From Wikipedia, the free encyclopedia

In phonology, particularly in historical phonology, dissimilation is a phenomenon whereby similar consonant sounds in a word become less similar. It has been claimed that dissimilation results in a form that is easier for the listener to perceive (whereas assimilation results in a form that is easier for the speaker to pronounce), with the implication that such results are in fact the cause of the change. These interpretations misunderstand the nature of the arguments and the reasoning behind them, however. While synchronic patterns of speech production may well be motivated by the need to produce speech with as little effort as possible whilst maximising the perceptual distinctiveness of the results (a goal-directed process), sound change operates blindly, like natural selection in evolution. The perceptal causes of dissimilation appear to centre around contexts in which a single acoustic effect derives from 2 or more sources, but listeners attribute the effect to only 1 of those sources, factoring the acoustic effect out from other contexts. This factoring out of coarticulatory effects has been experimentally replicated numerous times. For example, in a word like Greek *phakhu- "thick" > Greek pakhu- (παχυ-), the aspiration in the original form from both the initial and medial consonants will pervade both syllables at the phonetic level in casual speech, making the vowels breathy. Listeners hear a single effect - breathy voicing on the vowels - and attribute it to only one of the stop consonants, assuming the breathiness on the other syllable to be a long-distance coarticulatory effect. The idea that dissimilation springs from articulatory awkwardness is easily discounted as repetitions of features across word boundaries do not tend to dissimilate, as they might be expected to do if the problem was simply a motor one of multiple articulations of the same or similar segments. Dissimilation strikes within words because for listeners to replicate the words in their own speech, the words must be resolved to a single series of time-aligned motor commands, and in parsing the acoustic signal in this way, coarticulatory effects are undone.

In any case, dissimilation, like assimilation, may involve a change in pronunciation relative to a segment that is adjacent to the affected segment or at a distance; and may involve a change relative to a preceding or a following segment. As with assimilation, anticipatory dissimilation is much more common than lag dissimilation; but exactly unlike assimilation, most dissimilations are in reference to a non-contiguous segment. Also, while many kinds of assimilation have the character of a sound law, relatively few dissimilations do; most are in the nature of accidents that befall a particular lexical item.

Anticipatory dissimilation at a distance (far and away the most common)

• Latin *medio-diēs ("mid-day", i.e. "noon"; also "south") became merīdiēs. Latin venēnum "poison" > Italian veleno. This category includes a rare example of a systematic sound law, the dissimilation of aspirates in Greek and Sanskrit known as Grassmann's Law: *thi-thē-mi "I put" (with a reduplicated prefix) > Greek ti-thē-mi (τιθημι), *phakhu- "thick" > Greek pakhu- (παχυ-), *sekhō "I have" > *hekhō > Greek ekhō (εχω — cf. future *hekh-s-ō > heksō— ‘εξω). Some apparent cases are problematic, as in English "eksetera" for etcetera, which may rather be contamination from the numerous forms in eks- (or a combination of influences), though the common misspelling ect. implies dissimilation.

Anticipatory dissimilation from a contiguous segment (very rare)

• The change from fricative to stop articulation in a sequence of fricatives may belong here: German sechs /zeks/ (the /k/ was originally a fricative). In Sanskrit in any original sequence of two sibilants the first became a stop (often with further developments): root vas- "dress", fut. vas-sya- > vatsya-; *wiś-s "clan" (nom.sg.) > *viťś > *viṭṣ > viṭ (final clusters are simplified); *wiś-su locative pl. > *viṭṣu > vikṣu. English amphitheater is very commonly pronounced ampitheater (though spelling pronunciation may be either some or all of the story here).

Lag dissimilation at a distance (fairly common)

• Latin rārus "rare" > Italian rado. Cardamom the spice commonly cardamon. In Middle English, in a whole list of words ending in -n but preceded by an apical consonant the -n changed to -m: seldom, whilom, random, venom. Eng. marble is ultimately from Latin marmor. Russian февраль /fevrˈalʲ/ "February" is from Lat. Februārius.

Lag dissimilation from a contiguous segment (very rare)

• Proto-Romance *omne "man" > Old Spanish omne > omre > Spanish hombre.
• PRom. *nomne "name" > nomre > Sp. nombre.
• English chimney (standard) > chim(b)ley.
• Proto-Slavic *"sveboda" "freedom" > Slovak "sloboda".

• Vasmer's dictionary

• Assimilation (linguistics)