Þrjótrunn: A North Romance Language


Umlaut & Breaking

In an earlier stage of the language, short -u-, -a-, and -i- and the glides -j- and -w-, and also -r in unstressed syllables sometimes changed the vowel in preceding syllables. The rules for these processes are complicated, because this happened in several steps, and interacted with other sound changes in the history of the language. Moreover, from the modern words, the triggering phones have often disappeared, so umlaut and breaking are considered irregularities today.

There are two types of umlauts in Þrjótrunn: i- and u-umlaut, and two types of breaking: a- and u-breaking (u-breaking could be considered a combination of a-breaking + u-umlaut). The following tables show the effect on the stem vowel, if umlaut occurs.

If the umlaut mutation is marked with a superscript '*' in an ending, then the second syllable of a disyllabic stem triggers the umlaut here. Depending on whether it is 'i', 'a' or 'u', umlaut and/or breaking occurs.


original   umlauted
a ö in first syllable
a u in other syllables
ai ey
vo ó
e ö
é æ
i y
í ý

By far the most common u-umlaut is a > ö.

Endings show this type of umlaut with a superscript 'u', which also shows that u-breaking takes place.

In disyllabic stems, u-umlaut propagates through an intermediate syllable with the vowel -a-, which changes to -u-. E.g. kafall- + [u]r = köfull.


original   umlauted
a e
á æ
au ey
e i
é í
o ö
ó æ
ö y
u y
ú ý
ve ý

Endings show this type of umlaut with a superscript small i or capital I. The capital I means the umlaut does not affect 'e' or 'é'.


When the stem contains -e-, this becomes -ja- (plain or a-breaking), or -jö- (together with u-umlaut, also called u-breaking).

Endings show this type of umlaut with a superscript a or u.


When a consonantal ending is added to a consonantal stem, or when a vowel drops, the then adjacent consonants may combine in non-obvious, but usually regular ways. This section gives an overview of some regular consonant sandhi in Þrjótrunn. If an ending is marked with #, it means the effect only happens at ends of words.

ð + ð dd
ð + t tt
f + ll fl
f + rr fr
f + v f
ft + ð ft
g + t kt
k + # g not always; if so, marked with -k/g
k(j) + ð kt
k + j kkj
k + v kkv
l + ð ld
ll + ð llt
ll + r# ll
ld + r llr
lt + ð lt
m + n fn in some words
m + ð md
n + ð nd
n + r# nn
n + t# tt
n + m ðm
nd + ð nt
+ r nr
ng(j) + ð ngd
lg(j) + ð lgd
nn + r ðr
nn + ð nnd
p + ð pt
+ ð rt
rr + ð rrd
rr + n rn
s + r ss
s + ð st
ss + ð sst
st + ð st
t + # ð not always; if so, marked with -þ
t + ð tt

Apart from these, triples will be shortened to pairs, i.e. for any X, there's an implicit rule that XX + X > XX. And whenever there is a sandhi rule X + Y > XX, this also means XX + Y > XX, unless listed otherwise. E.g. tt + ð > tt.

If endings start with -(u), an -u- is inserted for epenthesis only. This means the above sandhi rules usually also apply. If the phoneme after this -(u) is lost by the sandhi, the -u- is not needed anymore. E.g. n+(u)r# > nn, instead of *-nur or something else.

Further note that if stems end in a consonant + r, an epenthetic -u- is inserted before -r if no ending is added.

When two endings are added to a word causing the word to drop an epenthetic vowel, then the first ending will not drop its vowel when the second ending is added.

Typical changes also involve stems in v-/j-. These consonants are lost at ends of words and before consonant. Further, v- is lost before -u and j- is lost before -i. Sometimes, endings is -(u)r are -ir for j-stems.


Prepositional prefixes (in-, svb-, ē-, ab-, dē-, ...), which occur frequently in Þrjótrunn, generally do not assimilate to the stem they attach to, in contrast to what happened in Latin. E.g. 'rigir'. It should be noted, however, that the stems often have different forms from what would be used without the prefix, and that often they do not occur at all without a prefix.

October 28th, 2007
Comments? Suggestions? Corrections? You can drop me a line.