Talk:Affection (linguistics)

This needs to be explained in terms of phonology rather than spelling. I'm also not sure why this needs to be a separate article from i-mutation. User:Angr 17:39, 17 July 2006 (UTC)

It is indeed a close call, but I believe a separate article is probably warranted on the basis that in i-mutation, it is the vowel sound itself that is affected by the fronting (or raising). In modern Celtic languages, the vowel sound is rarely altered by affection. The addition of the "slender" vowel causes any adjacent consonant to be palatized and necessitates the addition of another "slender" vowel to the preceeding (or following) syllable in multisylable words. Celtic languages aren't my specialty so I don't know the phonologcial history of the phenomenon, but I'm sure somebody will come along eventually and explain it better.--WilliamThweatt 19:34, 17 July 2006 (UTC)

My hope in starting this article is precisely that - we have no serious work on Celtic philology, and I would like to tempt the people who know more to come and put beef on this. Hence I acknowledge Angr's point about the deficiencies of my first attempt but still think it was worth making. If this article does not grow, then somewhere down the line it may indeed seem worth merging it with i-mutation, but meanwhile, let's see what happens with it. --Doric Loon 19:49, 17 July 2006 (UTC)

I agree with WilliamThweatt's comment, the article is completely wrong in suggesting that there is a fronting of the vowel in any of the examples given for Scottish Gaelic. In each case there is a consonant change only. Here are the three examples with IPA representation of the pronunciation:- cat [kʰaxt] -> cait [kʰaçtʲ] - here there is no vowel change at all cù [kʰuː] => coin [kʰoɲ] (or [[kʰon], depending on dialect) - here there's a raising, but no fronting cur [kʰuɾ] -> [kʰuɾʲ] cuir - here there is no vowel change at all. (I'm using ɾʲ because the encoding used on this page can't express the IPA symbol for a labiodental tap, and ʲ is the accepted usage in Goidelic linguistics for what historically was palatalisation of a consonant). I think also that the statement "In fact, of course, notwithstanding the conventions of dictionary citation, it is the form with the "i" which is the shifted form" is erroneous. One reason is tha the conjugated forms of the verb cuir (cuiridh, chuireas, cuir, chuir, do chuir, chuireadh, chuirinn, chuireamaid, cuireadh, cuirinn, cuireamaid, cuiream, cuirear, chuirear, chuirte, cchuiriste, cuirte, cuiriste and cuirtear) all contain the "i" ("cur" is a noun, not a verb) so it seems odd to claim that the base form is the form without the "i"; another reason is that it is blindingly obvious that the gaelic verb in general is based on the second person active imperative singluar, and not on the verbal noun - for example it seems extremely unlikely that the verbs "tog", "smuain", "eisd", "buail", and "can" are derived from the verbal nouns "togail", "smuaineachadh", "eisdeachd", "bualadh" and "cantuinn" and surely no-one will suggest that the English borrowing "figs" ("fix") is derived from the noun "figseachadh" ("fixing") instead of the other way around. --Micheal213.122.47.26 17:46, 24 March 2007 (UTC)

Thanks for the useful examples with phonetic transcriptions. You might want to add these to the article.

My intention was never to say that the forms with <i> are the base forms in any kind of synchronic analysis, but rather that they are historically the unshifted forms. Compare English "a" and "an". The original form is "an", as Old English had the <n>, but it has been dropped in most cases except before a vowel. So in a diachronic analysis, the form with the <n> is the source, but by any understanding of modern English, the form without the <n> is the norm. The comparison with Rückumlaut in the passage was precisely to make the point that this is one of those interesting cases where the shifted form has become the base form, and the original unshifted form survives as the exception. Please do try to make this clearer if you can, but without making a big thing of what is really a minor point.

Also, the article does not say that the vowels change - but you are right, it should say more clearly that they don't. Historically, they DID - the origin of the whole phenomenon is vowel fronting in early Celtic. --Doric Loon 10:46, 27 March 2007 (UTC)

Uhm, in Celtic linguistics I've never seen the term "i-affection" in the context of Goidelic (or at least not modern Goidelic, there were regressive vowel height assimilations at the Primitive Irish stage), only with regard to British Celtic (Brythonic), where the process (or processes) is early, pervasive, regular and really similar to Germanic umlaut phenomena. There is actually both a- and i-affection - again, very similarly to Germanic. I'd suggest Kenneth Jackson's books.

What you have in Scottish Gaelic is not usually described as i-affection. There is a tendency for vowels, especially a, to be palatalised in front of slender (palatalised) vowels, but this process is purely phonetic. Florian Blaschke (talk) 21:56, 21 April 2008 (UTC)

A further problem is that it is historically inaccurate to claim that i-infection in the forms presented was always caused by an 'i' in the following syllable. In the word 'coin' 'dogs' the following vowel was in fact an 'e'. The more usual explanation for what is here called 'i affection' is that a following front vowel palatalized a preceding consonant (Primitive Irish: /coneh/ > /coɲeh/ and that this palatal consonant added a phonetic (never phonological!) high-front offglide to the preceding vowel. (Old Irish /coɲ/ > [cojɲ]) In the modern Goidelic languages, this phonetic glide has disappeared, leaving the orthographic i as a mere diacritic of the palatal quality of the following vowel. True vocalic alternations are however produced in Old Irish by the influence of following vowels. Thus the short high vowels /i, u/ are lowered to [e, o] under the influence of an /a(:)/ long or short or a long /o:/ in the following syllable, while the mid vowels /e, o/ are raised to [i, u] by a following /i(:}/ or /u(:)/ long or short. This can be seen in the paradigm of beirid, 1sg. ind. pres. abs. biru (raising) and fer, voc. sing. fir (nominative lowered historically). —Preceding unsigned comment added by Manicsleeper (talk • contribs) 20:47, 31 August 2010 (UTC)