Non-Repetition in German
Studied mainly within the Distribution of Verbal Prefixes and Nominal Suffixes
Many people have some knowledge of non-repetition in German, though not knowing the word "non-repetition". Germans say heute morgen for "this morning", gestern morgen for "yesterday morning", but for "tomorrow morning" they don't say morgen morgen, but morgen früh, avoiding the repetition. In English the repetition is less striking, because of the variation between 'tomorrow' and 'morning'.
When you to König, 'King', (pronounced with /-iç/) add -lich /liç/, the standard pronunciation of Königlich is differentiated to /ik/ - /iç/ instead of /iç/ - /iç/.
The two suffixes -chen and -lein both mean 'small'. Their distribution is complex: -chen is much more frequent than -lein, in fiction the relation is maybe 10:1; geographically -lein is limited to the South-Eastern part of the German language area (Germany-Austria-Switzerland); in certain cases -lein is considered more expressive than -chen. Besides all these different distribution patterns, there is a phonotactic one, basically non-repetitive: after an -l- near the end of the word basis -chen is preferred: Spiel-chen, Seel-chen, Ställ-chen, after -ch-, -g-, -ng-, -lein is more common: Bäch-lein, Tüch-lein, Äug-lein, Ring-lein. If you go to the details, the question is much more complicated, v. Frans Plank, Morphologische (Ir-)Regularitäten, Tübingen 1981, p. 155-158.
Non-repetition is here proposed as a better term than dissimilation, because the latter is reminiscent of the old debate about dissimilation, and therefore could make believe that the level be only that of diachrony and language history treating odd phenomena in single and often exotic loanwords, apparently displaying themselves in a totally unpredictable way excluding any possibility of achieving some more precise explanation.
German has several other instances, where it can be proved statistically, that - in situations with optionality - repetitive sequences occur relatively less frequent than do others with no or little repetition. It is concluded, that in the German language is active an overall tendency to avoid non-functional repetition. A few examples:
German suffixes used to create nomina agentis. There is a group of suffixes, -er, -ler, -ner, of which the latter two are created by false separation of words with -er. Examples: Künstler, Kriegsgewinnler; Raumbildner, Blechner. If the ending of the word root contains a n, it is likely that people will prefer the suffix -ler. The nearer to the end of the word root the n is, the less probable is the use of the suffix -ner.
German plural nouns have mostly one of these suffixes: -e, -er, -(e)n or -Ø ± umlaut. There are some phonotactic restrictions; leaving alone most of the details we can affirm that:
Masculines and neuters in -e have nearly always -(e)n. NOT -e.
Masculines and neuters in -er have -Ø or -n. NOT -er.
Masculines and neuters in -en have -Ø. NOT -en.
The so-called real verbal prefixes be-, ver-, ge-, zer-, er-, ent-/emp-. Our investigation has demonstrated, that they all (with the exception of ent-/emp-) are used less with a verb that begins with respectively b-, v-, g- z-, r-. Bebauen is less probable than verbauen.
Such non-repetitive patterns constitute a basic condition of the German language, but cannot be only a German phenomenon, and it is in fact the result of a series of previous reports of mine (see bibliography), dealing with Spanish, French, Italian, Romanian, Danish, and Etruscan, and a number of unpublished studies treating Latin, Greek, English, Hungarian, Portuguese, Catalan and Friulan that it is possible to demonstrate exactly corresponding phenomena in these languages. We seem to have identified a LINGUISTIC UNIVERSAL. Its effect, within derivational morphology, can be described preliminarily in these terms:
If, at a certain moment, there is present in a language a group of suffixes or prefixes, rather synonymous, syntactically and pragmatically equivalent, easily differentiable according to their phonetic shape - usually by their consonants or consonant clusters; and is there, among the possible bases, a sufficient amount of themes containing the same or similar consonant-combinations (as those in the affixes), the affixes will, in the course of a certain period, develop a competition or distributive pattern following a phonotactic non-repetitive principle. Such a regulation is generally not as strong as are the rules of the grammar, and so its effect can often be demonstrated only by statistical means, because of the presence of many exceptions due to other rules. Example: -chen / -lein.
Another tendency, on the base of the semantic situation, is to make the dominant affix exclusive, or to differentiate the use and function of the affixes. Example: -ler is preferred by complex bases, -ner by simple ones, they are both used almost exclusively with nominal bases (while -er is used with verbal bases). -ler can have a pejorative value.
The distributive phonotactic patterns here described are a special case of a general tendency to avoid the appending of an affix to words that already ends (resp. begins) with a sequence similar to the affix.
Other solutions are affix haplology or paraphrasing the expression. The latter possibility is, however, the most difficult to demonstrate, in part because an expression does not always reveal what it "substitutes", in part, at another level, because you cannot be sure that there issubstitution. Example of affix haplology: Hamsterer + -erin becomes Hamsterin.
Another type of study, of a more systematic kind, was presented for the first time in the article "Combinations of consonants in stressed syllables in German (An extension of the rules of combination)", by William F. Twaddel, Acta Linguistica (Hafniensia) I (1939): 189-199, II (1940): 31-50. Here T. is studying (all) the unstressed syllables in German words, and finds that these syllables relatively rarely contain a syllable structure C'VC', such as beb, fof, etc., with the same consonant before and after the stressed vokal. Relatively rarely means of course rarer than one would expect statistically by a "random" distribution. Twaddel has here found what I propose to call consonantal circumvocalic non-repetition.
Twaddel's path-breaking work has inspired few other researchers to conduct similar studies. However, there have been made general studies of syllable structure which have contained a little interest of the problem of repetition. Two examples are "Tendencies in the combination of prevocal and postvocal consonants in Swedish monosyllables", by Bengt Sigurd, Studia linguistica 12, Lund 1958, pp. 27-51. and "Studier over Franske enstavelsesord" ('Studies in French monosyllables') (1963), by Bjarne Westring Christensen - "Observations" (1963), Bulletin du Cercle Linguistique de Copenhague, 1941-1965 (Bull. VIII-XXXI), Kbh. 1970, p. 210-211. Apparently neither Sigurd nor Westring Christensen have seen Twaddel's article.