Wright's English Dialect Grammar (1905)

Natural languages and linguistics
Post Reply
Darren
Posts: 899
Joined: Mon Nov 18, 2019 2:38 pm

Wright's English Dialect Grammar (1905)

Post by Darren »

Joseph Wright was the guy for Old, Middle and Modern English around the turn of the century. He was also one of the first people to look at English dialects from a "philological" (i.e. diachronic) perspective, and only just in time – he says himself that within 20 years of him writing the book, i.e. 1925, he estimates that no local dialects will remain in pure form with regular distinct correspondences to Middle English phonemes. It's now another hundred years since then, so I can safely say almost all of this information is thoroughly obsolete. His English Dialect Grammar covers indigenous English of the British Isles, discounting the remaining Gaelic-speaking areas; the first section discusses phonology, the second section grammatical differences, and the third section is a comparative lexicon. I found it very interesting how innovative or conservative some of these dialect forms are, and thanks to the success of Emily's thread, I thought I'd go through roughly the same process.

Wright's transcription is a mishmash between 19th-century philological schemes and nascent IPA; he uses some familiar symbols like /ʃ ʒ ə ɐ/ but also older symbols like /ę̄ ȯ/ etc. I'll try and convert everything into the IPA for familiarity. As far as I can make out he uses ì i e ė ē ę̄ for /i ɪ ɛ e eː ɛː/ and similarly for the back vowels. For low vowels he has æ a à ɒ /æ a ɑ ɒ/ plus ɐ for traditional "/ʌ/". He uses mostly familiar consonants, although for /r/ he doesn't use specific symbols, but notes that it's a trill in Scotland and the northern islands, /ʀ/ in Northumberland and parts of north Durham, /ɻ/ in south Pembrokshire, Gloucestershire, and parts of Hertfordshire and Oxfordshire, and all of the south-west, causing retroflexion of following /t d l n/, and "practically the same as lit. English" everywhere else, which may be either /ɹ/ or /ɾ/ at the time; I'm not sure (although he definitely uses it for /ɾ/ when talking about t-flapping).

I should also note that I follow his conventions for place-names, which for reasons which may be obvious to the more quick-witted of you, do not follow the post-1974 English county boundary shifts.

He doesn't mention anything about America, so I'd be interested to see what American English features appear and which don't.

Vowels of stressed syllables

In this section he makes reference to Old English vowels, since they're more stable than Middle English vowels and some dialects preserve distinctions lost in the standard modern language e.g. moan ~ mown. He notes that closed and open syllables were an important distinction, and that where morphology produced morphs with one vowel in a closed and another in an open syllable (e.g. OE dæg ~ dagas), it's almost always the open syllable alternant that's preserved, both in the standard and in dialects. He also observes that when short æ (later a), e and o were followed by a single consonant + a suffixing containing a sonorant, in both dialects and the standard length is lexically variable.

Short vowels

a (æ, ɔ) in closed syllables

When in a non-marked position, OE æ gave ME a, which then developed as follows:
  • In Scotland /a/ in general, /ɑ/ in the south, /ɛ/ in the centre-east and Ulster;
  • In much of the country, variation between /æ/ and /ɛ/ is seen;
  • /ɑ/ in Peeblesshire, and sporadically in specific lexical items in other counties;
  • /ɔ/ is seen in a few lexical items like apple, gabble, lad in various counties;
  • /e/ in apple and narrow in some counties.
Other minor variants are lengthened /aː/, /æː/, /eː/ in after, apple, arrow, glad, path, and /ɔː/ for lad in much of Scotland, Lothian and Edinburgh. He doesn't mention anything about the LAD~BAD split, although /æː/ is a recognised variant.

Conditioned variation includes the combination ak, which gives:
  • /ɛ/ in parts of western and south-western Yorkshire (e.g. /bɛk/ back)
  • /ai̯ ~ ɑi̯/ in south-central and southern Lancashire and northern Derbyshire (/slai̯k/ slack)
  • In the sequence /aks/;
    • /ɛ/ in a wide swath of counties, especially prevalent in the north and Scotland;
    • axe /eks/ in Buchan, Abderdeen and south Scotland; and /eːks/ in numerous counties, plus axle /eːksl̩/
asp (including from earlier aps), ass (hilarious I know) and ast, which develop as normal (i.e. /ɛ~æ~a~ɑ/) in the north and north midlands, but in the south midlands and the south-east it gives /aː/, and /æː/ in the south-west. Scotland mostly has /ɛ~æ/. One notable lexical entry is ass (hilarious, I know) which is /ɑːs/ in Lothian and Edinburgh, unlike normal English which passed this word by during the TRAP/BATH split for unknown reasons.

OE asc, ME /aʃ/ generally develops as unmarked a, but may be fronted to:
  • /ai̯/ or /ɛi̯/, in e.g. ash, smash, wash (!) in parts of Devon, Lancashire, Wiltshire, Somerset and Yorkshire and /æi̯/ in east-central and western Wiltshire;
  • /aː/ in some northern counties;
  • /ɑː/ in Lothian and Edinburgh (the dynamic duo of a-backing so it seems);
  • /æː/ in Dorset;
  • /ɛ/ in an enormous swath of counties;
  • /ɛː/ in a similar area;
  • /ɔi̯/ in central-southern Lancashire.
The sequence wa has rounded in the literary language to /wɔ/, but in most dialects it doesn't have this effect. Likewise war, wear + consonant give /ɔː/ pretty much only in the standard; it mostly remains /aː(r)/ in dialects. Some local variants include /aː/ in wander, want etc., /ɔː/ in want etc. in numerous counties; /aə̯/ for w-ar in east-central Lancashire, north Derbyshire and north Staffordshire; in the same position /æː/ e.g. warm /wæːm/ in north-northwestern Yorkshire, southwest Lancashire and east Somerset; and /əː/ e.g. /wəːm/ in central-south and southern Lancashire.

When g in the sequence ag was not lenided (e.g. drag vs. day < daġ), /ai̯/ is seen in south-central Lancashire and /ɛ/ in west and southwestern Yorkshire and parts of Lancashire.

Nasalised am, an lengthens to /aː/ in Ayrshire, northern Northumberland, the Isle of Mann, Lothian and Edinburgh (no shit), central Buckinghamshire and east Sussex; /ɑː/ in Ulster and /æː/ in south-east Kent. In Scotland usually /ɑ/ is found. /ɔ/ is also found in similar areas in some lexemes with a high degree of frequency (think Jamaican mon).

and often has lengthening;
  • /aː/ or /ɑː/ in numerous counties;
  • /æː/ in south-east Kent (and, I can assure you, Australia);
  • /eː/ in south Scotland;
  • /ɔː/ in various areas;
  • In some cases quality changes instead;
    • /ɛ/ in southeast Kent, south Ayrshire, Antrim;
    • /ɑ/ in numerous counties especially Ireland and Scotland;
    • /ɔ/ in a lot of areas, especially Scotland, but also in a few words in the literary language like bond.
displats a wide variety, including some variation in the standard, with shifts in both directions, occasionally to ludicrous degrees;
  • /a/ in a list of counties that lasts almost an entire page, in various lexemes at least;
  • /aː/ or /ɑː/ are also frequent, as is short /ɑ/;
  • /æ/ in a fair few dialects; e.g. among /əˈmæŋ/, or /ɛ/ e.g. along /əˈlɛŋ/;
  • /eː/ in e.g. tongs /teːŋz/ (!!) in Buchan, Aberdeenshire, west Forfarshire and east Perthshire;
  • /iə̯/ in e.g. thong /θiə̯ŋ/ (!!!) in south Scotland, north Yorkshire, parts of Northumberland
  • /ɔ/ in another page of dialects, including generally the standard (along, strong, tongs, throngs), although some words have /æ/ in the standard but /ɔ/ in dialects like hang, sang;
  • Long /ɔː/ is also widespread;
  • /ʊ/ or /u/ is seen in much of the north;
  • /ʌ/ appears in various lexical items again in many counties, and has been standardised in monger and a few other lexemes.
  • Similarly for aŋk, we see:
    • /aː/ or /ɑː/ in, you guessed it, Lothian and Edinburgh, plus some other northern counties;
    • Regularly or sporadically /ɛ/ crops up a lot;
    • /ɔ/ appears in much of the Midlands.
Old English amb usually gave ME omb, umb (e.g. womb, comb), but a was retained in Scotland and the North, as well as in the word lamb in most dialects. For comb he directs us to the index, which gives varied forms such as /kei̯m/ (Antrim), /kiə̯m/ (Scotland and the North), /kjem/ (south Northumberland, north Durham), /kwoə̯m/ (Dorsetshire) and /kwɒm/ (west Wiltshire, east Somerset). Much less variation is seen for lamb /læm ~ læːm ~ lam ~ laːm ~ lɑm ~ lɔm/.

ME aȝt produces numerous forms; /a/ in draught, slaughter in some dialects but /au̯/ in goddamn Lothian and Edinburgh (from now on Lothburgh since they're always the same) and southwest Northumberland; in many counties; /æː/ in Leicestershire, Wiltshire and Dorset but /ɔː/ in remaining dialects including the standard. Later on we'll touch on ȝ, which gives /Ø, f, x, k, θ/ etc. For eight numerous dialectal forms arise; in full /axt eːxt ɪxt ei̯t iːt eːt ai̯ə̯t eə̯t aːxt ext æːxt æi̯xt æːt ɛxt iə̯t/.

OE ar, ear, ME ar + consonant gives a baffling variety of rhotic and non-rhotic forms; mostly /a(ː)r ~ ɛ(ː)r/ in Scotland and Ulster, /a(ː)r/ in the North, otherwise /aː/, with /aə̯/ and occasional /əː/ as well. Other variants noted are:
  • /ɑr/ in Lothburgh, south Scotland and Inverness;
  • /ær/ in darn, spark in south Scotland; and /æː/ in a surprisingly large amount of counties (/(j)æːm/ arm, /æːvəs/ harvest etc.);
  • As an extension of this, /ɛr er ɛə̯ eːr eː/ are found frequently in the north and Scotland (/kert/ cart, /weːrn/ warn, /peə̯k/ park etc.);
  • Even further along this path is /iə̯/ in Dorsetshire, east Herefordshire, south Oxfordshire and north Kent;
  • Rounding to /ɔr ɔːr ɔː/ is common in the North and Midlands;
  • In the North, raising to /əː/ is frequent (/bəːli/ barley, /wəːt/ wart etc.), which has crept into the standard in earn.
Of similar variety is al (incl. OE eal). The l is often lost; in all varieties it's lost in calf, chalk, half, talk, walk, and in almost all in balk, salve. The usual outcome is /aː/ or as usual in the standard /ɔː/; short /a/ is seen often in the north (/kaf/ calf etc.) and long /aː/, /ɑː/ almost everywhere, including in numerous words in the literary language like half, salve, calf. Fronted /æː/, /æə̯/, /ɛə̯/ are seen in the south and mid-south. Rounded /ɔː/ is common in the standard and extended to calf in the north, including Lothburgh, and also /oː/, /oə̯/ in the North.
al(l) in final position gives forms with either loss or retention of l; /al/, /aə̯l/, /aːl/, /ɑːl/ are common; fronting to /æə̯l/, /æːl/, /ɛə̯l/ occurs in the South and the latter also in Yorkshire. Again rounding is very common; /ɔl/, /ɔə̯l/, /ɔːl/, etc., with some northern areas raising to /oːl/, /uə̯l/. With loss of l, the gamut is roughly repeated, but with no fronted forms; /a(ː)/, /ɑ(ː)/ are common in Scotland while rounded forms /ɔ(ː)/, /ɔə̯/, /oː/ predominate elsewhere.
alt sometimes retains the l, sometimes not. With retention of l are /al/ in the North, /aːl/ or /æl/ in the South, and /ɔl/ or /ɔːl/ in the middle. Alternatively we have /aː/, /ɑː/ in Scotland, and plain /o/, /oə̯/, /ɔː/, /oː/ in the mid and North. The lexeme halter is distinct, with some dialect forms coming from an i-umlaut form (ME heltir, heltere); whence Antrim /ˈhelðər/, Buchan/Aberdeenshire /ˈhɪltər/. Similarly ald gives a variety, but with considerably more rounding for some reason; thus we find alongside expected northern /a(ː)l/, /ɑ(ː)l/ diphthongised /au̯l/ (much of the North, Scotland), /ɛu̯l/ (Ulster), /ɔu̯l/ throughout and in most words like bold, cold, old, fold, hold, gold etc. in the standard (but not bald), and similarly /ʌu̯l/, /əu̯l/, through /uə̯l/ (Wiltshire, Kent, Lincolnshire, Somersetshire, Dorset) to /wɔl/ (south Devon /wɔld/ old, Kent /twɔld/ told), /wɔə̯l/ (Dorset /bwɔə̯ld/ bold; likewise Somersetshire, Devon, Kent); even /woːl/, /wʌl/. Fronting also occurs with /bæld/ bald (south Scotland), /e(ː)l/ in mid/northern Scotland. With loss of l we see much the same, but with no fronting or pre-glided forms; /au̯ aː ɑ ɑː ɔ ɔː ʌu̯/ etc.


Summary:
  • Plain a: /a ɑ æ ɛ e ɔ aː æː ɛː ɔː/
  • a-k, aks: /ai̯ ɑi̯ ɛi̯ ɛː e eː/ + normal evolutions
  • a-sp, ass, ast: /aː æː ɛ æ ɑː/ + normal evolutions
  • a-ʃ: /ai̯ aː ɑː æi̯ æː ɛ ɛi̯ eː ɔi̯/ + normal evolutions
  • w-a: /ɔ aː ɔː/ + normal evolutions
  • w-ar: /aə̯ æː ɔː əː/ + normal evolutions
  • a-g (not lenided): /ai̯ ɛ/ + normal evolutions
  • a-m, an: /aː ɑː æː ɑ ɔ/ + normal evolutions
  • a-ŋ: /a aː ɑ ɑː æ ɛ eː iə̯ ɔ ɔː ʊ u ʌ/ + normal evolutions
  • a-ŋk: /ɑː ɛ ɔ/ + normal evolutions
  • a-mb is sometimes treated as open-syllable a, sometimes as o; cf. lamb/comb/womb
  • aȝ-t: /a aː ɔː au̯ æː/
  • ar: /ar aə̯r aː aːr ɑr ær æː ɛr ɛə̯ er eː eːr iə̯ ɔr ɔː ɔːr əː/
  • al: /a aː ɔː æə̯ æː ɛə̯ ɔə̯ ɔː oː/
  • al-t: /al aːl æl ɔl ɔːl aː ɑː ɔ ɔə̯ ɔː oː/
  • al (final): /al aə̯l aːl ɑːl æə̯l æːl eə̯l ɔl ɔə̯l ɔːl oːl uə̯l a aː ɑ ɑː ɔ ɔə̯ ɔː oː/
  • al-d: /al au̯l aːl ɑl ɑːl ɛl ɛu̯l eːl ɔu̯l oə̯l ɔːl ol ou̯l oə̯l oːl uə̯l ʌu̯l wɔl wɔə̯l woːl wʌl a au̯ aː ɑ ɑː ɔ ɔu̯ ɔə̯ ɔː ou̯ oː ʌ ʌu̯/ (!!)
Next up, a in open syllables. This hopefully should demonstrate that the range of variation is miles beyond anything that exists today, even taking colonies into account. From now on I'll be condensing information a lot more, but I wanted to give a taste of the real variation first. Comments are welcome; I'd be interested to see how many of these are retained.
anteallach
Posts: 349
Joined: Sun Aug 12, 2018 3:11 pm
Location: Yorkshire

Re: Wright's English Dialect Grammar (1905)

Post by anteallach »

Darren wrote: Sun Feb 02, 2025 4:47 am Conditioned variation includes the combination ak, which gives:
  • /ɛ/ in parts of western and south-western Yorkshire (e.g. /bɛk/ back)
  • /ai̯ ~ ɑi̯/ in south-central and southern Lancashire and northern Derbyshire (/slai̯k/ slack)
I have actually heard back with something like [ai], from a speaker who had several other recessive dialect features. This was actually in South Yorkshire but dialect features do not always respect county boundaries.
User avatar
Emily
Posts: 415
Joined: Fri Aug 03, 2018 6:24 am
Contact:

Re: Wright's English Dialect Grammar (1905)

Post by Emily »

/æi̯ ~ æj/ before /ʃ/ was standard (or at least common) in the us in the early 20th century, i watched a flash gordon serial from the 30s recently and they are pronouncing his name with that glide left and right
Darren
Posts: 899
Joined: Mon Nov 18, 2019 2:38 pm

Re: Wright's English Dialect Grammar (1905)

Post by Darren »

a, æ (ea) in open syllables

This is a generally lengthening environment. The usual result in RP is /ɛi̯/; other significant variants are
  • /ai̯/ in the South (whence its appearance in AusEng);
  • In northeast Norfolk, /æː/ or /ɛː/;
  • /ɛə̯/, /eː/, /ɪə̯/ in the North;
  • But /ɪɛ̯/ in north Cumberland, /ɪa̯/ also in the north Midlands;
  • /eː/ or /iː/ in Cheshire;
  • /ɪə̯/ or /iː/ in Gloucestershire;
  • In counties with /ɪə̯/, /ɛə̯/ is also often used by younger generations (i.e. people who are now 150 years old);
  • In only some words, /aː/, as west Worcestershire /l/ ale, Berkshire /ʃm/ shame;
  • And getting more adventurous, the lexeme lane is prone to rounding; /ɔi̯/ (southwest Yorkshire, south Lancashire, mid Northamptonshire), /oː/ (much of the north and sporadically in the rest of the country), /(w)uə̯/ (the Midlands). Lame, shame also show /ɔː/ in similar parts.
In word-initial position (or post-vocalic in h-dropping behave), a rising diphthong is common, especially where /ɪə̯/ or similar is the norm; e.g. /jadz/ adze (S. Northumberland), /jaːkaːn/ acorn (S. Somerset); also more rarely /jɛə̯/, /jeː/, /jɪ/, /jɪə̯/, /jʌ/, /jə/. This sometimes appears in post-consonantal position as well, although without any obvious conditioning; e.g. /bk/ bake (SE. Northumberland), /k/ care (E. Oxfordshire).

Shortening of open a is fairly common, although sporadic. Low vowels reflect an early shortening (before the GVS), while late shortenings have mid front vowels. Lexemes affected include acorn, adze (incl. the standard < OE adesa), bake, bathe, crane, game, grave, hazel, late, made, make, sake, same, shake, shame, shape, take, which all have /a/ or /æ/ in various dialects. /ɑ/ is seen in crane in Scotland and take in Lancashire. /e/ and /ɛ/ are found in practically all of these words and many others, while /ɪ/ and /i/ are limited to a few lexical items mostly in Scotland.

æ (a) followed by a single consonant plus a suffix containing a sonorant (m n l r) is also often a shortening environment, with variation in the standard e.g. saddle, gather vs. ladle, father. Shortening is seen in various dialects (outside of the standard) in ladle, father, water, cradle. In some cases shortening is evidently late (post-GVS), e.g. /ɛ/ or /e/ in cradle, saddle, father etc. This may be raised further to e.g. /ɪ/ in father (north Durham), bramble, father, gather. Alternatively there's rounded /ɔ/ in gather (west Worcestershire), rather (south Staffordshire, Leicestershire), water (southwest Yorkshire, south Lincolnshire). Other minor variants are short /ʌ/, /ə/ e.g. Perthshire /krədl/ cradle.
Dialects where lengthening occurs have /aː/ in hammer, water in addition to standard father, rather. There's also early (pre-GVS) lengthening to /ɛi̯/ (or whatever else open a gives) in father (southeast Lancashire, north and mid-east Staffordshire, northwest Lincolnshire) and rather (northeast Staffordshire), plus standard ladle etc. Long /ɔː/ is found in water in the standard, plus in hammer in Lothburgh.

ær, ar in open syllables generally gives /ɛə̯/, /ɪə̯/ in England and /eːr/ in Scotland. Other variants are /aə̯/, /aː/, /æː/, /ɛi̯r/, /ɛə̯r/, /er/, /eːr/, /ɪa̯/, /ɪə̯/, /ɪə̯r/, /əː/.

Old English æġ, ME ai has fallen together with ei in all dialects, although not always with long ā as happened in the standard. Results include:
  • /ai̯/ is commonplace, with breaking to /aɪ̯ə̯/ before /l/ or /r/. This is simplified to /aə̯/ in e.g. fair (east Oxfordshire, east-central Wiltshire, east Somersetshire), further /aː/ (central Shropshire, Oxfordshire, northeast Norfolk);
  • /æi̯/ is also frequent, with pre-liquid /æɪ̯ə̯/, /æː/ as expected;
  • /æː/ also occurs, as for long/open ā, in the same areas;
  • The standard and many dialects have /ɛi̯/, while dialectal /ɛə̯/ is also common;
  • The remaining variants for long ā are all seen; /ei̯/ (Scotland), /eː/ (all over the place), /ɪa̯/, /ɪə̯/ (north and midlands), even /iː/, /əː/.

Similarly ag, aw give ME au, generally giving /aː/ or /ɔː/. Some notable quirks include /æː/ in words like claw, draw around Wiltshire and Devon, and the retention of g in the north-midlands in words like /nɛə̯g/ gnaw (Yorkshire, Lancashire), /mɛə̯g/ maw (southwest Yorkshire), /sag/ (southwest Yorkshire) + /sɛə̯g/ (SW Yorkshire, S. Lancashire) + /seːg/ (west Westmorland, S. Lancashire) all saw.
Sol717
Posts: 95
Joined: Mon Apr 29, 2019 2:38 am
Location: Kiwistan

Re: Wright's English Dialect Grammar (1905)

Post by Sol717 »

Darren wrote: Fri Feb 14, 2025 10:23 pm And getting more adventurous, the lexeme lane is prone to rounding; /ɔi̯/ (southwest Yorkshire, south Lancashire, mid Northamptonshire), /oː/ (much of the north and sporadically in the rest of the country), /(w)uə̯/ (the Midlands). Lame, shame also show /ɔː/ in similar parts.
These originate from a Middle English collateral form lone, displaying a pre-nasal rounding of Germanic */ɑ/ to that was widespread in Anglian Old English before becoming restricted to western England during the ME period, except in relic lexemes like those you mentioned. This rounded */ɑ/ merges with Gmc./OE */o/, thus sharing its development; for instance, traditional West Riding dialect has /lɒɪ̯n/ "lane" with a identical vowel to /lɒɪ̯z/ "lose" and /θɒɪ̯l/ "endure", from OE losian, þolian; so too Scots has /loːn/ like /loːz θoːl/.
Wright was undoubtedly aware of this since it is mentioned in his Grammar of the Dialect of Windhill, but incorporating such explanations into the EDG would've massively increased the time and effort needed to complete it (for both author and reader/user).
User avatar
äreo
Posts: 161
Joined: Thu Sep 05, 2019 1:33 pm

Re: Wright's English Dialect Grammar (1905)

Post by äreo »

It's incredible how much variation there used to be. Even now English has enough vowels dancing around to make one dialect sound crazy compared to another, but it clearly used to be much harder to slot things consistently into TRAP, BATH, LOT, etc.
User avatar
Raphael
Posts: 5021
Joined: Sun Jul 22, 2018 6:36 am

Re: Wright's English Dialect Grammar (1905)

Post by Raphael »

äreo wrote: Fri Feb 21, 2025 10:54 am but it clearly used to be much harder to slot things consistently into TRAP, BATH, LOT, etc.
Out of curiosity, is there a list anywhere of all the short, all-caps words that are used to describe or define English vowels?
Lērisama
Posts: 258
Joined: Fri Oct 18, 2024 9:51 am

Re: Wright's English Dialect Grammar (1905)

Post by Lērisama »

Raphael wrote: Fri Feb 21, 2025 11:02 am
äreo wrote: Fri Feb 21, 2025 10:54 am but it clearly used to be much harder to slot things consistently into TRAP, BATH, LOT, etc.
Out of curiosity, is there a list anywhere of all the short, all-caps words that are used to describe or define English vowels?
When I forget I check the list on Wikipedia



In other news, I'm greatly enjoying this thread, but don't have any comments bar a sadness at the diversity we've lost.
LZ – Lēri Ziwi
PS – Proto Sāzlakuic (ancestor of LZ)
PRk – Proto Rākēwuic
XI – Xú Iạlan
VN – verbal noun
SUP – supine
DIRECT – verbal directional
My language stuff
User avatar
Raphael
Posts: 5021
Joined: Sun Jul 22, 2018 6:36 am

Re: Wright's English Dialect Grammar (1905)

Post by Raphael »

Lērisama wrote: Fri Feb 21, 2025 12:05 pm
When I forget I check the list on Wikipedia
Thank you!
Post Reply