Euphony, Harmony, Disharmony in language.
Around 1978 (in an unpublished dissertation on Romance linguistics), I used the composite term vowel‑(dis)‑harmony about (a part of) the phenomenon of euphony or harmony, which became criticized for methodological reasons. The issue concerned the distribution of two endings used in the formation of ethnic names to city names, in Italian, Spanish, Catalan and Friulian; the endings are ‑ano/‑ino, ‑an/‑in or -án/-ín. They are distributed so that the form with the vowel a dominates by city names that have an i in the end of the word, while the form with i dominates by city names that have an a in the last part of the word.
In the linguistic tradition the term vowel harmony is used about phenomena which consist in the fact that the speakers of certain languages prefer words with vowels taken from only one category rather than words with vowels from different categories. A classic example is Hungarian: There are two categories of vowels, front vowels (e, é, i, í, ö, ő, ü, ű) and back vowels (a, á, o, ó, u, ú - acute accent indicates length). When the first vowel in a word is a front vowel, all the following vowels in the word must be of that same category. The two categories can not be mixed in a single word. Words can have these vowel sequences: tanuló, munkás or szürke, gyümölcs, but not *szürka or *boröck; There are some exceptions, especially with i and í.
Consonant harmony is, in principle, a fairly similar phenomenon, but reportedly less common. It consists in repetition of the same or almost the same consonant. In the suggested dissertation in 1978, I would discuss the concept of harmony, which is not a very clear term. Long after, I have seen that Togeby, in his article about dissimilation from 1964, p. 649, has had similar considerations of what harmony can mean, using the term une certaine harmonie consonantique about distributions like chant‑onner ‑ touss-otter (Thus avoiding repetition, in this case not choosing *chant‑otter), in other words, also Togeby calls a verbal form harmonic when it has relatively little repetition, opposed to tradition (this seems to have escaped not only my attention but everyone's).
Already Paul Passy, in his dissertation Étude sur les changements phonétiques et leurs caractères généraux, 1891, describes the concept of dissimilation harmonique in these words: "comme pour éviter la trop fréquente répétition du même phonème". Hence harmony can arise from little similarity of two consonants, because the repetition of consonants is perceived as disharmonic, cacophonic. A few pages before, p. 186 et seq., Passy refers to assimilation harmonique, of which french chercher (from sercher) is an example while Latin caeruleus (for *caeluleus, derived from caelus) is an example of harmonic dissimilation. The terminology can be considered unfortunate ‑ as it may raise doubts, but for Passy (and for Togeby, and for the present author) it is in line with the notion that harmony cannot be regarded as a simple and unambiguous phenomenon, claiming e.g. that great resemblance = much repetition = great harmony, little resemblance = little repetition = feeble harmony. Passy also sees no problem in that assimilation and dissimilation alternate, interfere, or support each other: .. il peut fort bien y avoir assimilation à un moment, dissimilation à un autre; ou encore assimilation dans un sens, dissimilation dans un autre. (s. 101).
The Norwegian Hans Vogt also uses »consonant "harmony"« about the phenomenon that words like *pri:r with the same sonant before and after the vowel, do not appear in Norwegian, see his "The structure of Norwegian monosyllables", 1942, p. 22. Again, Heymann Steinthal, Gesammelte kleine Schriften I, p. 150, expresses very clairly this apparent ambiguity: "Übellaut (i.e. disharmony, OHV) entsteht bald durch das Zusammenstoßen gar zu heterogener Elemente, bald durch die Berührung gar zu homogener; der Mangel des Wechsels Bewirkt den Prozeß der Dissimilation" (quoted from Brugmann 1909).
In recent times, in the same year as my mentioned work, M.M. Vihman has used the term in her article "Consonant harmony: Its scope and function in child language"; a few examples: Spanish television ‑> widsezo:n, comiendo ‑> kabiendu, Estonian pati ‑> papi, kristi ‑> tsitsi (pp. 304, 313, 315). Harmony is here = repetition of consonants.
The (apparent) confusion occurs by using "harmony" in two discording meanings:
(I) = Pleasant sound, including avoidance of "unpleasant" repetitions.
(II) = Concord of two sequences, i.e. = repetition.
A further contributing factor to the confusion may be that many people conceive it as natural that repetition of the same or similar vowels can provide a pleasant effect on the ear, acting harmoniously, while repetition of consonants is not considered to possess this property. In linguistic contexts we talk about vowel harmony, not necessarily in the sense that the effect on the ear is pleasant, but only that the repetition is present.
(In music, "harmonic" is used, for instance, about triads consisting of a prime, a third and a fifth. However, these harmonies in music are not sequences, but groups of three notes struck simultaneously.)
BRUGMANN, Karl, "Das Wesen der lautlichen Dissimilationen" ‑ Abhandlungen der Phil.‑hist. Klasse der Schs. Gesell. der Wiss., 27, H. 5 (p. 139‑178), Leipzig, 1909.
PASSY, Paul, Étude sur les changements phonÉtiques et leurs caractères généraux (diss.), Paris 1891.
STEINTHAL, Heymann, Gesammelte kleine Schriften I, Berlin, 1880.
TOGEBY, Knud, "Qu'est‑ce que la dissimilation?" ‑ Romance philology, 17 (1964): 642‑667. Berkeley, 1964. (reprinted in his Immanence et structure, Copenhagen, 1968, p. 96‑121. = Revue romane, Numro spcial 2.)
VIHMAN, Marilyn May, "Consonant harmony: Its scope and function in child language" - Universals of human language, 2: Phonology, J.H. Greenberg, ed., Stanford (California) 1978, pp. 281-334.
VOGT, Hans, "The structure of Norwegian monosyllables" - Norsk tidskrift for sprogvidenskap, 12, 1942, s. 5-29.