[Spellyans] 2013 SWF Review

Daniel Prohaska daniel at ryan-prohaska.com
Thu Apr 18 14:56:44 IST 2013

Dear all, 

Here I try to give some ideas to the first point of the list:

1.)         geminates:

                            a.)         rules for doubling geminates;

                            b.)         lack of ultimate double ‹nn›;

                            c.)          lack of clarity;

			    d.)         vowel length in words which double geminates in plural e reduction of ‹ll›, ‹mm› & ‹rr› medially;

                            f.)          suffixes in ‹-el›, ‹-en›;


SWF Review Issues:

1.)         geminates:

                            a.)         rules for doubling geminates;


Here are the rules currently in use in the SWF as Specified in the SWF Specification (2008, abbr. “SWF Spec”).


SWF Spec (p. 18): “For speakers of some varieties of RMC (including Kernewek Kemmyn), double 
consonant graphs also indicate where to pronounce geminate or long consonants.”
SWF Spec (p.18): “Speakers of Revived Late and Tudor Cornish do not pronounce geminate consonants. Instead, the historical geminate sonorants ‹ll mm nn rr› are 
pronounced as short fortes [lh bm dn rh] in their varieties.”
SWF Spec (p.19): “The SWF indicates geminate or long liquids ‹nn ll rr› in writing where they are actually pronounced long in conservative RMC, or as fortis [nh lh ɾh] in RLC and Tudor Cornish. This principle is extended to the phoneme /m/ which is also spelled ‹mm› where it is historically intrinsically long (as it is in most native words). Sonorants are doubled in those places where pre-occlusion of [mː] and [nː] occurs in Late Cornish – namely, in stressed syllables. Therefore ‹mm ~ bm, nn ~ dn, ll, rr› will be written in monosyllabic words and their compounds as well as in the stressed syllables of polysyllabic words. Vowels that precede double sonorants or pre-occluded ‹bm dn› in such words are short.” {...}
{...} “In compound words, an initial element like penn- ~ pedn- or kamm- {camm-} ~ kabm- {cabm-} retains its double (or pre-occluded) consonant because of secondary stress. Therefore, the SWF writes ‹mm ~ bm, nn ~ dn, ll, rr› rather than ‹m n l r› in compounds like pellgowser ‘telephone’. Double consonants are lost, however, in compounds like diwar ‘(pair of) legs’ (from diw ‘two’ + garr ‘leg’), since in such cases, the stress falls on the first syllable, and the final element does not receive a secondary stress.”
SWF Spec (p.22): “The double voiceless stops ‹pp tt kk› are only written in medial position between two vowels or between a vowel and the consonant ‹y› [j] — places where the stops may be pronounced as geminate or long [pː tː kː].”
SWF Spec (p.23): “The SWF does not write ‹-pp -tt -kk› word-finally because this would suggest a physically impossible pronunciation, such as [hatː] for hatt. Therefore, in the SWF, vowels are always short before ‹p t k› apart from a few easily recognisable loanwords like strok ‘stroke’, stret ‘street’. In medial position, ‹-pp- -tt- -kk-› are retained in words like klappya {clappya} ‘chatter’ and plurals like hattys, hattow ‘hats’, where the stop may be pronounced as a geminate. As with geminate or fortis sonorants, the SWF writes geminate stops only where they are pronounced (by those speakers who distinguish long and short consonants).
Monosyllabic loanwords which have final [k] after a short vowel will be spelled with ‹-ck› rather than ‹-k› in the SWF. This enables the SWF to distinguish words like klock {clock} [klɔk] ‘clock’ which contain short vowels from words like klok {clok} [klɔːk] ~ [kloːk] ‘cloak’ which contain long vowels. Plural forms (klockys {clockys} ‘clocks’; klokys {clokys} ‘cloaks’) and compounds containing these words will preserve the spelling ‹-ck› or ‹-k› found in the root.
Medial [kː] from historical /k + h/ in comparatives, superlatives, and subjunctives will be spelled ‹kk›, as in tekka ‘prettier, prettiest’ and dokko ‘may carry’. The graph ‹ck› will be used medially in loanwords like klockys {clockys} ‘clocks’. Both medial ‹-kk-› and ‹-ck-› may be pronounced as long or geminate [kː] by speakers who recognise distinctions in consonant length.”
SWF Spec (pp. 28/29): “Generally, double fricative spellings indicate that the preceding vowel is short and that the consonant itself is pronounced as a geminate in some varieties of RMC.
The SWF does not write ‹ggh›, ‹ssh›, or ‹cch› (the KK graphs used for [xː], [ʃː], [tʃː]) because:
1. KK ‹ssh› [ʃː] and ‹cch› [tʃː] represent marginal phonemes, and are only found in a few loanwords. In words like passyon ‘passion’, the historical trigraph ‹ssy› which suggests the pronunciation [ʃː] (‹ [sːj]), will be used instead.
2. In the SWF, the digraph ‹gh› in medial position is unambiguous. As discussed in § 4.6 above, a ‹gh› written between two vowels can only stand for the sound KK users pronounce as a geminate [xː]. The vowel which precedes such a medial ‹gh› will therefore be interpreted as short according to RMC rules about vowel length. Where a non-geminate /x/ phoneme occurs between vowels (in words like laha ‘law’, KK lagha), it is realised as [h], and is accordingly spelled ‹h› in the SWF.
However, the trigraph ‹tth› is used for geminate [θː] in comparatives, superlatives and subjunctives like kottha {cottha} ‘oldest’. This graph is attested in the traditional Cornish corpus, and is useful for speakers of some varieties of RMC, since it clearly marks the preceding vowel as short.”
SWF Spec (p. 82): “Other departures from current KK usage include vocalic alternation, a feature of traditional Cornish texts by which ‹y› in monosyllables often becomes ‹e› in polysyllables, and the introduction of phonetic rather than morpho-phonemic spelling in words whose roots contain geminate ‹mm nn ll rr›.”
SWF Spec (p. 84): “In contrast to KK, the SWF writes geminate consonants only where they are actually pronounced. This means that ‹mm nn ll rr› will be restricted to stressed syllables, resulting in alternations like:
bleujen, pl. bleujennow ‘flower(s)’
dyllans, pl. dylansow ‘emission’
In compounds, ‹mm nn ll rr› will be written in syllables carrying secondary stress, like the initial syllable of pellgowser ‘telephone’, but not in word-final position in words like diwar ‘(pair of) legs’.
While this approach may seem tedious at first to people who are used to morphemic spellings, it has a major advantage as well: it tells learners exactly how to pronounce the consonants in question. If they are written double, then they should be pronounced long. In the case of the examples given above, the forms bleujennow and dyllans have long consonants, while the forms bleujen and dylansow do not. The SWF therefore supports correct KK pronunciation of these words.
The Digraphs ‹mm ~ bm›, ‹nn ~ dn›, ‹ll›, and ‹rr› in the SWF 1.These forms will be written in syllables with primary stress:
RMC                                                   RLC                                                                  Meaning
kamm                             ~                kabm                                                              ‘crooked’
kemmer                         ~                kebmer                                                          ‘take’
rann                               ~                radn                                                               ‘part’
ranna                             ~                radna                                                             ‘divide’
pell                                 ~                pell                                                                 ‘far’
gallos                            ~                gallos                                                            ‘be able to ’   
garr                                ~                garr                                                                ‘leg’
terri                               ~                terri                                                               ‘break’

The Digraphs ‹mm ~ bm›, ‹nn ~ dn›, ‹ll›, and ‹rr› in the SWF
2. They will also be written in compound words, in syllables with secondary stress:
RMC                                                   RLC                                                                  Meaning
kammneves                  ~                kabmdhavas                                                ‘rainbow’
pennglin                       ~                pednglin                                                                         ‘kneecap’
pellgowser                                     -                                                                       ‘telephone’
korrdonner                                   -                                                                       ‘microwave’
3. They will not be used in unstressed syllables:
RMC                                                   RLC                                                                  Meaning
kemeres                        ~                kemeres                                                        ‘take’
diwvron                        ~                diwvron                                                                          ‘breast’
galosek                         ~                galosek                                                         ‘powerful’ 
diwar                             ~                diwar                                                             ‘pair of legs’
This third rule will prevent mispronunciations like **kebmeres or **diwvrodn.
For comparable reasons, word-final ‹-pp -tt -kk› do not appear in the SWF because voiceless stops can only be pronounced long in medial position. The SWF thus writes hat ‘hat’ and top ‘top’. In order to distinguish words like klok ‘cloak’ (which has a long vowel) and klock ‘clock’ (which has a short vowel), ‹-ck› will be used to represent the sound [k] after short vowels in monosyllablic English loanwords and their compounds. Medial ‹-pp-, -tt-, -kk-› will occur between vowels, or between a vowel and the consonant ‹y› in words like klappya ‘chatter’, hattow ‘hats’, and lakka ‘worse’. Medial ‹-ck-› will be used instead of ‹-kk-› in loanwords like klockow ‘clocks’.
Of the other KK geminates ‹ggh tth ssh cch›, the SWF writes only ‹tth›, in comparative, superlative, and subjunctive forms like kottha ‘older, oldest’ or hettho ‘may stop’. The forms ‹ssh› and ‹cch›, which are only found in a few loanwords, are replaced by ‹sh› and ‹ch›; the word passhyon will be respelled as passyon, but may still be pronounced as in KK. The sequence ‹ggh› is replaced by medial ‹gh› in the SWF. Since the SWF writes ‹h› rather than ‹gh› for the medial consonant sound in words like laha (KK lagha) ‘law’, a ‹gh› written between two vowels (as in sygha ~ segha ‘drier, driest’) unambiguously represents a geminate [xː].”

For the most part the SWF rules are sound. The only systematic discrepancy I see is where the unstressed part of a compound is concerned, e.g. ‹pellgowser› where stress is on ‹-gows-› and ‹pell-› is unstressed. In other words, such as ‹kemeres› the geminate that appears in the stress simplex ‹kemmer› is simplified and written ‹m› rather than ‹mm› (= **kemmeres). To be more consistent it may be advantageous to simplify the written geminate in ‹pellgowser› as well and write ‹pelgowser› and ‹kemeres›. All in all, with the SWF showing more features of a phonetic rather than a phonemic orthography, geminates should only be written where they are of phonological relevance, i.e.:


              -) Where one or more varieties of RC (‘Revived Cornish’) recommend a pronunciation with a geminate (or fortis) consonant;

              -) Where one or more varieties of RC recommend a short vowel in a monosyllabic word;

              -) Concerning /mː/ and /nː/ where RLC (‘Revived Late Cornish’) shows Pre-Occlusion to ‹bm› and ‹dn›;


                            b.)         lack of ultimate double ‹nn›;


This is problematic only for those RC speakers used to Kernewek Kemmyn (KK henceforth) or Breton. An ultimate, i.e. word-final double-n is written in the SWF if the final syllable carries the main stress of the word, e.g. RMC (‘Revived Middle Cornish’) ‹kedrynn› ~ RLC ‹kedrydn› ‘quarrel’. In such a case writing ‹nn› rather than ‹n› is of phonological consequence. The ‹nn› here shows that one or more varieties of Cornish recommend a pronunciation with a short vowel, with a long /nː/, or with a pre-occluded [dn] in RLC.


I think what KK-users who want word-final (‘ultimate’) ‹nn› mean that they want to derive the from a singular form, e.g. **gwedhenn (KK ‹gwydhenn›), that its corresponding plural form is ‹gwedhennow› (KK ‹gwydhennow›) rather than **gwedhenow.


There are possible arguments against the use of ‹nn› (and other written double consonants) in word-final unstressed position:


-) It was generally not used in traditional Cornish, especially in the case of ‹m(m)› and ‹n(n)›; 


-) It is not relevant to deriving the pronunciation of the singular form. No recommended phonology of RC proposes a distinction in pronunciation between word-final unstressed ‹n› and ‹nn›, not even KK which is the only orthography to spell this morpho-phonemic difference;


-) Writing ‹nn› in word-final unstressed position has the disadvantage of possibly prompting mispronunciations and mistaken placement of stress in the word. In KK teaching material (KK version of Kernewek Dre Lyther) a speaker pronounced the name **Yowann as [jʊˈwadn].


-) The argument that “Breton does it” is irrelevant to Cornish as No recommended phonology of RC proposes a distinction in pronunciation between word-final unstressed ‹n› and ‹nn›, not even KK which is the only orthography to spell this morpho-phonemic difference. This is not the case in Breton. Breton speakers today generally use an interdialectal orthography when writing their language. The South-Eastern Breton dialect of Gwened (Vannes) has word-final stress in polysyllabic words, so for this dialect the distinction between a word-final ‹nn› and ‹n› is phonologically relevant.


                            c.)          lack of clarity;


It would be helpful to have examples and an explanation as to which lack of clarity is meant here, in order to tackle the perceived problems. This point needs elaboration.


                            d.)         vowel length in words which double geminates in plural ‹e› reduction of ‹ll›, ‹mm› & ‹rr› medially;


This point also needs elaboration. It is quite unclear what the problem seems to be. I assume this concerns words like ‹gwedhen› and ‹fardel›. Their plural forms are ‹gwedhennow› ~ ‹gwedhednow› and ‹fardellow›. There are other words with final ‹-l› and ‹-n›, such as ‹aval› (pl. ‹avalow›) and ‹tremen› (pl. ‹tremenow›) which do not show a double consonant in the plural form. This is perceived as an inconsistency by those used to KK spelling, but not so for other Cornish speakers who are used to UC, UCR, KS or Modern Cornish. The question that needs to be answered before coming up with an orthographical solution whether and in what way the singular form should hint at the form and shape of the plural form (and possibly vice versa). Cornish shows a variety of possible plural (and singular) endings that in most cases need to be learnt along with the corresponding singular (or plural) form. There is no way to predict if a singular Cornish noun forms it plural by:

              -) adding an ending (e.g.: -ow, -yow, -yon, -s, -es, -ys, -i/-y, -as); 

              -) dropping an ending (e.g.: -yn, -en);

              -) prefixing dew- or diw- followed by soft mutation;

              -) by vowel-change or ‘umlaut’ (e.g.: troos, pl. treys; dans, pl. dens ~ dyns);

              -) by combining two of the above mentioned features;


It is not necessary to single out certain suffixes that show this feature. A plural form must be learnt along with the singular form. This is one of the difficulties of the Cornish language. A geminate should only be written where it is phonologically relevant.


-) Where one or more varieties of RC recommend a pronunciation with a geminate (or fortis) consonant;

-) Where one or more varieties of RC recommend a short vowel in a monosyllabic word;

-) Concerning /mː/ and /nː/ where RLC shows Pre-Occlusion to ‹bm› and ‹dn›;


                            f.)          suffixes in ‹-el›, ‹-en›;


See above.


On Apr 18, 2013, at 2:18 PM, Michael Everson wrote:

> On 18 Apr 2013, at 11:32, A. J. Trim <ajtrim at msn.com> wrote:
>> As you may know, the MAGA Kernow website now has a link to “Collated issues for SWF review.pdf” / “List of issues raised for the SWF Review”.
>> We have until the end of this month to comment.
>> There are apparently 56 issues.
> Yes, and the list is meaningless without context. I have said the following to Jenefer on the Corpus group list, and I am saying it now here.
> Jenefer, if it hasn't been made clear to you, you cannot expect linguistic experts to comment on nothing but a set of issue titles. Each title came with a rationale from its submitter. WIthout the full set of rationales, it is not possible to give you feed back. And it is not reasonable to ask linguistic experts to guess.
> Please supply a complete document. 
> Michael Everson * http://www.evertype.com/
> _______________________________________________
> Spellyans mailing list
> Spellyans at kernowek.net
> http://kernowek.net/mailman/listinfo/spellyans_kernowek.net

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://kernowek.net/pipermail/spellyans_kernowek.net/attachments/20130418/73b04be5/attachment-0001.html>

More information about the Spellyans mailing list