Talk:Lenition

Lenition vs. soft mutation

In my limited experience of discussion of Gaelic, I've seen "lenition" used much more often than in English-language discussion of Welsh. When people are talking about Welsh, they seem to tend to call the same phenomenon "soft mutation". I don't know why this should be. Has anyone else noticed this? and is it something we should mention? Marnanel 16:49, 23 Jun 2004 (UTC)

I can only assume it's a calque from the term traditionally used in Welsh grammar "treigliad meddal". A similar term is used in Breton to refer to the same phenomenon "kemmadur dre vlotaat". A bit similar I suppose to the way people often talk about "hard" and "soft" "g".. Neal (on Breton language Wiki)

It's a practical issue. The initial changes in the Celtic languages are generally referred to as 'mutations'. Irish has 2 (lenition (eg ceann > mo cheann) and nasalisation (ceann > ar gceann)), Scots Gaelic only one (lenition (ceann > mo cheann), Manx two (can't think of good examples off the top of my head). Lenition is by far the more common phenomenon (in terms of occurrence) so it's more talked about. Brythonic languages on the other hand have to cope with up to 4 initial mutations (soft mutation (voicing), aspiration (equivalant of lenition in Goidelic), hard mutation and mixed mutation. Soft Mutation here affects the most sounds so presumably it is the most 'talked' about one. Akerbeltz 22:20 30 June 2007 (GMT)

Miscellaneous comments

I've just removed a whole passage that is rather irrelevant to the discussion of lenition, even though it may be of some value in other articles. I'm quoting it after this. It reads a bit confusingly but the problem is that it mixes phonology with orthographical conventions (a big no-no), and the phonology part is rather sui generis. --Pablo D. Flores 11:00, 1 Feb 2005 (UTC)

Although probably not true to the dictionary definition, or the true meaning of the word, this term is also used, (just as is aspiration) to describe the effect of what sometimes happens when one adds an 'h' after a consonant to form a new consonant sound. The resulting sounds are marked by the passage of air during the consonant sound (producing a new sound), rather than directly afterward (producing the same sound with a hiss/buzz on the end). Irish Gaelic, for instance, uses these principles often. The 'b', in Irish, is very similar to the English 'b'; however, the 'bh' is like a 'b', only the lips don't quite touch, and thus it is like a 'v' sound, without using the teeth (lips only). Similarly, the Irish 'c' is like an English 'k'; the Irish 'ch' is like a 'k' that is not connected (air can still travel through during the sound), producing a sound somewhat like the 'h' in the word 'huge'. Sometimes, English consonants follow these rules, as with the 'th' in 'thing', the 'sh' in 'shell', and so forth.

Just for the record, one doesn't add 'h' after a consonant to form a new sound -- one may add 'h' to a consonant letter to produce a new digraph that may be read as another sound. But 'th' is not to 't' what 'sh' is to 's', and neither is what 'ch' is to 'k' in Irish... Orthography is arbitrary and absolutely irrelevant to the discussion of a phonological process (it's as relevant to phonology as discussing whether you write down the sounds using a pen, a pencil or a brush). --PDF

All instances of this do not completely follow these rules, as the tongue and mouth positions sometimes shift slightly to accomodate a better sound, and thus we get some interesting sound like the 'ch' in chalk, which seems to have its roots in the 'ts' in cats much more than the c in cats or the c in certify (note how the 'c' in some slavic languages makes this sound); with this same train of thought, the English 'j' sound, as in 'jam' has already been altered by these rules, and is the vocal version of the 'ch' as in 'chalk'; the unaltered version would make a sound like the 'ts' in 'cats', only vocalized, thus producing the 'ds' (like a 'dz') sound as in 'odds'.

This is rather confused (and confusing, if you don't know better). The sound represented by English 'ch' is a fricative; 'ts' is two consonants ('t' and 's'), and neither derives from the other in any sense (in English at least).

The 'ch' as used in some languages, does not follow these rules, as it still sounds like a 'k'. It seems probable that such forms follow the true definition of this word.

another sense

This word is also used more generally to refer to any assimilatory-type or loss-type phonological process. Lenition is the opposite of fortition. Then, 3 types of processes: (1) lenition, (2) fortition, (3) metrical. Maybe the idea of lenition comes from Roman Jakobson (?). — ishwar  (SPEAK) 15:18, 2005 Mar 31 (UTC)

the scale goes like what?

The section "the scale goes like this: unvoiced -> voiced -> fricative" is ambiguous: it is not immediately clear to a reader whether the progression is weak-to-strong or strong-to-weak. Perhaps someone who is more certain than I am would be kind enough to add some kind of indication?

Irish orthography

The discussion of matters of Irish orthography seems out of place in an otherwise linguistic article. And this area should be dealt with properly. Welsh and Breton orthographical choices deserve a mention as the contrast between the strategies that have been adopted by modern Gaelic languages is interesting. However, these matters deserve to be dealt with somewhere, and in fact the subject of language communities' orthographical strategic choices are insightful to linguists in showing native speakers' intuitions.

I don't think that anyone would disagree that the orthographies] deserve a separate article. The present text was parked here pending such an article, to be written by someone who knows rather more than I do about the subject (and maybe to provoke that person into doing so!). If you feel you could make at least a first cut, then please do so and others will improve. --Red King 23:14, 12 October 2005 (UTC)

Fortition needs a separate article

I have changed the redirect page at Fortition into a stub; I believe it needs its own article regardless of its rarity. Raifʻhār Doremítzwr 13:00, 11 September 2006 (UTC)

Hm, now that we've sorted out the Scots Gaelic lenition, I could add some on fortition in Gaelic. Due to the omnipresence of lenition the language disallows word initial friatives in most cases (unless caused by lenition) so fricative initial loanwords frequently undergo fortition, for example Norse hallur > talla, Scots yawl > geòla, Scots vervain > bearbhain etc. Akerbeltz 14:45, 23 February 2008 (GMT)

More examples needed

It would be nice to have some examples of diachronic lenition in the Celtic languages, as well, in the article. FilipeS 12:43, 20 October 2006 (UTC)

I think some examples of words that undergo this process would be nice, and would make this article more accessible to people with casual interest in linguistics rather than just people who already have a Ph.D.RSido 02:45, 18 February 2007 (UTC)

Hm I've added examples - in case the person who keeps editing the labials in IPA reads this page: Scots Gaelic does not have palatalised labials anymore, it only has a glide preceding a back vowel so my IPA is correct :) Akerbeltz 01:54, 30 June 2007 (GMT)

Merger?

Consonant mutation, Consonant gradation, Spirantization, Lenition, Fortition and Fortis and lenis all seem to be about the same kind of phenomenon. Perhaps they should be merged. FilipeS 21:28, 2 November 2006 (UTC)

I don't know about merging them, that would end up being a rather mixed up page. But perhaps a main page with a rough outline of consonant change outlining the general prinicple with links to mutation/gradation/lenition etc? Akerbeltz 13:51, 28 January 2008 (GMT)

Labials

Ok some suggestions please... someone keeps marking the Gaelic labials as palatalised but Gaelic doesn't have palatalised labials anymore, only a remnant glide before a back vowel. Any suggestions on how to keep that from happening? Akerbeltz 13:48, 28 January 2008 (GMT)

Maybe a historical note to that effect in the text? And commented-out notes for editors in the chart. kwami (talk) 20:58, 28 January 2008 (UTC)

Kwami, I'm getting a little annoyed with this. What exactly is your rationale behing suddenly claiming that Gaelic cannot lenite /ɲ/ to /n/??? Akerbeltz 10:41, 20 February 2008 (GMT)

As I explained on your Talk page, I'm not saying that the change doesn't occur, only that AFAIK it is not lenition. It would be lenition if it involved loss of secondary articulation, but since you say that is not the case, you've removed its only element of 'weakening'. If you can show that change from a more marked to less marked place of articulation is considered 'lenition', for example /θ/ → /s/, then not only should we restore this, but we should add an new type of lenition to the lead, and we would need to put this in a different category from /n̪ˠ/ → /n/. kwami (talk) 11:11, 20 February 2008 (UTC)

Lenition does not ONLY involve the loss of secondary articulation as any good book on phonology (and the wiki article itself) will tell you but any movement towars higher sonorisation or opening (try Roger Lass Phonology - An Introduction to Basic Concepts (Cambridge Textbooks in Linguistics) for example). Lass, and all other serious linguistic author that I've ever come across are all quite happy to accept the full set of celtic lenition phenomena as lenition. In Goidelic phonology /l/ /n/ and /ɾ/ are without any doubt less marked places of articulation as they are considered to involve less tongue movement than their strong counterparts /ɲ/ /lˠ/ and /rˠ/ (including /ɲ/ as it involves moving the body of the tongue which is 'bigger' than the tip). There is also no reason to remove the labials. If you're being very pedantic, you could mark them as dual articulation but that's excessive. Akerbeltz 13:10, 20 February 2008 (GMT)

It's fine as you have it now (though I reworded it somewhat). I won't bother tracking down your ref, since from your summary it would appear that it doesn't support your point. My problem was that /ɲ/ → /n/ is neither loss of secondary articulation nor increase in sonorisation. What we now have sounds good, and it may be what such authors had in mind, but it is OR on our part and we won't be able to defend it if anyone challenges it—unless we can find a ref that actually says that reduction of place markedness is considered lenition. kwami (talk) 00:57, 23 February 2008 (UTC)

Agreed. It's probably just a bit understudied. Not lenition as a whole but the fine detail especially since it's not desperately common synchronically and reliable diachronic phonlogical data older than 100 years is hard-ish to come by. I'll keep looking for a specific reference though for loss of place. Akerbeltz 14:14, 23 February 2008 (GMT)

Sounds good, and I doubt we'll be challenged in any case. kwami (talk) 18:34, 23 February 2008 (UTC)

Lenition and Eclipsis

I agree about the strange "opening type" thing - never heard of it either. Eclipsis though, if you read the lead, is relevant though it could probably be presented better, perhaps rather than lumping it under Orthography (which should probably be Celtic orthographies or something) we need:

• Mutations in Celtic languages
  • Lenition in CL
  • Sonorisation in CL
  • Other types in CL
  • Orthographic representation in CL

Something like that anyway. Akerbeltz (talk) 15:56, 16 February 2010 (UTC)

When it comes down to it, I actually think we should keep the Celtic mutations out of this article. They have their own article. This article should be about lenition in general as a phonological phenomenon, things like intervocalic spirantization and voicing. Of the Celtic languages, only Manx really has lenition in the sense of this article, things like /"baD@/ for baddey. I'd get rid of the sections "Grammatical lenition", "Orthography", "Consonant gradation", and "Fortition", and just have links to Consonant mutation, Consonant gradation, and Fortition in the See also section. +Angr 16:06, 16 February 2010 (UTC)

They illustrate the topic nicely though. To my mind it's just a little poorly laid out but I really don't see a need for throwing it all out just because these topics have "their own page". People who come here want to get a good idea of the cross-linguistic phenomena without having to read through lots of language specific pages. Akerbeltz (talk) 17:13, 16 February 2010 (UTC)

The thing is, they don't illustrate the topic nicely. They digress into what is really only a distantly related topic that happens (occasionally) to have the same name. +Angr 17:32, 16 February 2010 (UTC)

When you say the Celtic Mutations, do you mean ALL of them or just everything that's not séimhiú as such? Akerbeltz (talk) 17:37, 16 February 2010 (UTC)

I mean all of them except for the (non–grammatically determined) intervocalic voicing/spirantization in Manx. Yes, historically they originated as a kind of lenition (broadly defined), but I think it's just confusing for the reader to encounter them here, at least in such detail. At most, they should be mentioned in passing as a morphosyntactic phenomenon that historically originated as a kind of lenition but isn't any more. +Angr 17:44, 16 February 2010 (UTC)

That's OTT, lenition can be either syn or diachronic and Celtic just happens to be diachronic. Akerbeltz (talk) 17:50, 16 February 2010 (UTC)

That's an oversimplification. Celtic mutations aren't just diachronic lenition, and sometimes they aren't lenition at all (e.g. /s/ → /t/ in an tsagairt, /s/ → /f/ in O.Ir. in phiur, nasal mutation in Welsh, "provection" in Cornish and/or Breton (I can't remember if they both have it), etc. I really think going into any detail about Celtic mutations in this article will simply distract the reader from the main topic. We should be focusing on things like flapping in English, spirantization in Spanish, and that sort of thing. +Angr 18:02, 16 February 2010 (UTC)

No it's not. /s/ → /t/ IS lenition, it just so happens that the article has retained the form int in this case, so technically it's an tshagairt. /s/ → /f/ in the case of sister is an exceptional case of swesōr under lenition turning into hwehūr, the hw labialising into /f/ (hence mo/do/a fiur) whereas ar/far/a yielded ar/far/a siur. Nothing weird about that all. I must confess to a lack of knowledge on the Brythonic side, though I suspect similar processes at work. Akerbeltz (talk) 19:15, 16 February 2010 (UTC)

My point is, /s/ → /t/ and /s/ → /f/ are lenition in the sense of a morphosyntactically conditioned consonant mutation in a Celtic language, but they are certainly not lenition in the sense of a phonological increase in sonority. And intervocalic /t/ → [ð] in Manx is not lenition-the-morphosyntactically-conditioned-consonant-mutation (despite occurring in a Celtic language) but it is lenition-the-phonological-increase-in-sonority. Lenition thus means two very different things, and I think it would be less confusing for readers if we treated them in separate articles: this article can cover the phonological increase in sonority, and Irish initial mutations (or even a more general article on Celtic initial mutations) can cover the morphosyntactic consonant mutation. There are much clearer examples of diachronic lenition_phono that we can use, such as Greek /pʰ tʰ kʰ b d g/ → /f θ x v ð ɣ/. +Angr 21:40, 16 February 2010 (UTC)

Do me a favour and read the lead and maybe look up lenition in a linguistic textbook or two. It is *not* only defined as an increase in sonority but any slide down a gradation scale, including loss of articulatory features. If sonority were the only factor I'd agree with you but that is way too narrow a definition. Akerbeltz (talk) 22:00, 16 February 2010 (UTC)

Yes, phonological lenition can consist of changes other than an increase in sonority, like debuccalization and degemination. That doesn't change my main point: that the term "lenition" is used to describe two utterly different phenomena, and this article is failing to distinguish them. Even the lead is confusing the two issues by first saying, "lenition is a kind of consonant mutation that appears in many languages" and then proceeding to give examples of voicing and spirantization that are not mutations. +Angr 22:19, 16 February 2010 (UTC)

Style of this article

Does this article have to be so relentlessly technical ? Can no concessions be made for the general reader who is intelligent, interested in linguistics, and willing to learn, but does not have professional training ? I hope I fit that description, and I'm willing to meet you experts at least halfway, but there is really no way into this article for me. The main problem is the assumption that the reader is completely familiar with the phonetic alphabet. I realise that the phonetic alphabet is your professional tool, honed to a razor edge to allow you to discuss accurately the fine distinctions between almost identical pronounciations, and that it is absolutely necessary to use it in an article like this. But can't you throw me (a non-professional who is paying you the compliment of being interested in your area of expertise) at least a small bone ?

How about a paragraph or two of introduction which explains the topic at, say, first year University level, with a few examples using a less-technical pronounciation guide ? This could use the phonetic alphabet as well, so people like me can continually improve their grasp of it). I realise that whatever the intro paragraphs say, they won't be 100% agreement amongst you experts that it is accurate enough, but it's surely worth a little imprecision to reach a wider audience and show them how interesting your subject is ?

I hope I've persuaded some of you that people like me are worth bringing on board. I can't believe that everyone would hold with the view that Wikipedia is for experts only. It's certainly a great forum for experts to bring each up to date with the latest detailed knowledge. But surely it's a waste of a potentially excellent educational tool if professionals actively seek to keep others out ..?

A final point - I do practice what I preach. If I see an article on a science topic that I DO understand, which launches straight into postgrad jargon without taking a breath, then I do add a few introductory lines so that the intelligent non-expert can at least understand what the article is about. Surely some of you read articles about topics in other fields than your own, and appreciate a bit of help to get into them? Or would you really expect to be able to understand only articles on linguistics/phonetics ?

Chris Jones 27/May/2011 — Preceding unsigned comment added by 86.186.218.90 (talk) 08:24, 27 May 2011 (UTC)

Chris, the IPA isn't rocket science and there's a lot of stuff on it about, both on the web and Wiki itself. If you can't be bothered to even acquire a simple understanding of it, then I doubt you're interest in linguistics goes far enough for *any* article to ever suit your needs. That much regarding the use of the IPA. Not using it in a linguistics article is a bit like asking someone to write an article on trigonometry without using any angles.

As far as the wording of the lead goes, having re-read it, I concede you might have a point. Which bits of it do you find confusing in particular? I've changed it slightly, let me know if that's any clearer. Akerbeltz (talk) 11:11, 27 May 2011 (UTC)

Angr, fine, but is flapping really a good example itself? As far as lenition goes, it's one of the more obscure examples I'd say and I'm really not sure why you object to the mother example. Akerbeltz (talk) 13:15, 27 May 2011 (UTC)

I think flapping is a good example, because it occurs in English and is widely known (every first-year linguistics student in America has been taught about flapping, and even Brits who don't do it themselves immediately recognize it as a characteristic of American English). The trouble with the mother example is that fact that the t → d → ð is intervocalic is coincidence. Those changes didn't happen because the sound was intervocalic. PIE t changed to PGmc θ by Grimm's Law everywhere, not just intervocalically, PGmc θ became d/ð everywhere Verner's Law applied (so not everywhere that was intervocalic), and OE d became Modern English ð in some words where r followed, but again not generally in intervocalic position. —Angr (talk) 13:41, 27 May 2011 (UTC)

The thing is, the complaint - with some reason - was that the lead is not clear to non-linguists so the fact the US first year linguistics students are aware of it isn't helpful to the rest of humanity :)

Ok, so the mention of "intervocalic" was wrong, well spotted. Akerbeltz (talk) 14:45, 27 May 2011 (UTC)

By the way, I'm *not* throwing the rulebook but isn't the lead getting a bit long and technical again? WP:LEAD suggests no more than 4 paras and the way we're going it's going to be fairly long again. Akerbeltz (talk) 14:49, 27 May 2011 (UTC)

First-year linguistics students have been explicitly taught about flapping, but even non-linguists notice it. Most Americans (Canadians, Australians, Northern Irish) will notice the difference between their t in wait and their t in waiting, and most other English speakers are aware that Americans (Canadians, Australians, Northern Irish) do it. So it's something a lot people with no linguistic background will be aware of. As for length, I think the paragraphs are shorter than WP:LEAD assumes they'll be. But we could merge the synchronic example paragraph and the diachronic example paragraph into one, and maybe the mention of fortition could come out of the lead. —Angr (talk) 15:04, 27 May 2011 (UTC)

They may notice it but when compared to changes like d > ð or p > f, it's not an obvious case of lenition, if only because the difference between d/ð and p/f is very obvious to speakers of whatever English variant but flapping is, well, less obvious. Yes, I think we can shift fortition at the very least. Akerbeltz (talk) 15:11, 27 May 2011 (UTC)

The trouble is d/ð and p/f don't happen in English, and I think most readers will be less intimidated by an example in English than an example in, say, Biblical Hebrew. I really can't think of a better, more familiar, more obvious example of synchronic lenition in English than flapping. —Angr (talk) 15:24, 27 May 2011 (UTC)

Ah right, you're keen to have a purely synchronic one in the lead? How about the ð > v change in some dialects, aka brother to bruvva? But you're right, synchronically it's difficult. Akerbeltz (talk) 16:38, 27 May 2011 (UTC)

I don't see how th-fronting is lenition. The glottaling of t in cat as [kæʔ] found in many dialects is debuccalization and therefore lenition, I suppose, but that doesn't really leap out at one as an example of a consonant being made "weaker" in any way. —Angr (talk) 16:42, 27 May 2011 (UTC)

Aspirates in Sonorization Lenition

Shouldn't the original sound in the "sonorization" chart be the aspirated stop, which then becomes the unaspirated (then voiced, etc)? Jmolina116 (talk) 21:37, 3 October 2012 (UTC)

The starting point could be either but I guess since the topic of that chart is sonorization, the person who made the chart decided to just start at the point before the sonorization step. After all, the starting point could equally be /tt/ or /t:/ or /t'/ - but that's not the issue here. Akerbeltz (talk) 09:10, 4 October 2012 (UTC)

Scottish Gaelic section - two requests

The section on what happens with grammatical lenition in Scottish Gaelic is very clear - well done to those who worked on it. But two things strike me. First, most people will find it far easier to understand (especially the bits about liquids and nasals) if it were possible to have sound files attached to the key examples. Second, what I can't find anywhere in Wiki is a good explanation of blocked lenition. Of course, this is a general article on lenition in world languages, so possibly there is already too much on Scottish Gaelic - in that case maye this explanation should be put somewhere else. But it would be good to have it somewhere. --Doric Loon (talk) 08:16, 29 October 2012 (UTC)

Glad to heard that! I've added a section (sorry, no time for refs right now, bear with me, but shouldn't be controversial as such). Feel free to tweak stuff, I admit it was done in a hurry! Akerbeltz (talk) 10:00, 29 October 2012 (UTC)

I still think we're not helping anyone by lumping morphosyntactic lenition in with phonological lenition. Angr (talk) 22:26, 29 October 2012 (UTC)

And I still maintain that Joe Blogs looking for this thing called lenition won't have a clue which one they're looking for :) Akerbeltz (talk) 22:40, 29 October 2012 (UTC)

Joe Blogs isn't going to be looking for anything called lenition unless he's already had some exposure to linguistics. It's not like it's a term you encounter in newspapers. Angr (talk) 22:48, 29 October 2012 (UTC)

Thanks, Akerbeltz, for that speedy addition. I just read Blas na Gàidhlig, and to me it will do very nicely as a provisional citation. --Doric Loon (talk) 23:24, 29 October 2012 (UTC)

You're welcome Doric Loon! Angr, I beg to differ. At least in the Celtic countries anyone who's been to a Gaelic/Irish/Manx/etc class for more than a week will have heard of the term. Admittedly, most of them are confused by it but notwithstanding, a lot of non-linguistic people in that situation will be aware of the term. Akerbeltz (talk) 00:05, 30 October 2012 (UTC)

Having been to a Gaelic/Irish/Manx/etc. class for more than a week counts as having had some exposure to linguistics. And regardless of the amount of exposure to linguistics, readers will be less confused if this article knows what it's about rather than pingponging back and forth between phonological processes of lenition and morphosyntactic mutations called "lenition" for historical reasons. We really need two articles, Lenition (phonology) (or maybe Lenition (phonetics)) and Lenition (mutation). Angr (talk) 06:54, 30 October 2012 (UTC)

That's a really bizarre definition of "exposure to linguistics"... <shrugs> put it up for a vote and we'll see what the rest of Wikipedia thinks. Akerbeltz (talk) 10:22, 30 October 2012 (UTC)

Just a theory for ancient latin C and G sound - until the present

Letter	stop	→	aspiration	→	affrication	→	advancing	→	present
C	[k] or [c]	→	[cʰ]	→	[cç]	→	[tɕ] (→ [ʈʂ]?)	→	[ṯʃ]
G	[g] or [ɟ]	→	[ɟʱ]	→	[ɟʝ]	→	[dʑ] (→ [ɖʐ]?)	→	[ḏʒ]

Other letters:

Letter J = [j] → [ʝ] → [ʑ] → [ʐ] → [ʒ] (some European languages)
↓
[j] → [ʝ] → [ç] → [x] → ... → [h] (Spanish)

I thought about this for a while, and I hope there are some proofs for it in real... (יהודה שמחה ולדמן (talk) 09:23, 5 June 2013 (UTC))

Here is the approximate development of the Latin sounds into the modern Romance languages and English:

stop	→	palatalization	→	fronting	→	frontinɡ	→	depalatalization	→	deaffrication
[k]	→	[kʲ]	→	[t͡ʃʲ]	→	[t͡sʲ]	→	[t͡s]	→	[s]

Italian and Romanian never fronted to [t͡sʲ], and still have [t͡ʃ]. Castilian Spanish fronted further to [t͡θ] before deaffricating, giving [θ].

stop	→	palatalization	→	approximation	→	fortition
[ɡ]	→	[ɡʲ]	→	[j]	→	[d͡ʒ]

The Latin phoneme [j] underwent the same fortition to [d͡ʒ]. Portuguese, French, and Spanish all deaffricated both to [ʒ], and Spanish further devoiced to [ʃ], backed to [x], and in some regions debuccalized to [h]. An alternative Spanish development is [j] → [ʝ], spelled y. _{(suoı̣ʇnqı̣ɹʇuoɔ · ʞlɐʇ)} nɯnuı̣ɥԀ 23:33, 4 February 2014 (UTC)

Frozen forms

It strikes me as somewhat bizarre to present "sgian-dubh" as an example of "frozen forms involving the other two groups (labials and velars)" because the group involved here is dentals. However, it is a frozen form in the sense that "sgian dhubh" is now quite common while the compound word is always "sgian-dubh", so deleting this example maybe isn't the right approach - rewording "involving the other two groups" would perhaps be a good idea. Michealt (talk) 01:55, 1 June 2014 (UTC)

The word 'too' was missing, well spotted. But as for the fashion item, it's still pretty solidly -dubh in my experience. Of course, any old kitchen knife is indeed dhubh. Probably hard to determine until we get a proper corpus. Akerbeltz (talk) 11:32, 1 June 2014 (UTC)

Synchronic / diachronic

The head of this article makes a distinction between synchronic and diachronic lenition which, in the form described, seems to me to be entirely spurious. When Latin pater becomes Italian padre, that is diachronic, but when Spanish estamos becomes regional Spanish ehtamoh, that is synchronic? Why? Because Italian is "a language" and the Spanish variety is "just a dialect" (whatever that means)? Because the Italian example is reflected in the spelling? Because the Italian change took place earlier? I can't see any difference between these, nor can I see how lenition itself can be synchronic in any meaningful sense. Lenition is a change, and change is by nature "through time". (Fast change is through very little time, but you see my point.) What is of course true is that lenition can be an important part of a diachronic description, and it can be an important part of a synchronic description. The Italian example is a good example of the former, but for an example of the latter you would have to cite the Gaelic case further down the article. However, even then, this is not a distinction between two types of lenition, but rather between two contexts in which we discuss it. Perhaps the people who have been working on this article could revisit this?

The description of Gaelic is very good, by the way, and there are other excellent parts of this article. --Doric Loon (talk) 06:32, 18 January 2016 (UTC)

I agree that the Spanish example sounds diachronic. However, the American English example seems pretty synchronic. I can either choose to lenite the t in waiting or not. By doing so, I'm not splitting or merging phonemes (I don't have that power), so no diachronic sound change is happening. I would just delete the Spanish example and keep the AmE one, though perhaps others will disagree.

I'm unclear on how Celtic consonant mutation is synchronic. Perhaps because it's grammaticalized and doesn't split or merge phonemes (anymore at least). It's a little different from American English, anyway. — Eru·tuon 08:33, 18 January 2016 (UTC)

Celtic straddles both because it has the feature both syn and diachronically. For example, the Gaelic reflex of IE matar 'mother' is màthair i.e. diachronic lenition of /t/ > /h/. This is static and always the case, hence it counts as diachronic lenition. But in the right context, màthair can lenite further if it finds itself in the right grammatical context. For example the 1st pers poss mo causes lenition, so 'my mother' becomes mo mhàthair with lenition of /m/ > /v/. Similar pattern applies across Celtic. Akerbeltz (talk) 09:51, 18 January 2016 (UTC)

This may be a Celtic-coloured view of synchronic lenition but to my mind, most of what the article talks about is diachronic, apart from cases like Sardinian where initials either appear lenited or unlenited depending on context. For ehtamos to be synchronic, it would have to contrast with estamos in the same speech variety, whether dialect or language. I think the easiest fix is to actually change An example of synchronic lenition in English is to An example of diachronic lenition in English is, then most of the rest makes (more) sense. Akerbeltz (talk) 09:57, 18 January 2016 (UTC)

I think my point is that I find it hard to see that the lenition itself is any different. You have a fortis sound, and at some point somebody is lazy and pronounces it as lenis, and maybe that catches on, and gradually the lenited form is observable throughout the language group, at first as an alternative, then eventually perhaps as the only surviving form. That's what's happening in all of these examples. Since that is a mechanism of language change, it seems to me that it is always diachronic. (Remember, diachronic does not necessarily mean historical. If language change is happening right now, that is current rather than historical, but it is still diachronic.)

Now of course, if you have lenited and non-lenited forms in the same language, whether as idiolects or in different regional varieties, or in different forms or the same word or whatever, you can make a synchronic comparison, and when historical lenition has become grammaticalized as in the Celtic languages, it is quite important to do that. But the lenition itself is not a different, synchronic sort of lenition. All that is different is the reference point you find convenient when looking at it.

So I would be much happier if, instead of talking as though there were two distinct types of lenition, we were to talk about how lenition can be analyzed through a synchronic or a diachronic lens. Does that make sense? --Doric Loon (talk) 12:12, 18 January 2016 (UTC)

And just as an afterthought: I wonder if what you mean when you talk about "synchronic" lenition is actually "productive"? If it is still productive, that means new examples can be appearing in the language each day. OR, do you mean that lenited and non-lenited forms appear in allophonic or even phonemic alternation? Maybe these are better terminologies for the point you are wanting? --Doric Loon (talk) 12:25, 18 January 2016 (UTC)

No, they're not allophones, at least not in Celtic. coimpiutair becomes mo choimpiutair and on the odd occasion, Gaelic speakers have been know to accidentally apply this when speaking English (I recall and instance of someone asking for two hickets).

I guess in the sense of the phonetic practicalities, yes, there is no technical difference between syn and diachronic lenition. In either, /b/ can become /v/ and there's nothing about syn or diachronic lenition that says it cannot become /v/. The only difference, as I can see it, is the question of productivity of the rule i.e. instances of non-allophonic lenited and unlenited minimal pairs. Akerbeltz (talk) 13:05, 18 January 2016 (UTC)

PS I do not object to reorganising the page, avoiding this syn/diachronic split into sections and simply make reference under the respective language examples/sections to it being one or the other or a mix. Akerbeltz (talk) 13:09, 18 January 2016 (UTC)

Love the hickets!

We're not really far apart, and I'm probably banging on about something that doesn't really matter. But I do think it's helpful to explore further how any linguistic phenomenon can be viewed either synchronically or diachronically (implying that the phenomenon itself is neither).

My favourite example is English an. Old English ane was shortened to a in modern English, but we kept an before a vowel because it was easier to say. But modern English speakers don't know that, they see the more common a as the base form (it comes first in the dictionary, after all) and think they are adding an /n/ in an apple. They're not - the /n/ was always there. So, a diachronic analysis (one seeking to explain origins) sees an as the norm, shortened to a before a consonant. A synchronic analysis (e.g. in a study of speaker psychology or of the needs of the TEFL class) sees a as the norm, lengthened to an before a vowel. Both analyses are absolutely correct within their own terms.

I think the same is true of Gaelic mhàthair. A diachronic analysis says that the /m/ and the /t/ are lenited for exactly the same reason, namely that they came between two vowels in Protoceltic, and therefore they are a single lenition phenomenon. We don't do anything to PRODUCE the "mh" - it's been there for as long as the "th" has. But of course if we have màthair as our dictionary form, we have an illusion of leniting màthair to produce mhàthair. And if that's what speakers think they're doing, then presumably that's what the synapses in their brains really are doing, history notwithstanding. So we have two different explanations which contradict one another and are both correct. It's a question of whether you look at the problem synchronically or diachronically. Which I think is fun.

But probably I'm being pedantic, so duilich a charaid, I'll stop now. --Doric Loon (talk) 15:09, 18 January 2016 (UTC)

Well, pedantry is fine by me... but you've now steered us onto the thin ice of mental representations of language in native speakers i.e. is Gaelic lenition genuinely a rule about intervocalic consonants in PC still kicking about in people's heads. One thing that to my mind suggests that it can't quite be that is the phenomenon of de-lenition in Goidelic where loanwords are de-lenited to non-historic stops to avoid disallowed initial fricatives, such as vervain > bearbhain. How could this work if instances of lenition were just cases of stored allophones? And that aside, how do you propose improving the article, practically speaking? ;) Akerbeltz (talk) 16:55, 18 January 2016 (UTC)

Is lenition a form of assimilation?

In the cases of lenition I can think of, there is always some feature of the consonant that becomes more like its surrounding sounds. Voicing assimilates the voicing of neighbouring sounds (sonorants or vowels), frication assimilates the continuant feature, and so on. So is lenition considered a form of assimilation in linguistics, or are they separate? And if the latter, why are they separated? Rua (mew) 11:29, 13 January 2020 (UTC)

@Rua: Yes, lenition is considered a type of assimilation. In the old days of autosegmental phonology it was represented as the spreading of [+voice] or [+continuant] or whatever from a vowel to a neighboring consonant. —Mahāgaja · talk 17:13, 28 October 2020 (UTC)

Surely that only applies to diachronic lenition. Grammaticalised lenition can occur in all sorts of phonological environments in which you would not expect lenition due to the phonetic environment. Take the Scottish Gaelic comparative feàrr /fjaːrˠ/ which lenites after as /əs/ to as fheàrr /ə ʃaːrˠ/, that's not a spreading of voice or anything. Or lenition of cù /kʰuː/ after aon /ɯːn/ to aon chù /ɯːn xuː/. Akerbeltz (talk) 20:23, 28 October 2020 (UTC)

Lenition, since it's rooted in articulatory phonetics, is first and foremost synchronic (e.g. Tuscan /kasa/ → [kaːsa], /la kasa/ → [lahaːsa]; Spanish /dama/ → [dama], /la dama/ → [laðama]), and can certainly be described as intrusion of vocalic aperture and/or voicing, ergo a form of assimilation. A structural alternation such as /f/-/v/ of leaf/leaves originated in rule-driven synchronic phonology as lenition, but it's now lexicalized, in no way a living lenition (or assimilation). In the case of something like Latin EMERITA > Spanish Mérida, both are in evidence: /t/ > /d/ in an early round of lenition long ago lexicalized, synchronic lenition /d/ → [ð] in the modern pronunciation [ˈmeriða]. Stage 1 /t/ > /d/ originated as and is the result of lenition, but synchronically today it's merely /d/. Stage 2 /d/ → [ð] is productive lenition as in /la dama/ → [laðama], describable, if one wishes, as a form of assimilation. Barefoot through the chollas (talk) 23:45, 28 October 2020 (UTC)

Sorry but anyone familiar with Celtic linguistics is just shaking their heads at this. In Celtic, it is solely a grammaticalized diachronic phenomenon so one can't just claim it's synchronic because it suits one particular view. Akerbeltz (talk) 09:42, 29 October 2020 (UTC)

@Akerbeltz: "Lenition" means different things in different contexts. There's (synchronic) phonological lenition, which my answer to Rua's question above refers to, and there's morphosyntactic lenition, which is phonological only diachronically, such as the Celtic consonant mutations. Manx actually has both, and they have different effects. On the phonological side, it has voicing and spirantization of intervocalic stops, e.g. brattagh 'flag' /ˈbraðəx/ (= bratach), tappee 'quick' /ˈtavi/ (= tapaidh); and on the morphosyntactic side it has lenition just like Irish and Scottish Gaelic, e.g. e hie 'his house' (= a thaigh) and daa phaitçhey 'two children' (= dà phàisde). —Mahāgaja · talk 11:22, 29 October 2020 (UTC)

I'm not saying there isn't synchronic lention which can be seen as a form of assimilation. But you cannot categorically state in the lede of an article on lenition that it is a form of assimilation across the board. At best you could state that "synchronic lenition is (usually?) a form of assimilation" but then you'd have to explain in the lede what synchronic means. To my mind, this entire question about what lenition really is needs its own section outside the lede. Akerbeltz (talk) 12:06, 29 October 2020 (UTC)

@Akerbeltz:, starting "Oh my", this was written earlier, in response to your note beginning "Sorry". It looks now as though we're both headed in the same direction, i.e. actual phonological lenition, results of phonological lenition. If so, good. -- Oh my. This sort of critique is best addressed elsewhere IMO, but most linguists not specialized in Celtic are aware of (and some shake heads at) the problematic usage of standard terminology sometimes employed by some working in the Celticist tradition. Whoever wrote the third paragraph of the article made a valiant attempt at cleaning up the confusion/conflation by comparing Welsh cath - y gath - y cathod (yet even with that effort, a reader with a basic grasp of phonology might well wonder if today's alternation is really a "change of /k/ to [ɡ]" or /k/ → [g] of times gone by that eventually results in a change of /k/ to /g/ in the erstwhile leniting position).

"In Celtic, it is solely a grammaticalized diachronic phenomenon."

Exactly. Relic forms, now phonologically unpredictable, resulting from lenition that no longer applies. As above, also the grammaticalized diachronic /f/-/v/ alternation of Modern English -- leaf/leaves but Toronto Maple Leafs, and no longer applicable in the enabling (causal, if one is so disposed) V_V environment: proof/prove/proving but proofing, et voilà: the lenition /f/ → [v] is no more. The /f/-/v/ of proof-prove, belief-believe etc. is no longer describable phonologically as lenition.

To the point of the original question, lenition can be viewed as a form of assimilation; the grammaticalized results of lenition can be viewed as assimilatory in origin. Barefoot through the chollas (talk) 15:19, 29 October 2020 (UTC)

Point of order, if anyone is abusing the term lenition in ways it wasn't meant to be used, it's non-Celticists as the term was coined within the continental European school of Celticists in the early 20th century, check the OED ;) That aside, I am ok with your proposal of lenition can be viewed as a form of assimilation; the grammaticalized results of lenition can be viewed as assimilatory in origin :) Akerbeltz (talk) 17:31, 29 October 2020 (UTC)

I agree that the lead shouldn't say "lenition is a form of assimilation", because it's more nuanced than that. The article does already say in the Sonorization section that intervocalically, "lenition can be seen as a type of assimilation of the consonant to the surrounding vowels, in which features of the consonant that are not present in the surrounding vowels (e.g. obstruction, voicelessness) are gradually eliminated." And it occurs to me that there are, at least diachronically, cases of spirantization that are not necessarily intervocalic and that could be considered phonological lenition but not assimilation, such as the spirantization of voiceless stops under Grimm's law and of both voiceless aspirated stops and voiced stops in Koine Greek. —Mahāgaja · talk 18:27, 29 October 2020 (UTC)

Stops turning fricatives is still assimilation, namely spreading of the [+continuant] feature to the consonant. Rua (mew) 18:30, 29 October 2020 (UTC)

A thought: Has everyone read or re-read recently Honeybone's "Lenition, weakening and consonantal strength: tracing concepts through the history of phonology" in Brandão de Carvalho et al. Fortition and Lenition? It should be helpful in some of this discussion, and Patrick has kindly set up a copy on the web, easy to find. Barefoot through the chollas (talk) 20:19, 29 October 2020 (UTC)

@Rua: But sometimes there's no [+continuant] nearby to spread. PIE *pnew- becomes PGmc *fneusaną, but /p/ and /n/ are both [−continuant]. And when Ancient Greek πτερόν becomes Modern Greek φτερό, it would have to be dissimilation because the only place where unaspirated /p/ becomes /f/ is before /t/, which is also [−continuant]. —Mahāgaja · talk 22:11, 29 October 2020 (UTC)

Lenition: voiced consonant > unvoiced consonant?

I've been reading Terry Crowley's An Introduction to Historical Linguistics 2ed. On p. 39 he says: "voiced sounds [e.g. b] can be considered 'stronger' than voiceless sounds [e.g. p]". However, this article says that [p] changes to [b] by lenition/weakening. So which is right? Or are they talking about different things, perhaps diachronic vs synchronic sound change? — Preceding unsigned comment added by 208.127.198.234 (talk) 13:25, 26 October 2020 (UTC)

What's the larger context in which Crowley is stating that? Barefoot through the chollas (talk) 17:54, 26 October 2020 (UTC)

In terms of tension of the muscles in the mouth, we distinguish between fortis and lenis, whereby in English /p/ is fortis (strong) and /b/ is lenis (weak). However on a different level, voicing makes a sound louder and you could understand that as being stronger. My personal subjective feeling when thinking about my own speach is that /b/ is a stronger sound than /p/. So although "lenition" means "weakening", I would prefer to use the technical term: "lenis" means something specific, but "weak" could be understood different ways. --Doric Loon (talk) 12:15, 27 October 2020 (UTC)

Agreed. That's why I asked for the context of Crowley's statement. Barefoot through the chollas (talk) 12:48, 27 October 2020 (UTC)

T-glottalisation is completely wrong as an example

[t̚] > ʔ IS lenition, e.g. [ɡɛt̚ ˈbak] > [ɡɛʔ ˈbak].

But [tʰ] > [ʔ] 1. doesn’t happen on its own, only by analogy to other contexts and 2. would be pretty weird to call a “weakening” even if it did happen on its own

The way it would’ve happened is that “get” became pronounced as [ɡɛt̚] or [gɛʔ] in all contexts except when followed by a vowel. This was then generalised so that it was also pronounced like that before a vowel e.g. [ɡɛʔ ˈəʊvə]. This is not lenition, this change was motivated by making pronunciation consistent across different contexts, as opposed to being lazy with the tongue.

This then would’ve spread to instances of /t/ within a word because for instance “wet a” would be pronounced [ˈwɛʔ ə] so logically “wetter”, which is phonemically the same (even with the same morpheme boundaries) would also be pronounced [ˈwɛʔə]. At that point it makes sense that “better” would be pronounced [ˈbɛʔə] even though the “bet” is not a morpheme on its own. Then stuff like “letter” would get pronounced [ˈlɛʔə]. Again none of this is weakening/lenition; this happens by drawing analogy to other words.

To say that [ˈlɛtʰə] > [ˈlɛʔə] happened by lenition is plain ridiculous. Transient Being (talk) 13:34, 2 September 2024 (UTC)