Regularity of assimilation/dissimilation

Gavril

Senior Member
English, USA
Hello,

Are assimilation and dissimilation generally thought to be irregular processes (i.e., processes that only occur in certain words), or are they expected to apply to all relevant phonetic environments of a language?


I don't know of many examples of regular assimilation. For example,

- Latin barba "beard" (< *farba < *bhardha) is probably the result of assimilation, yet words with a very similar phonetic structure, such as Old Latin forbea "food", fail to show this assimilation

- Latvian debesis "heaven, sky" can be derived from earlier *neb- (cf. Latin nebula "cloud") via assimilation of *n- to non-nasal -b-, but no such assimilation has occurred in Latvian naba "navel"

Similarly, most examples of dissimilation that I know of are irregular (or at least are not supported by more than one example). E.g., Latin tenebrae "darkness" is thought to be dissimilated from earlier *temebrae, but membrum "limb" did not become *nembrum.

Is irregularity thought to be the exception or the rule when it comes to assimilation/dissimilation of sounds?

Thanks for any help
 
Last edited:
  • ahvalj

    Senior Member
    Manuals tell that Latin has at least one example of a stable dissimilation: the suffixes -āli-, -clo-, -blo- and -bli- change l to r when attached to a stem containing l:
    navālis, aequālis — militāris, lūnāris
    animal, tribūnal — exemplar, calcar
    piāclum, perīclum — lucrum, sepulcrum
    stabulum — lavābrum
    amābilis — salūbris


    When the stem contained both l and r, the suffix kept its original form: plūrālis, līberālis. The exceptions like glaciālis and lētālis can be explained as having been formed after this dissimilation stopped working.

    Debesis in Lithuanian can be influenced by dangus "sky", not sure about Latvian. In namas instead of *damas "house" the direction is opposite.

    Update. Also, it is important that your observations are valid only for the distant assimilation/dissimilation: the adjacent sounds change according to strict laws and this latter type of changes has countless examples in virtually any language.

    Update 2. An intermediate position between adjacent and distant assimilation occupies the umlaut, i. e. an anticipation of some aspect of the vowel quality. As a rule, umlauts are rather regular across the languages, though their results can be levelled back after the assimilation stops working, e. g. after the disappearance of the original vowel that caused it.
     
    Last edited:

    Gavril

    Senior Member
    English, USA
    Update. Also, it is important that your observations are valid only for the distant assimilation/dissimilation: the adjacent sounds change according to strict laws and this latter type of changes has countless examples in virtually any language.
    Right, I should have clarified that the question is about distant (rather than adjacent) assimilation/dissimilation.

    Update 2. An intermediate position between adjacent and distant assimilation occupies the umlaut, i. e. an anticipation of some aspect of the vowel quality. As a rule, umlauts are rather regular across the languages, though their results can be levelled back after the assimilation stops working, e. g. after the disappearance of the original vowel that caused it.


    Umlaut seems like a straightforward case of distant assimilation (assimilation of back vowels to front vowels, and assimilation of lower vowels to high vowels), insofar as it is independent of intervocalic consonants.
     
    Last edited:

    ThomasK

    Senior Member
    Belgium, Dutch
    What I wonder about is: is there some logic for dissimilation? I have the feeling that assimilation is almost natural, almost like wear and tear, whereas dissimilation ... ? For example: is there some reason for "nivel" (niveau) developing into "level"? Peregrino (the through-the-field-wanderer) into pellegrino? IN the latter case I suppose the sounds are phonetically close, but n/l? Any ideas, anyone?
     
    Last edited:

    Swatters

    Member
    French - Belgium, some Wallo-Picard
    Sonorant are continuous consonants (i.e. not like /t/ or /d/ where the airstream stops completely during their articulation) that don't produce turbulences (as fricatives like /s/ or /z/ would).

    Liquids are a subset of sonorants, as are glides (/j/, /w/, etc) and nasals (/n/, /m/, etc).

    So /n/ and /l/ are pronounced with the tongue in a similar position and the same strength of airflow, so switching from [n] to [ɾ] or [l] just involve the relatively small change of closing your velar flap when producing them
     

    ThomasK

    Senior Member
    Belgium, Dutch
    Great to read those explanations, thanks! I knew about the plosives and the fricatives and the glides, but did not know (remember?) the sonorants. I think that subcategory is not common in Dutch), have never heard the term as such. But sonorants make life so much more beautiful, I guess... ;-)
     

    apmoy70

    Senior Member
    Greek
    Dissimilation is one of the characteristics of Demotic Modern Greek (note that this register is not the same with Standard MoGr, demotic (i.e. of the people) is the language variant (L) that evolved naturally from Koine > Byzantine Gr > Demotic Greek).
    It expects the dissimilation of the consonant clusters /pt/ & /kt/ > /ft/ & /xt/ and the height dissimilation of /e/ > /i/ in the environment adjacent to /a/ or /o/ e.g:
    -Classical Greek masc. «κτίστης» ktístēs --> founder, builder, restorer > Demotic Greek «χτίστης»xtis.tis] -->builder, stone mason (deverbative from the ancient v. «κτίζω» ktízō)
    -Classical Greek v. «πταίω» ptaí̯ō --> to nudge, crash into, stumble, err, have bad luck (PIE *pi(e)h₂-u-ie/o- to strike cf Lith. pjauti, to cut) > Demotic Greek «φταίω»fte.ɔ] --> to be to blame, be at fault, be in the wrong
    -Classical Greek masc. «ἀετός» ăĕtós --> eagle > Demotic Greek masc. «ἀητός» & «ἀϊτός» [ai̯ˈtɔs] (both spellings are common, the former is preferred over the latter in school orthography) --> eagle.

    In Standard MoGr due to the fusion of Demotic (L) with the artificial Katharevousa (H) register, we have various collateral words (words with dissimilation when they derive from the Lower Register, that coexist with ancient ones re-introduced from the Higher Register) thus:
    -Standard MoGr «φταίω»fte.ɔ] --> to be to blame, be at fault, be in the wrong coexists with «πταίσμα»ptez.ma] (neut.) --> (legal lang.) misdemeanour < Classical Gr «πταῖσμα» ptaî̯smă (neut.) --> stumble, trip, false step, mistake, error, fault, failure, misfortune
    -Standard MoGr «αητός» [ai̯ˈtɔs] (masc.) --> eagle coexists with «αετοφωλιά» [a.e.tɔ.fɔˈʎa] (fem.) --> eagle's nest («ἀετός» ăĕtós + «φωλέᾱ» pʰōléā --> lair, hole of wild animals (probably Pre-Greek) which becomes «φωλιά» [fɔˈʎa] (fem.) with synizesis), «αετορράχη» [a.e.tɔˈɾa.çi] (fem.) --> eagle's ridge, steep mountain («ἀετός» + «ῥάχις» rʰắkʰis (fem) which becomes «ράχη» [ˈɾa.çi] (fem.) in MoGr)).
     

    Cenzontle

    Senior Member
    English, U.S.
    What about assimilation by adding one (e.g., Romeo > Romero)?
    That would be an intrusive 'r', I think, but that would not be assimilation, but more like reduplication...
    Neither, in my opinion.
    "Assimilation" refers to a phonological process, sounds influencing sounds.
    But don't you suppose (Spanish?) romeo > romero is the result of a morphological substitution,
    by analogy with the many other (Spanish?) words in "-ero"?
     

    Cenzontle

    Senior Member
    English, U.S.
    can you expect that kind of morphological substitution when it is a proper name?
    Oh, I was thinking of the common nouns, Old Spanish "romeo" and modern "romero", synonyms of "peregrino".
    What about Spanish mn > mr ... ?
    That dissimilation is fairly regular; I can't think of an exception.
     
    < Previous | Next >
    Top