants by Old Norse times.

The pronunciation of hv is unclear, and may have been (the Proto-Germanic pronunciation),  or. Unlike the other three groups above, it was retained much longer in all dialects, and never developed into a voiceless sonorant in Icelandic, but instead "hardened" to a plosive. This suggests that it was not a voiceless sonorant, but retained stronger frication.

Orthography
Unlike Proto-Norse, which was written with the Elder Futhark, runic Old Norse was originally written with the Younger Futhark, which only had 16 letters. Because of the limited number of runes, the rune for the vowel u was also used for the vowels o, ø and y, and the rune for i was used for e. Medieval runes came into use some time later.

As for the Latin alphabet, there was no standardized orthography in use in the Middle Ages. A modified version of the letter wynn called vend was used briefly for the sounds, , and. Long vowels were sometimes marked with acutes, but also sometimes left unmarked or geminated. The standardized Old Norse spelling was created in the 19th century, and is for the most part phonemic. The most notable deviation is that the non-phonemic difference between the voiced and the voiceless dental fricative is marked — the oldest texts as well as runic inscriptions use þ exclusively. Long vowels are denoted with acutes. Most other letters are written with the same glyph as the IPA phoneme, except as shown in the table below.

Accent
Primary stress in Old Norse falls on the word stem, so that hyrjar would be pronounced. In compound words, secondary stress falls on the second stem (e.g. lærisveinn, ). :1

Modern Swedish, Danish, and Norwegian have two registers reflected in differing pronunciation of the stressed syllable of words. In Swedish and Norwegian, the registers are reflected in different tones (i.e. through tonal word accent), while in Danish the difference is the presence or absence of stød, a glottal gesture considered a kind of creaky voice that seems to have been documented by Swedish sources as early as the 16th century. What is here called class 1 is reflected as tone 1 in Norwegian, acute accent in Swedish, and presence of stød in Danish, whereas class 2 words have tone 2 in Norwegian, grave accent in Swedish, and no stød in Danish. No sign of any tonal system is found in Icelandic or Faroese.

Not all cognates occur in the same register classes in all three languages, partly due to language-specific restrictions on the contexts in which the two classes can occur. For example, stød can only occur in stressed words that have long vowels and end in a voiced consonant, whereas in Swedish and Norwegian, monosyllables can only take tone 1/acute accent. In general, however, class 1 words are those that are monosyllabic in Old Norse, while class 2 words are those that are polysyllabic. Exceptions, including minimal pairs, have arisen for various reasons:
 * 1) Suffixing a definite article onto class 1 words does not change their class, suggesting that the register distinctions developed before the definite article fused with the noun. Hence Old Norse and-inn "the duck" > Swedish class 1 anden, but andi-nn "the spirit" > class 2 anden.
 * 2) The Old Norse non-syllabic morpheme -r after a consonant later became syllabic through the epenthesis of a vowel, without changing the class.  Hence Old Norse bœndr "farmers" > Norwegian class 1 bønder, but baunir "beans" and bœnir "prayers" both > Norwegian class 2 bønner (homophonous with bønder except for the pitch difference).
 * 3) Many later polysyllabic loanwords in the various languages have acquired class 1.