The Cambridge History of the English Language Volume 5 Part 5 potx

Ossi Ihalainen
Btrwick upoaiwud
SOmUcs
Map 5.1 English dialect districts, 1887 (redrawn from Ellis' map in his On
Early English
Pronunciation,
1889)
2
3
6
The dialects of England since 1776
Line 5: the
tbeeth
line.
The northern limit of the use of the standard form
of
the
and the hiss (th), i.e. [6], in conjunction with suspended (t
v
) as the
definite article, till
the
returns to the north of
line
7.
Line 6: the s.
hoose
line. The southern limit of the pronunciation of the
word
house
as

[hu:s].
The
hoose
line is also the northern limit of
the
[haus]
pronunciation of house.
Line 7: the northern
tee
line.
The northern limit of the suspended (t
v
) for
the definite article.
Line 8: the southern
sum
line in northern England or the southern limit
of the unrounding in words like
some.
Here the direction is 'travelling
from Scotland into England'.
Line 9: the northern
soom
line. The northern limit of any variety of the
[sum] pronunciation (which may be mixed with unrounded pro-
nunciations), 'on proceeding from the Midland counties to Scotland'.
Line 10: the limit between 'L [Lowland] Scotch and N [Northern]
English speech'. The linguistic border is 'not precisely coincident with
the political boundary of England and Scotland': for instance,
' Berwick-on-Tweed and its Liberties, extending 2 to 4 miles into

Bw. [Berwickshire], are linguistically part of England', whereas parts of
Cumberland and Northumberland are assigned to Scotland (Ellis 1889:
21).
Where the linguistic boundary should run seems to have been a
controversial question. In this matter, Ellis' views differed from those of
Murray and Bonaparte (see Glauser 1974: 49-55 for a discussion). This
suggests that the linguistic situation around the border was rather
complex, with spill-overs into the neighbour's territory. Since Ellis' day
the political border seems to have become linguistically more important,
with northern England becoming linguistically more sharply dif-
ferentiated from Scotland.
5.6.2 Ellis'
divisions
On the basis of the ten transverse lines, Ellis divides the dialects of Great
Britain into six principal divisions, which are further divided into forty-
two districts. The districts are further divided into varieties. The
2
37
Ossi Ihalainen
divisions and districts, but not the subdivisions, are indicated on a map
attached to Part V of Ellis' On Early
English Pronunciation
(1889). The
divisions and districts are the following:
I The southern division: districts 1—12
II The western division: districts 13 and 14
III The eastern division: districts 15—19
IV The midland division: districts 20-9
V The northern division: districts 30—2
VI The lowland division: districts 33-42

The main divisions and the districts in England and Wales are shown in
map 5.1 from
On
Early
English
Pronunciation,
Part V. Districts 1 to 3 of
division I represent the ' Celtic Southern', that is Welsh, English, and
division VI English as spoken in the Lowlands of Scotland. These will
not be discussed here. In addition to the' transverse lines' that are used to
differentiate between the principal divisions, Ellis lists a number of other
characteristics that he found within each division. Some of these will be
discussed below. The following chart summarises the criteria used to
differentiate between the divisions, that is, the main dialect areas.
Ellis
1889
South
West
East
Mid
North
Reverted
r
yes
(uncertain)
no
no
no
soom
no

no
no
yes.
yes*
/'
no
no
no
no
yes*
/'and
<*' [6]
no
no
no
yes*
no
hoose
no
no
no
no
yes
The asterisk (*) indicates that the feature characterises much of the area
but not necessarily all of it. Thus much of the north-west of England
does not have the /' realisation of
the
definite article, and there is a small
area in Northumberland where
some

is sum rather than
soom.
Further
differences between the general north and Northumberland include such
lexical oppositions as
summat
vs
something,
thou
vsje,
seet
vs
sight,
slape
vs
slippy 'slippery',
nor
vs
than,
wool
vs
ool
'wool'. The '/' in connection
with th'' realisation of the definite article does not occur in the east or the
south midlands.
Perhaps the most surprising anomaly here is that, besides the fact that
these areas are geographically separated from each other, there is
nothing to separate Ellis' west from his east. Both are 'straight no'
238
The dialects of England since 1776

dialects. On a lower level of abstraction, of course, the differences are
striking. Problems like these pinpoint the difficulty of finding classi-
ficatory criteria that support our judgements about linguistic areas,
are general enough to cover large areas and yet have considerable
discriminatory power.
That a small number of criteria do not identify areas that are mutually
exclusive simply shows something about dialect areas in general and
should not give us concern. For instance, Ellis'
soom
and sum areas
overlap in the east midlands (as they still do) to form a mixed area, a
transition zone (Chambers & Trudgill 1980: ch. 8).
Ellis admits he is not quite happy with the reverted r line in the west.
He says that it exists in his district 13 (the southern part of the western
division), but adds that it is 'generally inconspicuous and often
uncertain, so that it would not be possible to correct line 3' (1889: 176).
The north Herefordshire sample immediately following this passage
shows that reverted and non-reverted realisations of /r/ alternate. The
SED reports r-retroflexion in almost the whole of Ellis' western
division (LAE map Phil 'arm').
The distinction between the south and the west division is further
justified by the observation that some important southern and south-
western (i.e.'Wessex') characteristics (such as the retraction of the r-
sound or the retention of ME
at)
are non-existent or weakened at best in
Ellis'
western division (D 13 and D 14). There is, of course, a lot to
connect the southern part of Ellis' western division to at least the mid
southern variety

—
both have finite
be
and periphrastic
do,
for instance
—
but Ellis' observations about the western division fractures in such
words as
they,
road,
write and
doubt
and their connection to standard
English rather than any indigenous English dialect support his view that
underlying much of the western division English is some type of' Book
English' rather than a ' pure' dialect. The main characteristics of Ellis'
main divisions will be briefly discussed next.
The southern division
The defining characteristic is the 'reverted' or 'retracted' r. Southern
districts
1—3
are called ' the Celtic Southern'. Since this variety occurs on
what Ellis calls 'Celtic territory' - that is, in parts of Ireland and Wales
- it will not be discussed here.
Although he still seems to use this label in its historical sense, Ellis is
aware that the south is linguistically less unified than it used to be. The
Ossi Ihalainen
reverted
r

still prevails over the southern division, ' but the older main
characters, as shewn in D 4, all of which were probably characteristic of
the whole division, fade out gradually to the e. of D 4, and become
complicated with other characters to the w.' (Ellis 1889: 23). To Ellis,
then, the mid southern variety of southern English, which occupies ' the
principal seat of the Wessex tribe' (Ellis 1889: 36), is a paradigmatic,
historically pure representative southern variety of English.
LINGUISTIC CHARACTERISTICS OF THE MID SOUTHERN
(I.E.
'WESSEX') VARIETY
D 4
Linguistic features
of
the mid southern variety
of
southern English
include 'reverted' or 'retracted' r, voicing of initial /s/ and
/f/,
the
realisation of
thr-
as
dr-,
the use of /ai/ in words like
hay
and
may
and the
centralisation of
the

first element in the diphthongs /au/ and /ai/. (The
first element is said to be Bell's vowel number 22.)
The main grammatical characteristics are: finite be {I be
'I
am'),
prefixed participial forms
(a-done
'done'), periphrastic
do
(I
do go
'I go'),
pronoun exchange
(Her told
I'
She told me'),
dn
for 'him' and
'
it', he' it'
(as in, Where's the knife
?
He's in the kitchen
-
where you left
uri),
utch ' I'.
A
point of historical significance about this list is that the older south-
western «-less participial forms, as

in
i-do 'done' and
i-go
have been
replaced by an «-ful form.
16
Ellis'
western division (districts 13 and 14)
Ellis characterises the western division type
of
English
as
basically
Southern English with Welsh influence (D 13), giving in the west the
impression of being 'book English spoken by foreigners or
a
mixture of
S. and M. (D 14), where Southern forms are much used'.
The western division is bounded by the reverted r line and the sum
line;
that
is,
this variety does not retroflex the / sound and has
sum
rather
than soom.
The samples included show that Ellis' western dialects are rhotic,
but the r
is
not exclusively the retracted or reverted variety of Southern

English. As was pointed out above, Ellis nevertheless felt that there was
not enough evidence to make it possible to correct line 3 in the west.
The western division covers portions
of
Monmouthshire, Here-
fordshire, Shropshire in England, and of Breconshire, Radnorshire and
Montgomeryshire in Wales. Hereford is divided: South-Eastern Here-
240
The dialects
of
England since
1776
ford belongs
to
D 4 (the mid
southern),
and the
west
of
Hereford
English becomes more like Welsh English.
The
western division
' represents
on
the east comparatively late,
and on the
west very modern
invasions
of

the English language
on the
Welsh' (1889:
175).
Ellis finds
D
13
an
'imperfect dialect' with
a
considerable amount
of
Welsh influence:
'In D
13
the groundwork is S. English, which has been
altered
by
Celts
in a
different
way
from
D
10, 11"
(i.e., Cornwall,
Devon, West Somerset).
PHONOLOGICAL FEATURES OF DISTRICT13
Among the phonological characteristics of D 13 Ellis mentions
the

'fine
(6)'
for
[A]
('
fine
(6)'
is
a central, schwa-type vowel)
and
the diphthongs
/ai/
and /au/,
which have
a
'fine
(6)'
as
their first element.
(For
a
phonetic interpretation
of
Ellis'
(6), see
Eustace (1969)). Furthermore,
Ellis finds
the use of
[ai]
for

Middle English
ai, a
south-western
characteristic, 'uncertain',
and
initial
%
and v
(i.e.
voicing
of
initial
fricatives) almost extinct;
dr-
for
thr-
(as in
three)
is
lost.
In
other words,
some
of
the
strongest south-western characteristics
are
doubtful here.
As
a

regional idiosyncrasy Ellis mentions
the
form
/a9/
'with'. This
may seem like an irrelevant detail
at
first
sight, but
it
is worth noting that
with
is
usually realised
as
w?
in the
south-west.
The
samples show
forms like
I
be,
/jsnt/ ' isn't'
and
her's' she
is'.
Ellis finds Welsh intonation 'influential'
in
parts

of
the
western
division.
For
instance, Monmouthshire English
is
described
as
'book
English with Welsh intonation
and
Herefordshire
or
Gloucestershire
tendencies' (1889:183). Pitch movements
in
Welsh-influenced English,
as
in the
pronunciation
of the
word likely,
are
likened
to
pitch
movements
in
Norwegian. These intonational features

are
noticeable
even today; they create
a
strange impression
of
West Country grammar
being spoken with
the
' wrong' accent.
The samples illustrating
the
varieties
of
English spoken
in
District
13 actually suggest
to
the
reader clear grammatical affinities
to
south-
western English.
But
this aspect of the data is
not
elaborated
on
by Ellis.

DISTRICT14OF
THE
WESTERN DIVISION
According
to
Ellis, the reverted
r is
totally absent. Unlike
in D
13,
where
this feature
was
'uncertain', 'Southern'
/ai/
(i.e.,
/ai/ for
Middle
English
ai)
in
words like day does occur
in D
14.
The
SED
data
on
Middle English ai, published
in

AES
maps
119 to
130, shows that this
241
Ossi Ihalainen
pattern still obtains. What is interesting here is that D 13 separates D 14
from the larger /ai/-pronouncing area to the south of the western
division. It is possible, then, that D 14 is a residue of an earlier larger
area cut off from it by some later developments. As in D 13, /ai/ and
/au/ have the 'fine (6)' as their first element. On the other hand,
features like the plural present indicative
-en {We bin
'We are',
We do-en
'
We
do') and negations of the type bina
'
aren't' clearly point to the
Midlands and thus distinguish D 14 from D 13.
The eastern division (districts 15—19)
The eastern division covers the whole or greater part of Bedfordshire,
Buckinghamshire, Cambridgeshire, Essex, Hertfordshire, Huntingdon-
shire, Middlesex, Norfolk, Northamptonshire, Rutland, Suffolk and
the London metropolitan area. According to Ellis, eastern division
English has '
a
closer resemblance to received speech than in any other
division' (p. 188). The pronunciation in this area is not quite uniform,

but the differences are'
so
slight that it has been found extremely difficult
to obtain satisfactory information'. This is basically
a
non-retroflex,
sum-
area, but in the north there is a mixed
sum/soom
region. This mixture
seems to have persisted to our days, as can be seen from Chambers &
Trudgill (1980: 129-137). Cockney, which is described as eastern and
east metropolitan in origin, is treated as part of district 17 ('South
Eastern'), but its independent status is recognised by Ellis' division of
the south-eastern varieties into 'Metropolitan English' or 'London
Town Speech' and 'Rural Speech'.
District 19 is East Anglia, one of the areas that, as we have seen, had
attracted the attention of early writers on English dialects. Ellis points
out that Norfolk and Suffolk English are widely known for their
intonation, but regrets that there is no way of describing these
characteristics. Another salient point is the so-called 'French (y), of
which every one speaks' (p. 260). This is the sound in words such as
moon.
He concludes that the Norfolk sound is of recent origin and
different from the Lancashire
moon
vowel. Ellis' symbol for the East
Anglian
moon
vowel is a lower-case upright y with the subscript

v
This
is defined as
' a
modification of Fr[ench] u in a direction not precisely
ascertained' (1889: 87*). The sound is apparently often fractured by
beginning with the mouth too open. Ellis concludes his discussion of
the
moon
vowel by stating that' the exact analysis of this curious sound
is still to be made'. Dialect writers represent this sound by
ew,
as in tew,
242
The dialects of England since 1776
or by u as in
mune.
More recent transcriptions of East Anglian English
show a centralised vowel, which may be diphthongised (Kokeritz
1938/9:
41-8; Lodge 1984: 110-20). Kokeritz, however, finds that in
Suffolk English the
moon
vowel shows a great deal of variation and that
it is often diphthongised, as in [jeu] 'you'.
Other East Anglian characteristics referred to by Ellis are the
following: words like name have a monophthong /e:/ (at least in
Norfolk) and words like
boat
(i.e. words with an original OE

a:)
have
/o:/.
The
ride
and
bouse
vowels show considerable variation.
IF
and
v
are
commutable, but Ellis believes only the use of
w
for v to be indigenous,
whereas v for w is a hyperurbanism resulting from an attempt to speak
received English. East Anglian English has a 'euphonic r', that is, an r
sound in contexts like
draw-r-ing
'drawing'. It is one of the few local
dialects that do not drop aitches. (For the subsequent spread of h-
deletion to East Anglia, see Trudgill (1974,1983: 76-7, 1990: 50).)
Considering how common the glottal stop is in this area today, it is
somewhat significant that Ellis makes no mention of it. If this is not
inattention, the development of the glottal stop in this area must have
been recent and very rapid.
A distinctive grammatical characteristic is the use of uninflected third
person singular forms like
He know
it. Interesting from the viewpoint of

the development of dialect areas is Ellis' observation that, although
positive forms of finite
be,
as in I
be tired
did not occur, negative forms
did. This shows that the present finite
be
area was much larger in Ellis'
day. It also shows that the retreat of finite
be
was gradual, with certain
contexts retaining older forms longer than others. There are similar
observations by Ellis from the east midlands, where finite
be
is obsolete
today. The samples given by Ellis also show traces of the a- participial
prefix. But these appear to be exceptional and the prefix was apparently
being more and more confined to the south-west.
The midland division (districts 20-9; west midland 22, 25, 28)
The midland division, which is bounded on the south by the northern
sum line (line 1) and on the north by the northern
theeth
line (line 5),
covers all Cheshire, Derbyshire, Leicestershire, Lincolnshire, Not-
tinghamshire, Staffordshire, the north of Worcestershire and most of
Warwickshire, south and mid Lancashire, the north-east of Shropshire.
It also extends into Wales, covering 'all detached or English Fl. [i.e.
Flint],
a small part of main, or Welsh FL, and of Dn. [i.e. Denbigh]'.

2
43
Ossi Ihalainen
Dialectally the midland division falls into two distinct and apparently
unrelated sections, an eastern comprising D 20 (Lincolnshire) and a
western comprising the rest (p. 290).
Ellis points out that the midland area is not homogeneous and one
cannot look for'
any
one pervading character', but it must be' defined by
negatives': it does not have southern, western, eastern or northern
characteristics. The linguistic points that Ellis regards as particularly
significant include short
u,
a for [u:], the diphthongisation of
[i:],
which
according to Ellis is 'the first step in the change of [i:] to [ai], and the
development of OE
u\
(standard English [au]), which in some part of the
west midlands has undergone a further development to a mon-
ophthongal
[a:].
Ellis finds r, when not before a vowel, totally vocalised
in D 20, although he admits that this sound caused great difficulty even
to phoneticians, and reliable information from lay assistants (who could
not always keep spelling and pronunciation apart) was hard to get.
Finally, Ellis found that in the midland division h was universally
dropped.

Grammatical forms of interest are the definite article, the present
indicative plural marker
-en,
the form for I
am,
hoo
and
shoo
for
she.
The
definite article has four forms, [&9, 5, 9] and suspended (t
v
) in D 21-
D 27, but there is much variation in their use. The plural marker, as in
you
know-en'
you know', is universal in D
21,
D 22, D 25 and D 26. In D
23 it occurs in a few contracted forms {Anyo ?' Have you ?'
Dunyo
?' Do
you'?).
In D 24 it is only found at the borders of D 22 on the west and
D 26 on the south. In D 27 it seems to be practically lost, but Ellis feels
this is a recent development. In D 28 it is 'plentiful'. In D 29 it chiefly
exists in contracted forms, and ' more in the west than east, but even in
Leicestershire there are traces of it'.
The form I am separates the midland division from the northern

division, which has I
is.
Invariant
be,
as in I
be
'I am', is seldom used, and
most frequently in the negative I
ben't;
it is confined to the parts of the
southern midlands which border on the south division. The pronoun
hoo
'she'
is prevalent in D
21,
D 22, D 25 and D 26, although here also
her
may be used for
she.
The form
shoo
occurs in district 24.
The above features can still be found in various degrees in the SED
material. The present indicative marker
-en
shows an area that
is
basically
the area given by Ellis. Thus the SED shows that with true verbs (SED
VIII.5.1.

'They go to church', IV.6.2 'They keep hens') the stronghold
of
the -en
suffix is Cheshire, northern Staffordshire and west Derbyshire,
although it occurs in the adjoining parts of Lancashire, Yorkshire and
244
The dialects of England since 1776
Shropshire. The use of
-en
with true verbs like
go
and
say
appears to be
confined to the north-west midlands. But the
-en
suffix with
be,
as in
they
bin,
I
bin,
is centred further south on Shropshire.
The northern division (districts 30—2)
The northern division is bounded on the south by the northern
theeth
line (transverse line 5) and on the north by line 10. The area covers the
entire North and East Ridings with some of the West Riding of
Yorkshire, northern Lancashire, most of Cumberland and North-

umberland, all Westmorland and Durham. Much of
this
is characterised
by the following features. ME
u:
is retained, as in o't
wrang house
/ot ran
u:s/
'of the wrong house' (1889: 520). With the exception of the area
north of
line
7, the definite article is the suspended (t
v
). I
is is
the regular
form for I
am
in most of
the
northern division. Words like
nose
and
moon
have [iu] or
[ia].
'The letter
r
occasions considerable difficulty', but Ellis

concludes that in the east post-vocalic r 'practically disappears' and even
in the west 'its power is very small'. R is retained post-vocalically in
Northumberland, where it is realised as a uvular sound (as opposed to
the more usual 'gently trilled' r of northern English). This is the
' Northumbrian' burr, first commented on by Daniel Defoe in his Tour
thro'
the whole Island of Great Britain (1724-6). It is perhaps of some
interest that Ellis believes the uvular r to be ' rather a defective utterance
than a distinctive dialectal pronunciation' (p. 495). The distribution of
post-vocalic r that emerges from the SED material is surprisingly
similar to Ellis' description (see e.g. LAE map Phil 'arm').
Of historical interest is Ellis' observation that 'the guttural (kh) [i.e.
the voiceless velar fricative] has practically vanished from the N.'
However, on passing the Scottish-English border, both the guttural
(kh) and r become 'strong'.
Finally, Ellis feels that parts of 'north Cumberland' and 'north
Humberland' belong to Scotland linguistically. Ellis' view of the
geography of the Scottish—English linguistic border differs from those
of Murray and Bonaparte.
5.6.3 Realisation of Ellis' test
sentence
' You
see now
(that) I'm right' in
the main
divisions
As an illustration of
some
of the differences revealed by Ellis' evidence, I
list realisations of 'You see now (that) I'm right', which is part of the

dialect test. The phonetic exegesis applied to Ellis' palaeotype is that of
245
Ossi Ihalainen
Eustace (1969), but it should be borne in mind, as Eustace reminds us,
that the palaeotype is often ambiguous, occasionally intentionally vague
and never easy to interpret. In several places Ellis himself points out that
the rendering is doubtful or arbitrary or the symbol ambiguous.
Division I, the southern division, district 4 (Montacute,
Somerset; Ellis 1889: 85)
[3i:
d3
zi:
nAu oat/Ai bi: jAit]
Division I, the southern division, district 9 (Wingham, East
Kent; Ellis 1889: 142)
[jeo
si:
neu dat/oi a:j
JDU]
Division II, the western division (Lower Bache Farm, near
Leominster, Herefordshire; Ellis 1889: 176)
[ju
si-
n3u 3i bi- rait]
Division III, the eastern division, district 19 (Great Yarmouth,
Norfolk; Ellis 1889: 278)
[ja
si:
neu
Aim

rAit]
Division III, the eastern division, district 19 (Stanhoe, near
Wells-next-Sea, Norfolk; Ellis 1889: 264)
[JAU
si:
nau 53t a:im jait]
Division IV, the midland division, district 28 (Hanmer, English
Flint; Ellis 1889: 453)
[ja si:n
n3U
az 3im ri't]
Division V, the northern division, district 30 (Goole and
Marshland; Ellis 1889: 522)
[ju
si:
nu: at aizriit]
Some important linguistic characteristics of the main divisions are
brought into sharp relief
by
this simple sentence. For instance, 'I am' is
realised in different ways (7
be,
I
are,
I
am,
I
is),
there are differences in
verb agreement with

see,
and different forms of
that
appear (northern at
vs west midland
as).
There are differences in the pronunciation of and
incidence of /r/, differences in the pronunciation of diphthongs and so
forth. The following points are worth specific comment. Speakers of
Somerset English who have periphrastic
do
today find it obligatory to
mark the second person singular present indicative form with -st,
246
The dialects of England since 1776
realised as [s], as in
thee
's
know
'you know'. Therefore, it is likely that
the [Si: d3 zi:] realisation of'thee dost see' in the Somerset sample above
is an instance of assimilation. Ellis also has doubts about the precise
phonetics of the diphthong of the Montacute
now.
However, this is a
particularly problematic region from the viewpoint of/au/. I have also
heard diphthongs with fronted first elements of different heights in this
part of Somerset and it seems that Ellis' reluctance to take a stand on the
'pure' pronunciation of this sound is a sign of good judgement.
5.7 Wakelin on Ellis' divisions: the extent to which Ellis' areas are

still recoverable
Wakelin (1977: 102) believes that 'when English dialects are classified
again (if they ever are) their remnants will be seen to correspond
remarkably well with Ellis's results'. The data of the SED would in fact
make it possible to look at the present form of some of Ellis' districts.
Unfortunately, not much work along these lines has been done.
However, research on south-western English done by scholars like
Fischer, Wakelin, Viereck and Klemola suggests that Ellis' districts
could still be captured from the SED material, although possibly
reduced in size. In his study of Stafford English Gibson (1955: 306)
compares the relevant SED data with Ellis' and concludes that 'the
dialect situation has not altered a great deal since 1889'. And to give one
quite specific example, the area where Ellis found the plural present
indicative -n suffix with true verbs (i.e. verbs other than the auxiliaries,
have
and
be)
is practically identical with the area that emerges from the
mid-twentieth-century SED material, with Cheshire and northern
Staffordshire as its heartland (Ihalainen forthcoming b).
5.8 Dialect areas today
Our knowledge of the various dialect areas in the late twentieth century
is largely based on the evidence provided by the SED. The SED data
have been interpreted from the viewpoint of dialect areas by Wakelin
(1977, 1983), Fischer (1976), Viereck (1986), Lass (1987)
and
Klemola
(forthcoming a). Glauser has studied the Scottish—English linguistic
border by using evidence he collected specifically for this purpose.
Rohrer (1950), too, collected his own evidence for a study of the border

between the north and northern midlands. Viereck (1980) interprets
Guy Lowman's corpus from the 1930s to establish boundaries in the
247
Ossi Ihalainen
south-east, south-west and East Anglia. The complexity of what is
arguably the most important modern isogloss, Ellis'
sum/soom
line, was
investigated, in the Wash area, in great and revealing detail by Chambers
& Trudgill (1980: 129-37). Their analysis clearly shows how far
abstracted from reality lines drawn on maps can be.
5.8.1
Classificatory criteria
Criteria used in classifications of dialects are mainly phonological, but
Glauser (1974), Fischer (1976) and Viereck (1986) show that, in spite of
occasional reservations by dialectologists, lexical material can be used to
define dialect areas. The emphasis on phonological criteria derives
naturally from the fact that the greatest amount of variation can be
found in phonology, and phonological criteria can be used to
differentiate between quite small areas as against grammatical and
syntactic features that may unite areas showing a great deal of
phonological differentiation. For instance, the area where
be
is used as a
finite
verb,
as in
They
be tired 'They're tired' (south-west and south-
western midlands; see LAE map Ml) the vowel of

five
has at least seven
different realisations, ranging from [fae:v] to [fDiv].
Table 5.1 compares the main criteria used by Ellis, Wakelin and
Trudgill to define English dialect areas. The capital letters indicate the
feature concerned.
5.8.2 Studies of specific dialect boundaries
Three linguistic borders have been extensively studied: the Scottish-
English border (Glauser 1974), the south-western border (Fisher 1976;
Wakelin 1986a) and the border between the northern and north
midland dialects (Rohrer 1950).
The north
The Scottish-English linguistic border was investigated by Glauser
(1974).
On the basis of lexical evidence he concludes that the English
side that used to share features with Lowland Scotland is now
assimilating with northern England, with dialect words receding north.
The political border has thus become a strong linguistic barrier. Glauser
believes that the importance of the geographic boundary as a linguistic
divider will increase in the future.
Rohrer (1950) investigated the border between the northern and
northern midland dialects by asking eighty-three questions of more than
248
The dialects of England since 1776
Table 5.1. Criteria
used to define
English
Feature
sOme
paRK

riNG
hOUse
gAte
lOng
night
mAn
pAst
bAt
Seven
blind
fEW
coffEE
Hill
miLk
the
finite be
periphrastic do
Ellis
1889
X
X
X
X
Wakelin
1983
X
X
X
X
X

X
X
X
X
dialect areas
Trudgill
1990/Traditional
X
X
X
X
X
X
X
X
Trudgill
1990/Modern
X
X
X
X
X
X
(x)
X
a hundred informants from seventy-four villages in Yorkshire. He
found that the border between the north and north midlands runs along
the Wharfe, roughly. This was later confirmed by the SED material
(Hedevind 1967: 38). The differences between the two varieties are
shown in features such as the following:

foal
eat
cow
bone
spoon
Midland
[foil]
[en]
[kau]
[bugn]
[spurn]
The south-west
Northern
[fual]
[iat]
[ku:]
[bian]
[spian]
Depending on the features that they have regarded as significant,
various scholars have defined the south-west slightly differently. Yet
they seem to agree that this is typically an area where people say / ben't
sure instead of Tm not sure, vinger (with a prominent retroflex r) for
249
Ossi Ihalainen
'finger',
Where be em
to?
{Where be mun to?)
for 'Where are they?' and
What's do

that for? tat' Why did you do that?' They use
he
and its object
form
en
to refer to things as well as persons. It might be noted in passing
here that, although phonological, lexical, morphological and syntactic
features seldom co-occur to form clear-cut dialect boundaries, it appears
that the w«g<?r-pronouncing area is more or less identical with the area
where people say
What's do
that for? for 'Why did you do that?'
In his book
The Southwest
of
England
(1986a), Wakelin includes in the
south-west,' with its several sub-varieties', Cornwall, Devon, Somerset,
Dorset, Wiltshire and South Avon. Avon north of Bristol and the
western extremities of Gloucestershire, Berkshire and Hampshire are
regarded as forming a marginal area. This seems to be a conservative
estimate in that some scholars might extend the south-west slightly
further east and north, but the differences are not great (see Viereck
1980;
Wells 1982). Lass (1987: 220-3) defines a set of 'core' south-
western counties. These are: (part of) Cornwall, Devon, Somerset,
Wiltshire, Dorset and Hampshire (as opposed to the south-eastern
counties of
Berkshire,
Surrey, Kent, Sussex and southern East Anglia).

Interestingly, voicing of initial fricatives, the stock linguistic device
of Elizabethan dramatists to mark rusticity, is still found to be an
important dialect feature. However, today it characterises south-western
English rather than southern English in general.
Fischer (1976: 358) analyses lexical evidence provided by the SED
with the aim of establishing the dialect areas in the south-west of
England. The results are summarised as follows:
The South-West
as a
dialect area comprises the region lying southwest
of a line running approximately through Gloucestershire, western
Oxfordshire and Berkshire, and eastern Hampshire. A great many
dialect
words
used everywhere
in
the South-West or
in large parts
of it
confirm its homogeneity and coherence and separate it from other
speech areas further north and
east.
Yet despite this unity the region as
a whole must be subdivided into three smaller areas, namely Area 1
(West Cornwall), Area 2 (East Cornwall and Devon) and Area
3
(the
remaining zone).
Fischer also recognises transitional areas. His '2/3' is a transitional
belt between west Somerset and east Devon, closely resembling Ellis'

district 10. Roughly, the relationships between Ellis' and Fischer's
findings can be seen from the following equations: Ellis 12 = Fischer 1,
Ellis 11 = Fischer 2, Ellis 10 = Fischer 2/3, Ellis 4 = Fischer 3
(Fischer's maps 305 and 308).
250
The dialects of England since 1776
Fischer finds that his area 2 is the
Kernlandscbaft,
the most homo-
geneous area of the whole region. He does not elaborate on this, but it is
perhaps worth pointing out that this area also stands out from the rest of
the south-west on phonological and grammatical grounds: for instance,
it is unique in the south-west in that it uses
us
freely as a subject pronoun,
has
[Y:]
in words like
food,
has a monophthong [ae:] in words like knife,
just to mention the most striking unique characteristics.
A comparison with Ellis' classification shows a striking resemblance
between the late nineteenth and the mid-twentieth centuries (compare
Fischer's maps 305 and 308 in particular). In the actual development of
dialect vocabulary Fischer finds two kinds of recession in the south-
west. There is evidence of a westward recession, something that we have
already seen in phonology and grammar with southern features like the
voicing of initial fricatives, the present-tense -th marker and the
pronoun 'ch 'I' retreating to the south-west. Perhaps somewhat
unexpectedly, Fischer also discovers dialect words receding eastwards,

with older forms being replaced by standard vocabulary in Cornwall and
Devon. Fischer feels the reason for this is the late arrival of English in
his area: dialect is less deeply rooted here than elsewhere in the south-
west and thus more susceptible to influence from standard English. On
the whole the south-west is a retreat area.
5.8.3 General surveys of contemporary dialect areas
Wakelin (1983), on the basis of phonological evidence provided by the
SED, argues for four dialect areas, which appear to be basically those of
the Middle English period. These are the north, the south-west, the west
midlands and the east midlands. The line that separates the north from
the rest of the country is the Humber—Ribble line based on the
pronunciations of the words
cows,
goose,
loaf,
coal, eat,
ground,
blind and
wrong.
Typical pronunciations of these north of the Humber are
[ku:z],
[gias],
[liafj,
[kual],
[iat],
[grund], [blind],
[rag].
The south-west is seen, rather vaguely, as the area west of Watling
Street, a view that Wakelin modifies in his
Southwest

of
England,
as was
seen above. Characteristics of the south-west are rhoticity, voicing of
initial fricatives, lack of w in words like
woman,
and
be
or bin for am.
Wakelin finds periphrastic
do {They
do
go to work)
in a restricted area in the
central south-west, but he believes periphrastic
do
to have occurred in
the area bounded by Watling Street although its distribution is now
limited to the central south-west. Klemola (forthcoming b) has recently
251
Ossi Ihalainen
studied the unpublished incidental material in the SED fieldworkers'
recording-books and found evidence for
a do
area that roughly coincides
with the area where initial fricatives are voiced.
Viereck (1986) searches for bundles of isoglosses based on dialect
words ('heterolexes') in the SED material. He discovers evidence for
the following lexically differentiated (recessive) areas: the north (with
the extreme north-west of England emerging as a separate area),

Lincolnshire, East Anglia, the extreme south-east of England, the
south-west and the west midlands. The Home Counties do not emerge
as a clearly focused area on the basis of lexical evidence, which can be
accounted for by the close affinity to standard English. The solid lines on
map 5.2 indicate heavy bundling of dialect words and can be interpreted
as major dividers. The blank, unbounded areas indicate lexical closeness
to standard English. There is a clear-cut blank in the Home Counties.
One is reminded of Puttenham's statement that the 'best' type of
English was ' the vsual speach of
the
Court, and that of London and the
shires lying about London within lx. myles, and not much above' [The
Arte of
English
Poesie,
1589). Viereck's analysis would thus appear to
support Gorlach's conclusion that 'the geographical basis of good
English has not really changed over the last 400 years' (vol. Ill, of
CHEL, forthcoming).
Trudgill (1990) differentiates between 'Traditional Dialect' and
' Modern Dialect' and sketches the distributional patterns of each in a
way that makes it possible to draw conclusions about recent changes in
English dialect areas. Trudgill's Traditional Dialect is the type of rural,
mid-twentieth-century, working-class English surveyed by the SED
(1962-71). The term Modern Dialect is not clearly defined, but it is used
to cover rural, working-class English today. Trudgill's criteria are
phonological. They are listed below:
Traditional
dialect
1

2
3
4
5
6
7
8
Long:
lang
vs
long
Night:
neet
vs nite
Blind: blinnd vs blined
Land:
lond
vs land
Arm: arrm vs ahm
Hill:
ill vs hill
Seven:
%even
vs
seven
Bat:
bat
vs bxt
Modern
dialect

1
2
3
4
5
6
7
but vs boot
arrm vs ahm
singer
vs
singger
few vsfoo
tense vs lax
ee
in
coffee
gate vs geht
milk vs mioo{t)k
Expressed another way, his key words for Traditional Dialects are
LONG:
lang
vs
long,
i.e. [larj] vs
[lDrj];
NIGHT:
neet
vs
nite,

i.e. [ni:t] vs
252
The dialects of England since 1776
3—5 heterolexes
6-8 heurokxes
9-14
heterokxes
50
100km-
SO milts
Map 5.2 Bundles of heterolexes in England (redrawn from Viereck 1986:
734)
Ossi Ihalainen
Traditional Dialects
Scots
Western Eastern
Central Central
Western
Eastern
Figure 5.2 Trudgill's classification of traditional dialects.
[nait];
BLIND
blinnd
vs
blined,
i.e. [blind] vs [blaind]; LAND
lond
vs
land,
i.e. [lDnd] vs [laend/land];

ARM
arrm vs ahm, i.e. whether the r is
pronounced or not;
HILL
:
///
vs
bill,
i.e. A-deletion; SEVEN
^even
vs
seven,
i.e. voicing of the initial fricative;
BAT
bat
vs bxt, i.e. [bat] vs
[bxt].
What is interesting about these criteria is that Ellis'
some
is not
regarded as a main divider of divisions and there is a definite emphasis
on items that are northern or northern and north midland
(lang,
neet,
blinnd,
lond).
Thus the early perception that the English spoken in the
north is somehow radically different from the English spoken in the rest
of the country is still reflected in Trudgill's classification.
The application of the eight test pronunciations

(LONG, NIGHT, BLIND,
LAND,
ARM, HILL, SEVEN, BAT) to the SED material defines thirteen
varieties of traditional dialect, the most basic distinction being the
division of England linguistically into north and south. These can be
seen from figure 5.2, while table 5.2 shows how the eight test features are
pronounced in these thirteen areas. The linguistic feature used as the
southern boundary of
the
linguistic north is the
/larj/
pronunciation of
long.
The south, the area south of the River Humber, is further divided
into the central and southern dialect areas. Southern dialects as against
central dialects have [x] in bat in the east and are r-pronouncing in the
west. An additional feature that distinguishes between these two
varieties is the pronunciation of the vowel in words likepatb. The vowel
is short in the central dialects and long in the southern dialects.
254
The dialects of England since 1776
Table 5.2. The
pronunciation
of Trudgill's
eight
diagnostic features in the
thirteen Traditional Dialect
areas
(from Trudgill 1990: 32)
Northumberland

Lower North
Lancashire
Staffordshire
South Yorkshire
Lincolnshire
Leicestershire
Western Southwest
Northern Southwest
Eastern Southwest
Southeast
Central East
Eastern Counties
Long
lang
lang
long
long
long
long
long
long
long
long
long
long
long
Night
neet
neet
neet

nite
neet
nite
nite
nite
nite
nite
nite
nite
nite
Blind
blinnd
blinnd
blined
blined
blinnd
blinnd
blined
blined
blined
blined
blined
blined
blined
Land
land
land
lond
lond
land

land
land
land
lond
land
la:nd
lxnd
tend
Arm
arrm
ahm
arrm
ahm
ahm
ahm
ahm
arrm
arrm
arrm
arrm
ahm
ahm
Hill
hill
ill
ill
ill
ill
ill
ill

ill
ill
ill
ill
ill
hill
Seven
seven
seven
seven
seven
seven
seven
seven
zeven
seven
seven
seven
seven
seven
Bat
bat
bat
bat
bat
bat
bat
bat
bat
bat

bat
ba:t
ba:t
baet
Interestingly, using a totally different set of criteria, Trudgill comes up
with major dialect areas that resemble those of
Ellis.
That is, the major
dialect boundaries divide the country into northern, central (that is,
midland) and southern areas. There is a close resemblance between
Trudgill's traditional dialect areas and Ellis' English dialect districts.
Trudgill uses traditional spelling (supplemented by a number of
diacritics) to show pronunciations. One cannot help noticing that many
of these are remarkably similar to the spellings found in early dialect
texts:
for instance, such northern pronunciations as '
the
rang spee-oon'
(the wrong
spoon),
'
a
stee-an hoos' (a
stone house)
could have come from
Meriton's 'Yorkshire Dialogue' (1684).
Trudgill's criteria for classifying Modern Dialects are the following
pronunciations: the vowel in but, r in arm,
ng
in

singer,
ew
in few,
ee
in
coffee,
a ingate
and /in
milk.
These features are incorporated in the test sentence
'Very few cars made it up the long hill'. The diagnostically interesting
realisations of these variables are
[A]
VS
[U]
in
but,
rhoticity, [q] vs [rjg] in
words like j/«g«r,j-dropping in words like few
(i.e.
[fju:] vs [fu:]), tensing
of the
final
vowel in words like
coffee
(i.e.
[kDfi] vs [kDfi:]), monophthong
in gate (i.e. [geit] vs [ge:t]), /I/ in words like milk (i.e. [milk] vs
pronunciations where the /has acquired a short [u]-like vowel in front of
it or is realised as a vowel, as in [miulk], [miok]). As in the case of

traditional dialect, A-retention is also used as a diagnostic, although its
255
Ossi Ihalainen
•NORTHERN
Central c.
North.
)
./••••"\ > '~
Midlands \ CENTRAL/
SOUTH
so too
km.
so miles
Map 5.3 Trudgill's modern dialect areas
256
The dialects of England since 1776
Modern Dialects
Northeast
Figure 5.3 Trudgill's classification of modern dialects.
distribution today appears to be quite restricted, with former h-
pronouncing areas like East Anglia becoming /6-less (Trudgill 1990: 50).
The most basic criterion in Trudgill's classification is the 'but/boot'
distinction. This is, of course, Ellis' 'sum/soom' boundary. One is
reminded of Ellis' prediction that this may turn out to be the most
important contemporary linguistic divider in England. Trudgill's
classification of Modern Dialects can be seen from map 5.3. How the
various Modern Dialects are grouped together is shown by figure 5.3.
The following list shows how Trudgill's test sentence 'Very few cars
made it up the long hill' is realised in sixteen Modern Dialect areas:
Very few

cars made
it up the
long
hill:
Northeast
Central North
Central Lancashire
Humberside
Merseyside
Northwest Midlands
West Midlands
Central Midlands
Northeast Midlands
East Midlands
Upper Southwest
Central Southwest
Lower Southwest
Veree few cahs mehd it oop the long hill
Veri few cahs mehd it oop the long ill
Veri few carrs mehd it oop the longg ill
Veree few cahs mehd it oop the long ill
Veree few cahs mayd it oop the longg ill
Veri few cahs mayd it oop the longg ill
Veree few cahs mayd it oop the longg ill
Veri few cahs mayd it oop the long ill
Veree few cahs mayd it oop the long ill
Veree foo cahs mayd it oop the long ill
Veree few carrs mayd it up the long ill
Veree few carrs mayd it up the long iooll
Veree few carrs mehd it up the long ill

257
Ossi Ihalainen
South Midlands Veree foo cahs mayd it up the long iooll
East Anglia Veree foo cahs mayd it up the long (h)ill
Home Counties Veree few cahs mayd it up the long iooll
(Trudgill 1990: 65-6)
As can be seen from works like Wyld (1956) and Dobson (1968), the
features Trudgill uses for the classification of Modern Dialects may have
long histories in spite of their relatively recent status as diagnostic
features. In what follows the emphasis will be on recent developments
and possible future changes in dialects and dialect features.
Three of the Modern Dialect markers, the final vowel of
coffee,
j-
dropping (i.e. the
dook
realisation of
duke)
and
mehd
for
made,
will be
briefly discussed here;
ngg
for
ng
in words like
strong
was dealt with on

page 217. Because of their wider implications, the remaining
'modern dialect' markers, with some additional features, will be
discussed in sections 5.8.4 and 5.8.5.
The [i:] realisation of the final
ee
in
coffee
- called y-tensing by some
scholars
—
was in fact regarded as a standard feature by Walker. He
points out that words like
vanity
might as well be written
vanitee
(1791:
24).
In early northern dialect texts, on the other hand, the final -y was
often spelt
a (verra
' very',
Jerra '
Jerry',
Sunda'
Sunday'), which not only
suggests that the vowel was lax but also that it was markedly northern.
The SED shows that in the mid-twentieth century the country was
about three-quarters -j-laxing in words like
every
and

ready
(LAE maps
Ph203,
Ph204). In rural dialects tensing appears to have been indigenous
especially in the south-west. Trudgill feels that the tense vowel is now
spreading rapidly into the lax -y territory. It has already reached such
northern urban centres as Liverpool and Newcastle, where it has
jumped over the intervening lax
-y
territory (Trudgill 1990: 77).
Yod-dropping, as in duke /du:k/ was observed and censured by
Walker (1791), who compares it with assibilation:
There is a slight deviation often heard in the pronunciation of this
word [i.e.
duke],
as if written
Dook;
but this borders on vulgarity; the
true sound of the « must be carefully preserved, as if written
Dewk.
There is another impropriety in pronouncing this word, as if written
Jook;
this is
not
so
vulgar
as the
former,
and arises
from

an
ignorance of
the influence of accent.
The spread of the /e:/ pronunciation, Trudgill's 'mehd' and 'geht',
in words like
made
and
gate
in northern English at the expense of the
traditional [ia] can be seen from Tidholm's study of Egton English
258
The dialects of England since 1776
(1979).
Tidholm provides the following figures for his three informant
categories Old, Mid and Young (Tidholm's 'Young' informants were
born after the Second World War):
name, made, etc.
13
e:
ei
Old
12-2
84-8
3-0
Mid
18-6
61-4
20-0
Young
5-4

59-5
35-1
Tidholm
explains the [e:] pronunciation in words like
made
[me:d] as an
approximation
of the standard English sound. Trudgill feels this sound
is
'destined to spread into Humberside, Central Lancashire and the
Lower
South-West in the not too distant
future'
(1990:
78).
5.8.4 Recent developments in dialect areas
Like Ellis (1889), Trudgill uses the pronunciation of r as a main
classificatory criterion.
17
However, there is a difference. While Ellis was
interested in the phonetic realisation of r (whether r was ' reverted' or
not),
Trudgill is concerned with the question of whether varieties of
English are rhotic or non-rhotic, a rhotic dialect being one where r is
pronounced in non-prevocalic positions, as in
car
and park. There was,
of course, a good reason for Ellis not to use rhoticity as such as a
criterion for classifying dialects: rhoticity, at least variable rhoticity,
appears to have been 'accepted' up to the twentieth century. Ellis'

remarks on the pronunciation of
r
suggest a situation that Gimson and
Eustace were able to document as late as 1965. They recorded the speech
of Miss Flora Russell (the niece of the ninth Duke of Bedford), whose
speech they regard as a ' good example of a certain kind of Victorian
English' (Eustace 1969: 34). The transcription of
Miss
Russell's speech
shows that she has an r (phonetically realised as a velar or bunched r)
before non-vowels (i.e. in words like
German)
in six cases out of the
possible twenty-two (Eustace 1969: 73). Eustace regards the r of Miss
Russell's type of English as a weakening of an earlier retroflex r before
its disappearance in pre-nonvocalic contexts. However, the variety
described in Jones (1909) is non-rhotic. Jones characterised this as a
variety used by 'educated people in London and the neighbourhood',
and clearly intended it to provide a pronunciation standard. That
rhoticity was seen as regional about that time is also supported by
evidence from dialect literature. (See, for example, the reference to a
story by John Read on p. 215.)
259
Ossi Ihalainen
Although Ellis was more interested in the phonetics of r than in
rhoticity as such, his transcriptions suggest that the ' reverted'
r
area was
rhotic. Assuming this to be the case, we can compare Ellis' reverted ur
area with Trudgill's southern 'Modern Dialect' r-pronouncing area.

This comparison shows that rhoticity has disappeared from the south-
east of England and shows a south-westerly recess movement. Because
of this rapid retreat, Trudgill concludes that rhoticity is ' unlikely to
survive for longer than a century or so' (1990: 77).
Since Ellis' 'reverted' r area is more or less the r area that emerges
from the SED evidence, it is reasonable to conclude that the recession of
r in the south of England is a post-1950s development. Combined with
the evidence about recent changes in Yorkshire English from Tidholm
(1979),
it would seem that English dialects were rather stable till the
generations born after the Second World War.
Comparing the SED material with his modern evidence, Trudgill
(1990:
76) concludes that the northern limit of rounding of OE
a:,
the
so-called Humber-Ribble line, which separated northern England from
the rest of the country, has disappeared completely. The distinction
between rounded and unrounded reflexes of Old English a: is still an
important dialect boundary, but it has receded northwards and now
divides Scotland from England. Glauser's study of northern vocabulary
shows the same kind of recession into Scotland. The political boundary
between England and Scotland has thus increased its importance as a
linguistic divider, with northern English becoming more English and
more clearly differentiated from Scottish English than before.
The earlier primary linguistic boundary in England, the Humber-
Ribble line, Trudgill argues, has now been replaced by the 'pahst'
vs 'passt' and 'up' vs 'oop' isoglosses, which run, roughly, from the
Wash to the Severn and separate the linguistic south from the linguistic
north (i.e. the north proper and the midlands): 'this is a line which most

English people are very well aware of and which they use informally to
divide "southerners" from "northerners'" (1990: 76).
These two dialect markers, as opposed to the northern failure to
round Old English
a:,
are of relatively recent origin and they show the
modern trend for innovations to spread from the south-east.
In words like past and
dance
the original short sound shows signs of
lengthening towards the end of the seventeenth century, and the
lengthened vowel is further lowered to /a:/ (Dobson 1968, n: section
50).
However, the relative social status of the short and long vowels
appears to have been still unsettled even in the late eighteenth century.
260