User:Kolmiel/Luxembourgish

This page is about the historical phonology of Luxembourgish.

It aims at giving an overview of the native phonetic developments that Luxembourgish has undergone from Old High German to the contemporary standard language (“Luxembourgish koine”).

Preliminary remarks
Every standardized language contains forms that are etymologically exceptional, since there will be dialectal variation and individual words will be standardized in forms differing from the bulk of the vocabulary. Luxembourgish contains an above-average number of such words because this language, which was a random dialect 150 years ago, has been heavily influenced by Modern Standard German as well as bordering varieties of German. Standard German influence is, of course, due to schools, books, and newspapers. Outside dialectal influence has been strong, too. It must be remembered that Luxembourg was long just another territory of the Holy Roman Empire, and that Luxembourg City used not to be a major cultural centre.

This should not be misunderstood to mean that the phonetic shapes of Luxembourgish words are widely unpredictable. But it does mean that exceptions—even several of them—need not invalidate a particular rule. Every rule given is liable to exceptions, but some are so more than others. This depends chiefly on whether a given development is shared by most of West Central German or not. I mention or discuss exceptions only sporadically.

That having been said, since part of this page is based on original research, individual rules may need to be corrected or made more precise. Please propose improvements wherever you can!

Central Franconian consonantism
Lux. developed from Old High German, but not from the kind that most of the surviving texts are in. Rather, it is based on the Central Franconian dialect, which did not undergo the following parts of the High German consonant shift:
 * Proto-Germanic -β-, -ɣ- → OHG -b-, -g- ;
 * PG d → OHG t;
 * PG p-, -pp-, -p- after -m- → OHG ph ;
 * PG k-, -kk-, -k- after any consonant → OHG ch — also lacking in northern Upper German and Standard German.

Northern OHG underwent final devoicing, a development that spread southward during OHG times.

Central Franconian was also affected by some developments more typical of Low Franconian (and Low Saxon). Although their influence on the contemporary language is minimal, it should be mentioned that they existed:
 * Early loss of intervocalic -h-: The sequence -eha- in words like OHG merges with the diphthong -ia- (Lux. ).
 * -ft- → -ht- : This (late OHG or early MHG) development may have been fully completed since it occurs throughout Low Franconian, too. Modern Central Franconian generally has -ft-, but the later development of -ht- was subject to outside influences that could not but have blurred the original picture (see 2.2.). The most notable relict forms in Lux. are, alongside non-native , and  from OHG.
 * Loss of nasals before spirants: This development was in all likelihood not identical with the North Sea Germanic spirant law, but might have been related to it. It might also have been just a sporadic tendency rather than a universal rule in northern OHG. The most notable examples in Lux. are (“us, our”) and archaic  alongside non-native (?).

Later developments
The following developments probably occurred during Middle High German times. Several of them are not shared by all dialects of Central Franconian, one or two are even restricted to Lux. and the dialects just to its east.


 * hs, ht → ss, t — The deleted h triggers a peculiar development of the preceding short vowel (see 3.2.1). Later on, h (spelt ch) was reintroduced before t in most cases under outside influences, except in south-western dialects of Lux. Some general exceptions remain, however, as well as hypercorrect forms ( from OHG, now obsolete Jesuiicht for ).
 * Loss of intervocalic -g- — Preceding short vowels are affected by this, undergoing lengthening or diphthongization (see 3.4.1).
 * Intervocalic -mb-, -ld-, -nd-, -ŋg- → -mm-, -ll-, -nn-, -ŋŋ-.
 * Intervocalic -n- → -ŋ- after certain vowels (velarisation) — See 3.4.1. for details.
 * Word-final -n is deleted when the next word begins with a non-alveolar consonant (Eifeler Regel). This is now restricted to grammatical endings and a limited number of frequent words (lexicalisation), but it probably originally affected every word-final -n.
 * Development of sibilants:
 * OHG sc → sch.
 * OHG rs → sch — Again, the deleted consonant (-r-) triggers a peculiar development of preceding short vowels (see 3.2.1).
 * OHG s before l, m, n, p, t → sch (spelt s- word-initially before -p- and -t-) — Unlike most of High German, is usually also pronounced word-internally before p and t, though originally only in southern and central Lux. This development is not active anymore, occurring neither in the superlative ending -st- nor in more recent borrowings.
 * Otherwise: OHG s → s.
 * OHG ʒ (= z, ) → ss.


 * Developments of geminates:
 * Northern OHG -bb-, -dd-, -gg- → -pp-, -tt-, -ck-.
 * After this, all geminates are simplified.

Umlaut
Post-vocalic sch from OHG sc triggered a secondary umlaut. Compare Lux. ,, from OHG , ,. The development is shared by all of Central Franconian as well as Limburgish.

Lux. shows sporadic irregularities concerning umlaut, meaning unexpected umlaut vowels as well as apparent replacement of an umlaut vowel with its non-umlaut equivalent. This situation may point to hypercorrection, though the reason seems quite unclear. Compare for example, Lux. from OHG undar versus Lux. from OHG hiudu.

Lux. underwent unrounding of rounded front vowels, which lead to the following mergers:
 * MHG œ, iu, öu merge with MHG ē, ī, ei;
 * MHG ö, ü merge with MHG e, i, though these two pairs develop slightly differently in closed syllables (see 3.2.2.).

Peculiar developments
Before OHG -r- + consonant and before hs and ht, short vowels undergo special developments, which are best described in a table:

Note that before hs and ht vowels had been shortened prior to this in the following way: Therefore the developments before hs, ht described above apply to the mentioned MHG long vowels and diphthongs as well.
 * MHG ā → a;
 * MHG ī, ie → i;
 * MHG ū, uo → u.

General developments
Otherwise the following developments took place in MHG closed syllables:


 * MHG a →
 * MHG a → — before L and nasals if word-final or followed by a voiceless consonant


 * MHG o →
 * MHG o → — before L


 * MHG ö →
 * MHG ö → — before L


 * MHG e →
 * MHG e → — before ʒʒ and ch


 * MHG u →


 * MHG ü →


 * MHG i →
 * MHG i → — before L

Note: is actually pronounced  except before velars, but the distinction is purely allophonic.

Exceptional forms are particularly frequent for i →, to the degree that words with possibly outnumber the regular forms, except before alveolars. Variants and the surrounding dialects show, however, that is indeed the regular development in all positions (except before L).

The resulting variation between and  was grammaticalized in several words. Compare, 3rd p. sg. present, or , pl..

MHG short vowels in open syllables

 * MHG a, o →
 * MHG a, o → — before nasals


 * MHG e →
 * MHG e → — before nasals


 * MHG u →
 * MHG i →

Pre-nasal →  is a secondary development. The exception is explained by the Eifeler Regel (see 2.2): Since the form without -n would not have been shorted, the one with -n kept the diphthong as well. There are two more exceptions (apart from a few regionally restricted words): and  (from MHG, ). This might mean that the shortening was blocked before older (cf. German, ).

Peculiar developments
For shortening before hs, ht see 3.2.1.

Deletion of invocalic -g- (see 2.2) triggered lengthening or diphthongisation of preceding short vowels (probably because the loss went through a stage of weak consonants, , and/or ). This lead to the following mergers: The latter two mergers are slightly problematic: MHG -ō- and -uo- are only distinguishable in Lux. insofar as -uo- is shortened to -u- before certain consonants (see 3.4.2). In several past participles -og- does undergo this shortening (cf., ). However, unexpected shortenings occur also in a few other irregular verb forms (cf., ). For -eg- as well, there are occasional irregularities, namely instances where it behaves like MHG -ē- (cf., ).
 * MHG -ag- merges with MHG -ā-;
 * MHG -ug- merges with MHG -ū-;
 * MHG -ig- merges with MHG -ī-;
 * MHG -og- merges with MHG -ō- or -uo- (?);
 * MHG -eg- merges with MHG -ei-.

A development shared in principle with Ripuarian is velarisation of -n- into. In Lux. this affects the following syllable types: Lux. velarisation is generally restricted to open syllables, that is to -n- followed by a vowel (unlike Ripuarian). Again, the reason must be the Eifeler Regel (see 2.2.).
 * MHG -ein- →
 * MHG -īn-, -ien- →
 * MHG -ūn-, -uon- →

General developments
Otherwise the following developments took places:


 * MHG -ā- →


 * MHG -ō- →
 * MHG -ē-, -æ- →


 * MHG -ū- →
 * MHG -ū- → — in an open syllable before a voiced consonant


 * MHG -ī- →
 * MHG -ī- → — in an open syllable before a voiced consonant


 * MHG -ou- →
 * MHG -ei- →


 * MHG -uo- →
 * MHG -uo- → — before, and  when not velarised


 * MHG -ie- →
 * MHG -ie- → — before, and  when not velarised

MHG -ā- was shifted further to in a handful of words:, , , , , as well as some preterite forms (where, however,  has replaced other vowels as well and was even used in originally weak verbs, cf.  from ).

Levelling of variation between inflected and uninflected stems
Even more than in most other Germanic languages, the Lux. developments described above led to a great deal of variation between inflected and uninflected stems, particularly due to the variations between closed and open syllables. When word-final unstressed -e was eventually lost (apocope), these variations also became widely unpredictable. Compare e.g. the expected inflections of the nouns (“tooth”, from MHG ) and  (“day”, from MHG, northern form of ):

In contemporary Lux. these inflections have been simplified in the following ways: In Zant, the dative forms have been lost (which is the predominant rule), while the alternation in the plural has been maintained. In Dag, a regularized plural was formed according to the main rule of the masculine, i.e. plural by umlaut ; whereas the old dative sg. and nominative pl. became a new word meaning “the bright hours of the day” (rather than “day” as a unit of time).

The predominant ways of levelling vary by part of speech:
 * Verbs unequivocally generalize the inflected stem (since all MHG verb forms had endings).
 * Adjectives usually generalize the inflected stem, though there are some exceptions with generalization of the uninflected stem.
 * Nouns are usually in the form of the uninflected nominative sg., while the plural may be derived therefrom or may come in its etymologically expected form (thus preserving the alternation).

Stem alternations may also be grouped according to whether they are usually levelled or not:
 * Alternations that are widely preserved:
 * umlaut (which is even used productively in some cases);
 * alternations based on final devoicing (though devoicing is ocasionally expanded to inflected forms).


 * Alternations that are sometimes preserved in noun plurals, but rarely otherwise:
 * alternations due to reduction of intervocalic -mb-, -ld-, -nd-, -ng- — cf. → pl. ;
 * -based vs. -based diphthongs from MHG ū, ī — cf.  (“mouse”, ) → pl.  ;
 * short vowels in closed vs. open syllables — cf. → pl..


 * Alternations that are widely levelled:
 * alternations due to loss of -g- — but cf. → pl. ;
 * alternations due to velarisation in open syllables — but cf. → pl. ;
 * vs. from MHG -a- before L and nasals.