Zound changes at the beginning of sentences

May 30, 2014 11:15 AM

Linguistics: Can the beginning of a sentence or phrase be a conditioning environment for sound variation?

Are there any examples of languages where being at the beginning of a sentence or phrase causes a variation in sound, such as voicing, aspiration, stress, tone, to a word?

It seems like something which could well be possible, but I can't think up any examples myself.

Knowledge of a language where this happens, or thoughtful opinions on whether it could, are welcome.

Thank you.

posted by Thing to Writing & Language (6 answers total)

Best answer: You mean a mere phonetic change because the word falls at the beginning of a sentence? In English at least, this certainly doesn't happen as a rule - but phonetic variations are common at the beginnings of sentences because of dialect/laziness/etc. Implosive voiced bilabials instead of plosive bilabials for example. Or depending on the prosody of the preceding sentence, you might voice a normally voiceless phoneme at the start of a new sentence.

I'm sure it happens more when a native speaker of one language is speaking a different language however where some phone is not in their native language, as in Spanish speakers who usually have to being sentences in English that begin with /s/ with /es/.
posted by Lutoslawski at 11:34 AM on May 30, 2014

Best answer: Hmm, most of the sound-change environments I'm familiar with come after (or at the end of) something. I think there are some languages that obligatorily begin a vowel-initial sentence with a glottal stop, and if that eventually led to a sound change in the quality of the initial vowel (such as glottalization), that would work.
posted by Harvey Kilobit at 11:35 AM on May 30, 2014

Best answer: The idea that phrasal stress falls on the highest or edgemost (L or R depending on language) word within a spell-out domain has been pursued by a few people. Feel free to PM if you would like references or more detail.
posted by karbonokapi at 12:15 PM on May 30, 2014

Best answer: I can find a bit of research saying "uh a little bit I guess". Like this shows the comparative lengths of some pairs of stops vary based on whether they're at the beginning of an utterance. But I can't find anything that's not unsatisfyingly subtle.
posted by aubilenon at 2:27 PM on May 30, 2014

Best answer: Not quite what you asked for, but apparently some diachronic consonant deletion in Australian languages is thought to have been at least partly conditioned by utterance-initial position:

Luise Hercus has investigated initial dropping in Arabana-Wangganguru, just to the south of the Arandic languages. Here just ŋ- has been lost, and that only from pronouns, kin terms, interjections (and adverbs that can be used in single word utterances). She suggests an answer to the question of why just these classes of words should be first to show initial dropping — 'the feature that is shared by these three categories of words is that in Arabana-Wangganguru they are precisely the words found most frequently at the beginning of an utterance. Their initial consonant was therefore in the initial position par excellence and most liable to loss.'
— Dixon, The Languages of Australia, 2010, p. 207

In over fifty Australian Aboriginal languages, however, consonants have been lost from the beginnings of words, leaving vowel-initial syllables. As detailed in Blevins (2001b), initial C-loss is not limited to any single genetic or areal group of languages within Australia ... Blevins (2001b) explores several factors which may have played a role in this sound change. ... However, in other languages, the most important identifiable conditioning factor for C-loss is utterance-initial position.
— Blevins, Evolutionary Phonology: The Emergence of Sound Patterns, 2004, p. 165

posted by stebulus at 5:58 AM on May 31, 2014 [1 favorite]

Response by poster: Thank you everybody for your answers. I will think about them further, but for the time being file this in "not wholly proven, but not impossible".
posted by Thing at 8:50 PM on June 2, 2014