See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/249955067

On the reversibility of mergers: /W/, /V/ and evidence from lesser-known

Englishes

Article in Folia Linguistica Historica · January 2003

DOI: 10.1515/flih.2003.24.1-2.23

CITATIONS READS

29 816

1 author:

Peter Trudgill
Université de Fribourg
125 PUBLICATIONS 8,486 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Non-standard language varieties View project

All content following this page was uploaded by Peter Trudgill on 09 April 2015.

The user has requested enhancement of the downloaded file.

 ON THE REVERSIBILITY OF MERGERS:
/W/, /V/ AND EVIDENCE FROM LESSER-KNOWN ENGLISHES1

PETER TRUDGILL – DANIEL SCHREIER – DANIEL LONG – JEFFREY P. WILLIAMS

Conventional linguistic wisdom has it that mergers cannot be reversed: “it is generally
agreed that mergers are irreversible: once a merger, always a merger” (Labov 1994:
311). The reason for this is clear: once two phonemes have converged, speakers have
no way of knowing which one of the two original units belongs in which one of the two
original lexical sets, and restoration is impossible. As is well known, however, there are
a number of reports in the historical linguistics literature of phonological mergers
which have been reversed. One often quoted example is that of the merger in English
of the lexical sets of MATE and MEAT, which is well-attested from earlier periods of the
language, but which is not found in any modern variety of English.
This “once a merger, always a merger” maxim has quite naturally led historical
linguists to consider how to explain these reports of mergers which have been reversed.
In earlier work on this topic, historical linguists (e.g. Kökeritz 1953) typically
employed explanations for this puzzling phenomenon which were based on dialect
contact. They agreed that mergers could not be reversed as such, but their thesis was
that while, say, MATE and MEAT were indeed genuinely merged in some dialects, the
merger was later undone as a result of contact between speakers of these dialects and
speakers of other dialects where it had not occurred. That is, speakers were able to
accurately repair the merger by consulting the distribution of vowels over lexical sets
in the speech of speakers of the non-merging dialects. Wyld (1956: 210) writes that we
have to assume that the MATE and MEAT part of the English vowel system was
“differentiated among different classes of speakers – whether in a Regional or a Class
dialect I am unable at present to say – into two types”, and that the unmerger was not a
sound change as such but “merely the result of the abandonment of one type of
pronunciation and the adoption of another” (1956: 211).
More recently, a brilliant and pioneering alternative explanation has been advanced
by Labov. This is that these mergers were never actually mergers at all but rather
“near-mergers”. That is, they may have been perceived and spelt and reported as
mergers because of a very close phonetic proximity between the two phonemes
concerned. Labov (1994: 349-70) discusses this issue at considerable length. He cites
several instances of speakers being able to produce a very small phonetic distinction

1 We are very grateful to Karen Lavarello-Schreier for her help with Tristanian and St Helenian English,
and to Anders Källgård for his magnanimous and invaluable help with Pitcairnese. We would also like to
thank Walt Wolfram for his information on earlier forms of English in North Carolina.
212

without being able to perceive it. How they do this, however, is, as Labov says (1994:
371), “not at all clear”. These small differences are big enough to be apparent to
investigating linguists but are not observed by speakers themselves. Trudgill (1974)
cites speakers in Norwich in minimal pairs tests claiming a merger of the lexical sets of
NEAR and SQUARE by reading aloud and commenting on, for example, the pair here and
hair as follows: ‘[he¢:], [hE£:] – yes, they’re the same’. Normally this very close
approximation, we have to suppose, represents a stage on the way to a complete merger
– as has indeed subsequently proved to be the case in Norwich (Trudgill 1988).
However, on occasion, the two phonemes, because a total merger has not actually taken
place, can at a later date subsequently move phonetically further apart again, leading to
reports of unmergers as in the case of MEAT and MATE.
All the reversed and therefore, according to the Labovian thesis, near-mergers
discussed in the literature so far have concerned vowels (cf. Labov 1994: 349-90). In this
paper we discuss a well-known but little discussed phonological merger in English which,
however, involves consonants. The merger is of especial interest because, if that is what it
was, it has clearly been totally reversed in the geographical area for which it was reported,
namely the southeast of England. In this paper we examine what is known about this
merger and attempt an examination of the viability, in its case, of the “dialect contact”
versus the “near-merger” theses. We accept that the modern examples of near-mergers
cited by Labov are entirely convincing, and we are very happy to accept this thesis as the
correct explanation for apparent unmergers in many historical cases. Here, however, we
ask whether the dialect-contact thesis may not in other cases be correct also.
The facts concerning this merger as they are generally reported are that in at least
many of the local varieties spoken in the southeast of England in at least the 18th and
19th centuries, prevocalic /v/ in items like village was replaced by /w/. Most reports
focus on word-initial /w/ in items such as village, victuals, vegetables, vermin, and
although many writers do not actually say so, the impression one receives is that [v]
then occurred only in non-prevocalic position i.e. in items such as love, with the
consequence that [w] and [v] were, presumably, in complementary distribution and /w/
and /v/ were no longer distinct.

1. Reports of the merger

Some of the reports of this merger are the following:

1) Ellis (1889) describes the southeast of 19th century England as being, at the
Traditional Dialect level, the “land of wee” (as opposed to /v/).
2) Wright (1905: 227) says that “initial and medial v has become w in
mid-Buckinghamshire, Norfolk, Suffolk, Essex, Kent, east Sussex”. This
geographical configuration might lead one to suppose that it was a feature of
Hertfordshire dialect also.
3) Wakelin (1972: 95-6) writes that the SED materials show that:
 213

“In parts of southern England, notably East Anglia and the south-east, initial
and medial [v] may appear as [w], cf. V.7.19 vinegar, IV.9.4 viper (under
adder), V.8.2 victuals (under food)..... The use of [w] for [v] was a
well-known Cockney feature up to the last century.”
Wakelin (1972: 96) hypothesises that “the area in which [v] > [w] was perhaps
coextensive with the voicing area”. This seems to be wrong, however, since the
voicing area he is referring to here is the area in which voiceless fricatives
became voiced fricatives in initial position. This area most certainly did not
include East Anglia.
4) Wakelin (1984: 79) also says that “Old East Anglian and south-eastern dialect is
noted for its pronunciation of initial /v/ as /w/ in, e.g., vinegar, viper; a very old
feature, which was preserved in Cockney up to the last century”.
5) Further examination of the published SED materials shows other sporadic
instances of this merger. The spontaneous responses to VIII.3.2, for instance,
show very with initial /w/ in Buckland and Coleshill, Buckinghamshire ; and in
Grimston, North Elmham, Ludham, Reedham, and Pulham St Mary, Norfolk.
Many of the other SED instances of /w/ are from reports in which informants
have labelled this pronunciation “older”.
6) Certainly Norfolk was one of the areas in which this merger lasted longest. In a
paper on vestigial dialect variants, Trudgill (1999) discusses the current high
stereotype-level of awareness of this feature in the county even though it has
totally vanished from actual usage. The merger is ‘remembered’ by the local
community decades after its actual disappearance. Most local people in the area
over a certain age ‘know’ that village used to be pronounced willage and that
very used to be pronounced wery. The longevity of this folk memory is rather
remarkable. As a child, Trudgill regularly associated with Traditional Dialect
speakers who were born as early as the 1860s. However, he never heard anyone
use this feature except as a joke or quotation. Discussions with older Norfolk
people suggest that it was in widespread normal unselfconscious use only until
the 1920s. The fact that modern dialect writers still use the feature is therefore
highly noteworthy. For example, Michael Brindred in his local dialect column in
the Norwich-based Eastern Daily Press of August 26th, 1998 writes anniversary
<anniwarsary>. This dialect feature has remained a stereotype for generations
after its disappearance from actual speech.
This Norfolk folk memory concurs with the SED materials and suggests that the
change [v] > [w] indeed took place only in syllable initial position, and that [v] was
still retained in words such as love. We can assume, therefore, that there was, as
suggested above, a single phoneme /w/ which had two allophones, [w] in
syllable-initial and [v] in syllable final position.
A merger of /v/ and /w/ is not too surprising. The functional load of this opposition
in English is rather low. Minimal pairs are very few. However, the merger has clearly
been reversed, as we said above: no native English speaker anywhere in England now
214

fails to contrast /w/ and /v/. The principle concerning the irreversibility of mergers
therefore suggests that we should consider very carefully if the reports of a merger can
in fact be correct. We have to ask: was there a genuine merger which has been reversed
or was it simply a near-merger?
Since the difference between apparent and genuine mergers is a matter of fine
phonetic detail, it would be useful if we could accurately reconstruct the details of the
reported /w/ – /v/ merger. It is impossible now, however, to reassemble the phonetic
details of what happened in England since there are now no native speakers of English
anywhere in the British Isles who have this feature and, if Norfolk is typical, nor have
there been for several decades. Neither is the merger to be found anywhere in the major
colonial varieties of English: American, Canadian, Australian, New Zealand and South
African English do not have it.
However, an avenue of exploration is opened up to us by the fact that the merger is
a feature which reportedly makes an appearance in the phonologies of a large number
of lesser-known colonial varieties of English spoken in small communities in the North
Atlantic, South Atlantic, North Pacific and South Pacific. It is to these varieties
therefore that our discussion now turns. Our thesis is that small isolated communities
may produce slower rates of linguistic change (cf. Trudgill 2001) and that, if this is the
case, these varieties may in some respects be more representative of earlier stages of
English as this was spoken in England than modern English English itself. If we can
show that these varieties genuinely do have a merger, then we can hypothesise that it is
present in these varieties as a result of having been transported to the locations
concerned in the phonologies of speakers from the southeast of England. That is, these
varieties have retained an originally south of England feature which has been lost in its
original homeland. Similarly, if it emerges that these varieties have a near-merger, we
can hypothesise that this, instead, was what was transported from England.

2. Reports of mergers in Lesser Known Englishes

We have located the following reports of the merger in lesser-known varieties of English:

North Atlantic
Bermuda
Ayres (1933: 10) reports a merger of /w/ and /v/ (see below).

Bahamas
Wells (1982: 589) reports for one white Bahamian English speaker “the phonemic
merger of standard /v/ and /w/ into a single phoneme with the allophones [w] and [v] in
complementary distribution. The [w] allophone occurs in initial position...but the [v]
allophone elsewhere”.

Montserrat
Wells (1982: 568) reports village as occurring with initial [w].
 215

St Vincent
Wells (1982: 568) writes that “Vincentians are among those for whom the use of [w]
for standard [v] has been reported”.

Bay Islands
According to Warantz (1983: 84), the phonology of the native English of the Bay
Islands off the coast of Honduras has “the merging of /w/ and /v/ in certain
environments”.

South Atlantic
Tristan da Cunha
Tristan English is reported as having the merger by Zettersten. In his discussion of
consonants he writes: “[v] > [w]: [wEri] ‘very’” (1969: 72).

St Helena
Schreier et al. (in press) write that “phonologically, the most salient characteristics of
basilectal St Helena English include ... the V-W merger”.

North Pacific
Bonin Islands
According to Long (1998; 2000a; 2000b), /v/ and /w/ are merged, and [v] and [w] are
in complementary distribution.

South Pacific
Pitcairn
Ross and Moverley (1964: 154) claim that in Pitcairnese /v/ is generally [w] in
word-initial, word-final and intervocalic position. They say that there is no evidence
that the merger took place in preconsonantal position, where /v/ is realised as [v].
Actually, however, since /w/ does not occur in preconsonantal position in English, this
would in fact appear to be evidence of a merger, since [w] and [v] are in
complementary distribution according to this description.
Norfolk Island
Flint (1964: 196ff.) reports for [w] in valley and invitation.
Palmerston
Ehrhart-Kneher (1996: 530) shows that Palmerston has both [v] and [w]. For words
derived from English/w/, her data show [w] as in ui ‘we’, uan ‘one’, and uash ‘wash’.
For etymonic /v/, there is variation between [w] and [v], apparently in complementary
distribution. In medial and final position, all examples show [v], as in ev ‘have’. In
initial position, however, /v/ consistently becomes [w] as in ueri ‘very’. This seems to
suggest a total merger of original /v/ and /w/.
216

3. The substratum problem

Thus in eleven different varieties of English in widely separated areas of the world, a
genuine merger of /v/ and/w/ has been reported. This would seem on the face of it to
represent good evidence that this same merger did in fact take place in the southeast of
England, whence it was exported to these colonial varieties overseas, but where it has
now disappeared. However, there is an important reason why these reports of the W-V
confusion cannot, without further examination, unambiguously be interpreted – as
colonial remnants of a phenomenon which has disappeared in the mother country – as
evidence for the merger in England. This is that there is an alternative explanation
which could be advanced, namely a substratum effect deriving from the influence of
other languages which these colonial varieties of English may have come into contact
with. Let us now consider these in turn.

The North Atlantic

The Bahamas
The Bahamas consists of an archipelago of about 700 islands to the southeast of
Florida, with a population of about 270,000. Serious English involvement began in
1648 when Bermuda was suffering from religious disputes and Captain William Sayle,
a former governor, decided to look for an island where religious dissidents could
worship. He sailed to the Bahamas with about 70 settlers, Bermudian dissidents and
others who had come directly from England. Their plantation colony was largely
unsuccessful, and a number of settlers returned to Bermuda. Other Bermudian migrants
continued to arrive, however, and the island of New Providence was settled from
Bermuda in 1656. After the American Revolution, from 1782 onwards, many
American Loyalists also fled to the Bahamas, which had the effect of doubling the
white population. The English of white Bahamians, then, has two main sources: the
Bermudian English of the original settlers, and the North American English of the
Loyalists. Some of the Loyalists were from the American South, but Abaco and
northern Eleuthera islands in particular were settled by Americans from New England
and New York (Holm – Shilling 1982). There was also some later white immigration
from the Miskito coast of Central America when this area was ceded by Britain to
Spain in 1786, and Andros island in particular was settled from there. There are
certainly linguistic differences between the different islands of the Bahamas to this day.
It is possible then that the /v/ – /w/ merger, rather than arriving directly from England,
came from Bermuda or from Central America, from where we have also reports of the
merger. Holm, however, argues for a different explanation. A majority of the modern
population of the Bahamas is of African descent. Black Bahamians too have different
origins (cf. Holm 1980), some being descended from slaves who actually arrived in the
Bahamas directly, others being originally from the Caribbean or the American South –
the post 1782 immigration had the effect of trebling the black population since many
Loyalists brought their slaves with them. According to Holm and Shilling (1982: vii),
Black Bahamian English is probably most like the mainland American Creole Gullah.
It is certainly “closer to white English than comparable varieties in the Caribbean
 217

proper, but much further from white English than the vernacular Black English of the
United States”. Holm argues strongly that the Bahamian V-W confusion is the result of
African influence. This is also an origin countenanced for the West Indies as a whole
by Wells, who writes “it is not clear whether this phenomenon ... arose independently
in the West Indies through the influence of an African substratum lacking /v/” (1982:
568). This is a possibility we have to consider since Welmers (1973: 52) states that
typical West African consonant systems have /w/ but no /v/, a point which is supported
by Clements (2000: 125). Ewe, on the other hand, has /w/ as well as /B/ and /v/
(Ladefoged 1968: 25), and also, according to Clements (2000: 127), /¸/.
Holm argues further that the presence of the merger in white speech is the result of
influence from the majority black population.

The Bay Islands

The Bay Islands are a group of 8 small islands about 55 km off the northern Honduras
coast. They were first sighted by Columbus in 1502 and were settled in 1642 by
English buccaneers. Between 1650 and 1850 Spain, Honduras, and England disputed
ownership of the islands. The islands were officially annexed by Britain in 1852 but
were then ceded to Honduras in 1859. White English-speaking Protestants formed the
majority of the population until about 1900, when Hispanic Hondurans from the
mainland began settling, but indigenous anglophones still form about 85% of the
population. The population is currently about 20,000, black, white and mixed. An
African substratum cannot be ruled out here, but seems less likely than in the Bahamas.

Bermuda
Bermuda is a British colony about 900 km east of Cape Hatteras, North Carolina, USA.
The first anglophones to arrive were English Puritans who were shipwrecked in 1609.
In 1612, 60 English settlers were sent to colonise the islands and Bermuda became a
crown colony in 1684. African slaves were transported to Bermuda as early as 1616,
and soon the black population was larger than the white one. Today about 60% of the
population are of African origin; whites are mostly of British origin, but descendants of
Portuguese labourers from Madeira and the Azores who arrived during the 1800s are
also to be found and some Portuguese is still spoken. There are noticeable differences
between the speech of blacks and whites, the former being more Caribbean in
character, the latter more like the English of coastal South Carolina (cf. Trudgill 1986).
Once again, an African substratum has to be considered.
A further complication is that there were a large number of historical connections
between these three different communities. There was considerable to-ing and fro-ing
of anglophones between the Bahamas and Bermuda, as well as between the Bay
Islands of Honduras, the Caymans, and the Bahamas (Parsons 1954). We cannot
exclude the possibility therefore that in any one of these territories, the merger arrived
not from England but from one of the others.
218

The South Atlantic

St Helena
There is a clear substratum problem with the English of St Helena as well. This island
is situated in the South Atlantic Ocean, 1,930 km west of Angola. Its nearest neighbour
is Ascension Island, approximately 1,100 kilometres to the northwest. St Helena’s
population of approximately 6,000 is of mixed European, African and Asian origin.
English is the only language spoken on the island. From the time when it was claimed
by the British East India Company in 1658, a concerted policy of settlement was
implemented, and Company employees (soldiers and servants) and ‘planters’ were
recruited to St Helena, along with slaves supplied on request by EIC ships. Little is
known about the origins of the British settlers but there is some evidence that most of
them came from southern England. Even less is known about the origins of the
non-white population, but various records show that slaves were imported from the
Guinea Coast, the Indian sub-continent and Madagascar, and to a lesser extent from the
Cape, the West Indies, Malaya and the Maldives. St Helenian English, which is perhaps
best described as a creoloid, is thus the result of contacts between regional dialects of
English English and many other languages. An African or other substratum origin for
the V-W confusion is thus possible.

Tristan da Cunha
This British dependent territory consists of six small islands which are about half-way
between southern Africa and South America. The only populated island, Tristan da
Cunha, has an area of about 100 square km and a population of about 290. It is said to
be the most remote permanently inhabited settlement in the world, the nearest
habitation being St Helena, which is about 2,300 km away. The islands were
discovered in 1506 by a Portuguese sailor, Tristao da Cunha. A British garrison was
stationed on Tristan da Cunha in 1816, as a result of fears that it might be used as a
base for an attempt to rescue Napoleon from St Helena (see above) and the islands
were formally annexed by Britain. When the garrison was withdrawn the following
year, three British soldiers, one of them with his wife and children, obtained permission
to stay behind and settle permanently. In the 1820s shipwrecked sailors and castaways
from all parts of the British Isles added to the population, and six women, some of
whom seem to have been freed non-white slaves, immigrated from St Helena in 1827,
one of them with four daughters. In the 1830s and 1840s several US American whalers
arrived, as well as three non-anglophone seamen: a Dutchman and two Danes. The
population increased rapidly and by 1842 the island community consisted of 10
families with 75 people. There were thus three influential groups in the community’s
early formation period. First, the British group, the colony’s founders, consisting of
soldiers, castaways and shipwrecked sailors from the British Isles; second, the women
who arrived from St Helena in 1827; and third, the US American whalers and
European sailors who settled between 1833 and 1849. There was very little trade with
passing ships in the second half of the 19th century and the influx of settlers declined
drastically, the only newcomers being a weaver from Yorkshire in the 1860s, two
 219

stranded sailors from Italy who settled in 1892, and two Irish sisters who arrived in
1908. Despite repeated language contact at various stages of the formation process, the
present-day population is entirely anglophone. Sociohistorically, there were three types
of contact: dialect contact between the British and American dialects spoken by the
anglophone founders; language contact between English and the native tongues of the
non-anglophone settlers (i.e. Dutch, Danish and Italian); and contact with the
English-based creoloid (or creole or pidgin) spoken by the women from St Helena. It is
possible that the V-W merger on Tristan could therefore have an origin in southern
England; and/or in influence from Dutch and/or Danish and/or Italian; and/or in
African influence via the English of St Helena.

The North Pacific

Bonin Islands
These Japanese-owned islands, known in Japanese as Ogasawara-gunto, are in the
central Pacific Ocean, about 800 km southeast of Japan proper. The population is about
2,000. The islands were discovered by the Spanish navigator Ruy Lopez de Villalobos
in 1543. They were claimed by the U.S. in 1823 and by Britain in 1825. They were
formally annexed by Japan in 1876, but after World War II, they were a placed under
U.S. military control, and returned to Japan in 1968. The originally uninhabited islands
were first settled in 1830 by fifteen people: five seamen – two Americans, one
Englishmen, one Dane, and one Italian; and ten Hawaiians, five men and five women.
This founding population was later joined by whalers, shipwrecked sailors, and drifters
of many different origins.
The English of the islands appears to be mainly American in origin and has many
similarities with New England varieties (cf. Long 1998; Long 2000a; 2000b).
Immigration from Japan is currently being followed by language shift to Japanese.
The English phoneme /v/ corresponds to three different variants in Bonin English.
The first variant [v] is distinguished as a separate phoneme and is commonly found
among speakers (middle-aged today) who grew up during the United States Navy’s
administration of the island, attending the Navy school and, in many cases, advancing
on to high school in Guam. This mid-20th century infusion of American English into
the community surely strengthened the usage of [v], but social circumstances in the
later 19th century (mentioned below), indicate that [v] may have been used on the
islands before World War II as well. The second variant, pronounced as [b], represents
a merger of /v/ and /b/. This variant is common among speakers (in their sixties and
above today) whose English was acquired prior to World War II, when native speakers
of Japanese formed the majority of the population. In this era, English was generally
restricted to private situations, whereas Japanese was the language of wider
communication, including formal education. Japanese has no /v/, which typically
becomes [b] in English loanwords in Japanese. The third variant of /v/ is [w], and its
origin is more difficult to pin down to a single specific cause. The variant would appear
to predate [v] and [b] on the islands, and is clearly heard on tapes of the speaker we
refer to as Clarence, who was born on Chichijima in the 1880s. He consistently uses
220

[w] in syllable-initial position and [v] in preconsonantal and syllable-final position, as

per the reports from southeast England.
Substratum influence concerning [w] is difficult to evaluate. Many specifics of 19th
century Bonin English will probably forever remain a mystery; only piecemeal
information exists. We do, however, know the following facts. The early settlers of the
island were a multiethnic band whose native tongues consisted of English, Portuguese,
Hawaiian, Chamorro and at least a dozen other European and Pacific Island languages.
Native speakers of English were a tiny minority in the community and almost all
households consisted of speakers of differing languages. A variety of English was used
as the common language of the tiny community, and it is often described as being
‘broken’. At the same time, however, some islanders had exposure to mainstream
(non-contact) varieties of English through off-island experiences such as attending
English schools on the Japanese mainland. Circumstantial and anecdotal evidence
indicates that a continuum existed in late 19th and early 20th century ranging from a
mainstream variety of English to a local contact variety. Within the context of the
islands’ settlement history, two possibilities present themselves to account for the usage
for /w/-/v/merger. One is that it results from the English spoken by the large number of
Polynesian (and other Oceanic language) speakers living on the island in the 19th
century. Hawaiian has no contrast between /v/ and /w/: according to Elbert and Pukui
(1979: 12-13) /w/ is [w] on Kaua’i and Ni’ihau, [v] on Hawaii, with [v] and [w]
occurring as allophonic variants on the other Hawaiian islands. There is evidence that
this leads to confusion between/v/ and /w/ in the English of Hawaiians and other
Polynesians, but nothing to indicate the kind of complementary distribution found on
the Bonins. The other is of course that it arrived in the phonologies a
native-English-speaking settlers: two early settlers (both males) who exerted a
tremendous degree of influence over the community were native speakers of English
whose home regions we can pinpoint. One was from Wallington, Surrey, England –
crucially in the southeastern merger area; the other from Bradford, Massachusetts in
the United States.

The South Pacific

Pitcairn
Pitcairn is an isolated British colony about 2,200 km southeast of Tahiti in the south
Pacific. The main and only inhabited island has an area of about 5 square km, and the
population in 1992 was 52. The modern population, as is well known, is descended from
the mutineers of the British ship HMS Bounty and their Tahitian companions. After a
lengthy stay on Tahiti, the crew, led by the first mate, Fletcher Christian, mutinied when
their voyage to the West Indies had got only as far as western Polynesia, and set their
captain William Bligh and a number of loyal sailors adrift. They headed back to Tahiti,
where they collected a number of local women and a few men, and from where, fearing
discovery by the Royal Navy, many of them set off again. They reached Pitcairn in 1790,
where, in the interests of secrecy, they burnt their ship. The island community survived
undiscovered until found by American whalers in 1808. It is therefore possible that the
 221

merger in Pitcairn English could be the result of Tahitian influence. There is no contrast
between /v/ and /w/ in Tahitian: Tryon (1970: 2) writes that Tahitian /v/ is “phonetically
[v], as in vine” but that “v is sometimes pronounced w, as a free variant; it is also realised
as [B]”. Even if the Pitcairn merger does have an origin in English, moreover, there is no
guarantee that this was in English English: one of the mutineers, Edward Young, came
from St Kitts in the Caribbean, some areas of which, as we have already noted, are
reported to have the merger themselves.

Norfolk Island
Norfolk Island is an Australian dependent territory in the southwestern Pacific, about
1,600 km northeast of Sydney. The island, with an area of 35 square km, has a
population of about 2,000. In 1856 the Pitcairn Islanders, descendants of the mutineers
on the Bounty, were resettled on Norfolk because of overcrowding on Pitcairn. Not all of
the islanders were happy, however, and eventually two separate groups returned to
Pitcairn. Norfolk island’s current population includes about one third who can claim to
be the descendants of mutineers, the remainder being descendants of later settlers, mostly
from Australia and New Zealand. Since Norfolk Island English has its origins on
Pitcairn, the same problems in ascribing an origin in England to this phenomenon occur.

Palmerston
Palmerston English is a variety spoken on Palmerston Island, Polynesian Avarau, a
coral atoll in the Cook Islands about 430 kilometres northwest of Rarotonga, by
descendants of Cook Island Maori and English speakers. What we know about the
settlement is that William Marsters, a ship’s carpenter and cooper from
Gloucestershire, England, came to uninhabited Palmerston Atoll in 1862. He had three
wives, all from Penrhyn/Tongareva in the Northern Cook Islands. He forced his wives,
17 children and numerous grandchildren to use English all the time. Virtually the entire
population of the island today descends from the patriarch. Ehrhart-Kneher (1996)
considers Palmerston English to be a dialectal variety of English rather than a contact
language. She writes that it appears to be a classic case of mixing and vernacularization
of a type which has “produced languages which, while new languages, are varieties of
English rather than new languages without genetic affiliation in the usual sense”. While
her analysis focuses on syntax and we cannot ignore her disclaimer that “the
transcription used here makes no phonological claims” (Ehrhart-Kneher 1996: 524),
she is nonetheless a trained linguist and has based her transcriptions on the IPA
alphabet. Once again, the substratum problem is evident since Penrhyn/Tongarevan has
/v/, but no /w/ (Clark 1976 : 20, quoting Yasuda, 1968).

4. Summary

It would be possible, then, to argue that these Lesser Known Englishes provide no
evidence as to the nature of the /w/-/v/ merger in England. In every case, it is possible
to argue that the merger was not imported from England at all but is the result of direct
or indirect interference from languages other than English which had no /w/ – /v/
222

distinction. After all, only a very few of the world’s languages have this distinction. An
examination of Maddieson (1985) shows that 76% of his sample languages have /w/,
but only 21% have /v/, and a mere 11% percent of the languages in his sample have
both /v/ and /w/.
Our own feeling, however, is that the presence of this merger in so many different
varieties of English in so many different parts of the world – as we shall see, we have
discovered 17 such varieties so far – is too much of a coincidence to be totally
explicable in terms of a substratum effect in all cases, and that it is therefore indicative
of an earlier merger in England. There is considerable evidence, we would maintain, of
there being a merger of /v/ and /w/ in pre-19th-century English in the southeast of
England. If we are correct in this, then all the reports that we have cited suggest that it
was a merger that was subsequently reversed, and not a near-merger.

5. The two-way transfer pattern

We now move on to another important piece of evidence which we have so far not
discussed. This evidence concerns what we can call an apparent two-way transfer
pattern. That is, there are a number of reports from England which seem to indicate not
a merger of /w/ and /v/ as a result of a sound change /v/ > /w/, but rather that two
different changes occurred: /w/ > /v/; and /v/ > /w/. The evidence is as follows:

1) Wyld (1956: 292) writes of the interchange of v- and w-:

“This was formerly a London vulgarism, but is now apparently extinct in the
Cockney dialect. Personally, I never heard these pronunciations, so well known
to the readers of Dickens, Thackeray, and of the earlier numbers of Punch. My
time for observing such points begins in the late seventies or early eighties of the
last [i.e. 19th] century, and I never remember noticing this particular feature in
actual genuine speech, though I remember quite well, as a boy, hearing
middle-aged people say weal for veal and vich for which, jocularly, as though in
imitation of some actual type of speech with which they were familiar.”
He then goes on to suppose that the pronunciation was extant until the 1840s or
1850s. Crucially, he also reports documentary evidence, some of it as early as
the 15th century, of a two-way transfer pattern, with examples of the same writer
using v for w and w for v, and cites instances such as vyne ‘wine’, and vyves
“wives’, as well as wyce ‘vice’ and woyce ‘voice’, suggesting a complex
interchange of the two phonemes. This early date is supported by Wakelin
(1972: 96), who writes “this change goes back to ME and is evidenced in
place-name spellings from much of southern England from the second half of
the thirteenth century onwards”. Wyld also argues (1956: 180) that the merger
was “probably not confined to London”.
2) Wyld also cites (1956: 179) the Scot James Elphinston, born 1721, as referring
to the two-way transfer. And he tells us (1956: 182) that the elocutionist John
Walker, born in London in 1732, discussed the V-W merger as occurring
“among the inhabitants of London, and those not only of the lowest order” and
 223

as operating in both directions: vind = ‘wind’ and weal = ‘veal’. Similarly,

Thomas Sheridan (1719-88) discusses veal > weal as well as winter > vinter.
3) This two-way transfer is also reflected, as Wyld mentions, in the stereotypical
19th century Cockney portrayed, famously, in the speech of Dickens’ character
Sam Weller in Pickwick Papers. Dickens writes, for example, <wery> for very,
but also <vith> for with:
I had a reg’lar new fit o’ clothes that mornin’, gen’l’men of the jury, said
Sam, and that was a wery partickler and uncommon circumstance vith me in
those days.......If they wos a pair o’ patent double million magnifyin’ gas
microscopes of hextra power, p’raps I might be able to see through a flight o’
stairs and a deal door; but bein’ only eyes, you see, my wision’s limited
(Charles Dickens, The Pickwick Papers (1837), chapter 34, emphasis added).
4) W. Matthews, in his book Cockney Past and Present, also cites the v-w
interchange as being typical of older Cockney (1972: 180-1).
5) The SED gives a few examples of w > v as well as the examples of v > w
already cited. Wakelin (1972: 96) cites watch from Somerset and wife from Kent
with initial [v].
6) Wells (1982: 568) refers to the “V-W Confusion of eighteenth- and nineteenth-
century London Cockney”.
7) There is also evidence from the lesser-known overseas varieties of English.
Holm claims for the Bahamas that “[w] and [v] “alternate freely” (1980: 56).
8) Washabaugh (1983) claims that in the Caymans Islands, a location we have not
previously mentioned, “/v/ and /w/ are sometimes used in the reverse of their
English reflexes, e.g., vejiz (wages) and inwestigeyt (investigate)” (1983: 178).
Kohlman (n.d.) agrees: “v and w, which in many words can scarcely be
differentiated, and in other words are interchanged. Hence we find ‘vessel’ is
‘wessel’, ‘virgin’ is ‘wirgin’, and ‘wood’ is ‘vood’(p. 13)”.

A two-way transfer /v/ > /w/ and /w/ > /v/ is a very mysterious change from a
historical linguistic point of view. It is a phenomenon which most historical
phonologists would consider extremely unlikely if not totally impossible. Two
simultaneous changes /w/ > /v/ and /v/ > /w/ must surely be out of the question. So why
does Dickens show both v for w and w for v? And why do we find so many other
reports suggesting the same thing?

6. Explanations

We can conceive of three different possible explanations:

1) There was a near-merger involving a distinction which Dickens did not hear, and
which led the writers cited by Wyld to employ interchangeable spellings. We
cite evidence below for why we do not believe this is what happened.
224

2) A single change /v/ > /w/ did genuinely take place. Occasional forms such as
vyves are the result of spasmodic hypercorrection leading to non-systematic
substitutions in the opposite direction /w/ > /v/ also. This point of view is
supported by Wakelin: “it may be assumed from the statements of various writers
that early New English spellings of v for etymological w reveal hypercorrect
pronunciations”. However, forms such as vyves are in Dickens not occasional. The
consistency with which Dickens converts /v/ to <w> and /w/ to <v> suggests
something much more regular and widespread than hypercorrection. Wyld, too,
gives numerous examples of substitutions in both directions.
3) A complete merger did indeed take place but that it was on some articulation
intermediate between [w] and [v]. It is a principle of phonological perception
that listeners notice what is different about accents other than their own, not
what is the same. Listeners also normally perceive segments which are alien to
their own variety in terms of segments which are native to it. This would have
led Dickens, who presumably did not have the merger in his own accent, and
who was presumably doing his best to report what he thought he heard, to
illustrate Weller as speaking in this highly improbable manner. Dickens would
have heard this intermediate articulation in very as ‘not [v] and therefore
necessarily [w]’ and in with as ‘not [w] and therefore necessarily [v]’.

The question then is: if there was an intermediate articulation, what was it? An
obvious candidate for an articulation intermediate between [v] and [w] is [B], which
combines the bilabial place of articulation of [w] with the voiced fricative manner of
articulation of [v]. This possibility occurred to Wakelin, who speculates that the merger
in England “may have taken place via a bilabial stage /B/” (1972: 96). For the
lesser-known Englishes which we are hoping to derive insights from, Wells (1982:
568) says of the Caribbean that a bilabial fricative ‘has been reported’. The concrete
reports of which we are aware are the following:

1) Ayres (1933: 10) says that in Bermuda there is “an intermediate sound like
Middle German “w” – it is usually a frictional labial sound”.
2) Kohlman (n.d.) writes for the Caymans Islands of “a soft blurring that is neither
‘v’ nor ‘w’”.
3) For the Bahamas, Shilling (1980) writes that “all white speakers ... used [B]
variably with “correct” v and w”. Wells, however, disputes this finding: The
essence of the admittedly frequent V-W Confusion is not ... the use of [B]
indifferently for both /v/ and /w/.
4) Turner (1949: 25) states that the pronunciation corresponding to both English /v/
and /w/ in Gullah – another variety we have not cited before – is /B/. McDavid
and McDavid (1951: 28) also tell us that “many white folk informants in and
near Gullah country replace both /v, w/ by one bilabial voiced spirant /B/”.
 225

Here again, there are problems of substratum effect as opposed to inheritance from
England. The Cayman Islands, for example, are a British colony of three major islands in
the Caribbean, about 290 km northwest of Jamaica. The population is about 25,000. About
a quarter of the Caymanians are European, mostly of British origin; about one quarter are
descendants of African slaves; and the remainder are of mixed ancestry. The Islands were
not claimed by any nation until they were ceded to England in 1670. Most of the early
settlers were British mariners, buccaneers, shipwrecked passengers, plus land-grant holders
from Jamaica, and African slaves. An African substratum therefore has to be considered
for the /v/-/w/ merger. The Gullah /B/ pronunciation is also the basis for the claim,
mentioned above, made by Holm (1980) for an African substratum effect in the
Bahamas. He notes (1980: 56) that Turner points out “that /B/ occurs in many West
African languages”.
At least one West African language, however, as we have already noted, has /w/, /v/
and /B/. To complicate matters further, Ladefoged also states that “the I.P.A. does not
provide for the symbolisation of the contrast between a labial velar approximant like
the English w, and a similar sound with closer articulation which may produce audible
friction” (1968: 25) which also occurs in some West African languages.
We show here that there is reason to believe that these speculations and reports
about an intermediate pronunciation are on the right lines, and indeed for Gullah at
least we have no reason to suppose that the report is anything other than phonetically
totally correct. We suggest, however, that, while the Dickensian representations and the
other reports of a two-way substitution in England can indeed be explained in terms of
the merger being on an articulation intermediate between [v] and [w], and while we
argue that these constitute evidence to suggest that the reports of a total merger in
England were indeed entirely correct, the articulation in question in southeastern
England may not have been – or may not only have been – [B].
This assertion is based again on data from Lesser-Known Englishes. In this case,
however, it is based on our own analyses rather than second-hand reports. These are
phonetic and phonological analysis carried out by Trudgill of six different
lesser-known varieties of English, three of which we have already discussed, which all
still have a genuine merger of historical /w/ and /v/ on precisely the same articulation.
This articulation, however, is not a voiced bilabial fricative but rather a bilabial
approximant. That is, it is an articulation which bears the same relationship to [B] as [¨]
does to [z], and [j] to [Æ], and [å] to [F]. There is no single IPA alphabet symbol for
this consonant so, just as one has to write [D¢] if one wants to distinguish the Danish
dental approximate from the English [D], so one has to use here – in order to
distinguish it from [B] – the symbol [B¢ ]. Laver (1994: 302) writes:
Apart from the symbols specified above, all other central approximants can
be transcribed by adding a subscript diacritic [ ¢] to the corresponding fricative
symbol, meaning ‘more open stricture’. The articulation involves no audible
friction, no lip rounding (except before rounded vowels), and there is no
approximation of the tongue towards the velum as there is for [w].
The six varieties are that we have analysed are:
226

1) Pitcairnese English, as investigated by Källgård (1993), who very kindly made

his tape-recordings, made on Pitcairn in 1980, available to us. Our analyses
imply that the transcriptions of Ross and Moverly are incorrect. Pitcairnese
speakers consistently use [B¢ ].
2) Tristan da Cunha, as investigated by Schreier (in press), where at least some
speakers have this articulation. That is, the analyses and transcriptions of
Zettersten are not totally correct.
3) St Helena, as reported in Schreier et al. (2003), and based on recordings kindly
made available to us by Karen Lavarello.
4) The English of the White community on Bequia, in the Caribbean Grenadines,
as investigated by Williams (1987). Bequia, the largest and northernmost of the
Vincentian Grenadines, is situated in the Windward Antilles approximately 14
kilometres from St Vincent. Bequia has an area of 18 square kilometres and a
population of around 7,200. Little has been written about the settlement history
of Bequia and virtually nothing has appeared on the history of white settlement
(however cf. Williams 1987). There are two separate white communities on
Bequia: Sugar Hill and Mt. Pleasant. Some of the whites in these two
communities derive from relocation schemes entered into by the governments of
Barbados and St Vincent beginning in the 1850s and lasting up until 1880. Like
most other white communities in the Caribbean, white Bequerians have tended
to remain in isolation and adhere to rules of colour endogamy.
5) The English of the White community on Saba, in the Dutch Antilles, as
investigated by Williams (1985, 1987). Saba is the smallest of the Windward
Netherlands Antilles, having a total land area of less than 13 square kilometres.
The population of Saba is approximately 1,600 although about 400 of these are
foreign medical students. Even though Saba has been a colonial possession of
the Netherlands intermittently since the mid-seventeenth century, varieties of
English have been stable vernaculars of communication among local Sabans
throughout the recorded history of the island (cf. Hartog 1988, Johnson 1989,
Williams 1985, 1987, in press). Saba has the highest ratio of whites to blacks of
any of the English-speaking West Indies. The earliest white settlement on Saba
was at the historic village of Mary’s Point that was relocated by the Dutch
government in 1933 (cf. Williams in press). Presently, the largest white
populations are to be found in the villages of Upper Hell’s Gate, Lower Hell’s
Gate, and Windwardside. All of these communities are descended from early
English-speaking settlers who came from St Kitts – many of these Presbyterian
Scots – and also Barbados – many of these escaped, or freed indentured servants
–-during the middle of the seventeenth century. The whites of Saba have
maintained separateness from the blacks on the island, with most blacks taking
up residence in the village of The Bottom following emancipation in 1863
(Johnson 1989: 10).
 227

6) Particularly valuable has been recent fieldwork on the basilectal English of the
white community on Anguilla by Williams and Trudgill. Anguilla is a British
dependency in the northern Leeward Antilles, approximately 235 kilometres east
of Puerto Rico. The most recent population census of Anguilla lists nearly
12,000 inhabitants. The majority of the islanders are classified as either black or
brown (mixed), with only a handful of ‘clear-skinned’ (white) people. The white
population of Anguilla, presently concentrated in the community of Island
Harbour, had two likely historical sources. One was the settlement of English
from St Kitts who began arriving on Anguilla around 1650 to collect salt and
grow tobacco (Burns 1954: 350). Some of these early settlers founded the
historical village of Sandy Hill whose only remaining feature is a cemetery. The
other source of the present-day white population of Anguilla was the shipwreck
of the English brigantine “Antelope” that was wrecked off the small island
called Scrub Island, only a few kilometres northeast of Island Harbour village, in
1771. The ship was sailing from Grenada to England when it ran onto the reef.
A trio of brothers and their wives, the Websters, survived the wreck and made
their way to the main island of Anguilla and established the community of
Island Harbour. Although the white population of Anguilla has had contact with
the black and mixed populations of the island, a pattern of colour endogamy has
prevailed up until recently.
This recent fieldwork of ours has revealed the same pronunciation, [B¢], but there is
the added bonus that, since this phonetic detail was now one of the objects of our
research, it was possible for us to witness, as well as just hear from tapes, that no
lip-rounding occurs in the pronunciation of the (identical) consonants of will and
village, except before the two rounded vowels /u:/ and /u/. (Laver (1994: 297-8)
describes [w] as a voiced labial velar approximant which has a “rounded lip position”.)
To cite just one of our informants: Mr Elbert Webster, born 1919, of Island Harbour,
Anguilla, employs this articulation throughout in items such as: away, well, Webster,
where, visit, vex, voice, everyone, television, over, love.
Our observations are also supported by Wolfram and Thomas (2002) who report on
an isolated area of coastal North Carolina, another location we have so far not
mentioned – in work we were not aware of until we had completed our research in
Anguilla – saying that “one of the noteworthy patterns apparently found in earlier
Pamlico Sound English is the merger of /v/ and /w/..... The LAMSAS Hyde County
interview with the speaker born in 1858 indicates a number of instances where [B¢] is
transcribed for both /v/ and /w/.” This feature is no longer found in the area.
It is of course vital that we should state unequivocally that we are entirely confident
that we are not hearing a near-merger here. We have had to consider this possibility not
only because of Labov’s very persuasive work but also because Kohlman’s (n.d.)
description of colloquial Caymanian English contains the phrase “the ‘v’s’ in such
words as ‘invited’, ‘several’, ‘have’, and the ‘w’s’ in ‘worship’, ‘work’, and ‘wife’, are
pronounced the same, or almost the same” [our emphasis]. The consonants involved in
the speech of our Anguillian informants are visibly and audibly identical as between
the two lexical sets.
228

An African substratum is less likely as a source for this pronunciation in the speech
of white than black Caribbean speakers, though we concede that it is possible.
(According to Ladefoged (1968: 25) there are complex relationships between several
different labial fricatives and approximants in many West African languages, with some
languages lacking both /v/ and /w/ but having instead a voiced bilabial or labiodental
approximant). We also have to consider the possibility of an Irish Gaelic substratum.
Many of the servants who were transported to the Caribbean from Ireland in the 17th
century were speakers of Irish, and Irish lacks /w/ (cf. O Dochartaigh 1984: 298).
However, especially bearing in mind our observations on Tristan, St Helena and
Pitcairn English, it is particularly crucial to note that this bilabial approximant
articulation is one which is extraordinarily rare in the languages of the world: Maddieson
(1984: 96) shows that only 1.9% of his sample languages have this consonant. This rarity
strongly supports our proposal that the best way in which we can explain the coincidence
of the presence of such an unusual articulation in the six varieties we have analysed is by
supposing that all or most of them inherited it from England.
We therefore conclude that these different communities, scattered as they are in
different widely separated parts of the world, and sharing an unusual articulation found
in only a very small percentage of the world’s languages, inherited this articulation
from a common source, namely the dialectal English of the southeast of England.

7. The role of dialect contact in restoration

It looks, then, as if there were two different types of merger in southern England. In
one, the merger was on [w] and [v] as allophones of a single phoneme. In the other,
there was a merger of historical /w/ and /v/ on [B¢ ] or [B] or both. How do we account
for this rather puzzling situation?
Given that there is reason to believe that there was a merger in the south of
England, we have to conclude that there was also a genuine reversal of the merger. And
we therefore have to turn, as an explanation for the undoing of this merger, to the
dialect contact hypothesis. There is indeed, happily, evidence that dialect contact is
exactly the correct explanation of how the W-V confusion was ‘repaired’. Reports that
we mentioned above from some of the Lesser Known Englishes, other than the seven
varieties we have just mentioned, show evidence of a partial reversal in progress. We
interpret this as evidence that the merger did occur but is currently in the process of
being reversed. This process has only just begun in Montserrat, for instance, where
Wells (1982: 568) says that the merger “is restricted to the speech of the older
uneducated population, and even for them lexically restricted, since vote has [v], not
[w]”. For Norfolk Island, Flint (1964: 196ff.) reports, as we have seen, [w] in valley
and invitation; he also, however, reports [v] in devil (1964: 208), implying that the
merger – which we can suppose from the evidence of Pitcairnese was formerly total –
has been reversed and that the /w/ for /v/ pronunciation has become very lexically
restricted. Even more importantly, our younger Anguillian informants lack the merger
also: what is happening in Anguilla now, we would argue, is what started happening in
the southeast of England 150 years ago.
 229

We therefore suppose the following. There was indeed an early genuine merger in
southeastern England of /v/ and /w/ on [B¢ ] or [B]. This merger on an articulation
intermediate between [v] and [w] led listeners who did not have it to report, and in the
case of Dickens, to portray /v/ for /w/ and vice versa. This merger was carried, perhaps
in the 17th century, to other parts of the world, mainly the early colonies such as those
of the Caribbean, in some of which it still remains. In southeastern England, on the
other hand it was reversed, as a result of contact with middle-class accents and accents
from further north and west in England which did not have the merger.
Again there is evidence from lesser-known Englishes as to how this took place:
what happened in the south of England is probably what – we can infer from Wells’
observations – is beginning to happen in Montserrat. There the contact-induced
de-merger process has not yet led to a total re-establishment of two separate lexical sets
but it has led to the replacement of the low status form [B¢ ] by the acrolectal [w] and [v]
articulations, which however continue for the moment to be widely used in members of
the “wrong” lexical sets. We can suppose that in England, too, the merger was reversed
by means of an intermediate stage in which the mainstream English forms [w] and [v]
were used instead of [B¢] but, for a while, allophonically rather than contrastively. This
later, chronologically intermediate system is the one which is illustrated in the SED
records and the subject of the Norfolk ‘folk memory’, and the one which was also
exported, probably in the 18th century, to later colonial Englishes, such as that of the
Bonin Islands.
As we said above, we are not at all here opposing the near-merger hypothesis as
such. We believe that the evidence put forward by Labov is utterly persuasive, and
indeed one of us has produced evidence of one case of a near-merger in his own work
(Trudgill 1974). Our conclusion, however, is that, while the near-merger thesis may be
valid in many cases, it is probable that in this particular case, and therefore perhaps in
others also, this is not the correct historical analysis. The consonants /v/ and /w/ were
genuinely merged on [B¢ ] and then unmerged again in parts of southern England, via an
intermediate stage in which [w] and [v] were allophones of the same phoneme. This
merger does not survive in any of the major varieties of English around the world. But
it has survived, due to isolation from the influence of mainstream varieties which did
not have the merger, as well as to small community size and consequent slow rate of
linguistic change, in a large number of lesser-known Englishes. Evidence from these
little-studied varieties suggests rather strongly that “once a merger” is not necessarily
always “always a merger”.

REFERENCES

Aceto, Michael – Jeffrey P. Williams (eds.)

in press Eastern Caribbean Englishes. Amsterdam: John Benjamins.
Ammon, Ulrich – Norbert Dittmar – Klaus Mattheier – Peter Trudgill (eds.)
in press Sociolinguistics/Soziolinguistik. Berlin: de Gruyter.
230

Ayres, Harry Morgan

1933 “Bermudian English”, American Speech 8/1: 3-10.
Burns, Sir Alan
1954 History of the British West Indies. London: George Allen and Unwin.
Chambers, Jack K. – Natalie Schilling-Estes – Peter Trudgill (eds.)
2001 Handbook of linguistic variation and change. Oxford: Blackwell.
Clark, Ross
1976 Aspects of Polynesian syntax. Auckland: Linguistic Society of New Zealand.
Clements, George N.
2000 “Phonology”, in: Bernd Heine – Derek Nurse (eds.), 123-160.
Day, Rita (ed.)
1980 Issues in English creoles. Heidelberg: Groos.
Elbert, Samuel H. – Mary Kawena Pukui
1979 Hawaiian grammar. Honolulu: University of Hawaii Press.
Ellis, Alexander J.
1889 On early English pronunciation, part V: the existing phonology of English dia-
lects. London: EETS.
Ehrhart-Kneher, Sabine
1996 “Palmerston English”, in: Stephen A. Wurm – Peter Mühlhäusler – Darrell T.
Tryon (eds.), 523-531.
Fischer, Steven – Wilfried B. Sperlich (eds.)
2000 Leo Pasifika: Proceedings of the Fourth International Conference on Oceanic
Linguistics.
Flint, Elwyn H.
1964 “The language of Norfolk Island”, in: A.S.C. Ross – A. Moverley (eds),
189-206.
Hartog, John
1988 History of Saba. Saba: O.K.S.N.A.
Heine, Bernd – Derek Nurse (eds.)
2000 African languages: An introduction. Cambridge: Cambridge University Press.
Holm, John
1980 “African features in White Bahamian English”, English World-Wide 1: 45-66.
Holm, John (ed.)
1983 Central American English. Amsterdam: Benjamins.
Holm, John – Alison Shilling
1982 Dictionary of Bahamian English. Cold Spring NY: Lexik House
Johnson, W.
1989 Saban lore: Tales from my grandmother’s pipe (3rd revised edition.) Saba:
Lynne Johnson.
Källgård, Anders
1993 “Present-day Pitcairnese”, English World-Wide 14: 71-114.
 231

Kohlman, Aarona Booker

no date Wotcha Say: An introduction to colloquial Caymanian. Grand Cayman: Cayman
ARTventures.
Kökeritz, Helge
1953 Shakespeare’s pronunciation. New Haven: Yale University Press.
Labov, William
1994 Principles of linguistic change. Vol. 1 – Internal factors. Oxford: Blackwell.
Ladefoged, Peter
1968 A phonetic study of West African languages. Cambridge: Cambridge University Press.
Laver, John
1994 Principles of phonetics. Cambridge: Cambridge University Press.
Long, Daniel
2000a “Evidence of an English contact language in 19th century Bonin Islands”, Eng-
lish World-Wide 20: 251-286.
2000b “Examining the Bonin (Ogasawara) Islands within the contexts of Pacific lan-
guage contact”, in: Steven Fischer – Wilfried B. Sperlich (eds.), 200-217.
Long, Daniel (ed.)
1998 The linguistic culture of the Ogasawara Islands. Japanese Language Centre Re-
search Reports 6. Osaka: Shoin Women’s College.
Maddieson, Ian
1985 Patterns of sounds. Cambridge: Cambridge University Press.
Matthews, William
1972 Cockney past and present. London: Routledge and Kegan Paul
McDavid, Raven I. – Virginia McDavid
1951 “The relationship of the speech of American Negroes to the speech of whites”,
American speech 26 [Reprinted in: Walt Wolfram – Nona Clark (eds.).]
Ó Dochartaigh, Cathair
1984 “Irish”, in: Peter Trudgill (ed.), 289-305.
Parsons, James
1954 “English-speaking settlement of the Western Caribbean”, Yearbook of the Asso-
ciation of Pacific Coast Geographers 16: 2-16.
Ross, Alan S.C. – A. Moverley
1964 The Pitcairnese Language. London: Andre Deutsch.
Schreier, Daniel
2003 Isolation and language change: Sociohistorical and contemporary evidence
from Tristan da Cunha English. Basingstoke: Palgrave/Macmillan.
Schreier, Daniel – Andrea Sudbury – Sheila Wilson
in press “English in the South Atlantic”, in: Ulrich Ammon – Norbert Dittmar – Klaus
Mattheier – Peter Trudgill (eds.).
Shilling, Alison
1980 “Bahamian English – a non-continuum?”, in: Rita Day (ed.), 133-146.
232

Trudgill, Peter
1974 The social differentiation of English in Norwich. Cambridge: Cambridge Uni-
versity Press.
1986 Dialects in contact. Oxford: Basil Blackwell.
1999 “New-dialect formation and dedialectalization: embryonic and vestigial vari-
ants”, Journal of English Linguistics 27/4: 319-327.
2001 “Linguistic and social typology”, in: Jack K. Chambers – Natalie Schil-
ling-Estes – Peter Trudgill (eds.), 707-728.
Peter Trudgill (ed.)
1984 Language in the British Isles. Cambridge: Cambridge University Press.
Tryon Darrell T.
1970 Conversational Tahitian: An introduction to the Tahitian language of French
Polynesia. Berkeley: University of California Press.
Turner, Lorenzo
1949 Africanisms in the Gullah dialect. Chicago: University of Chicago Press.
Wakelin, Martin F.
1972 English dialects: An introduction. London: Athlone Press.
1984 “Rural dialects in England”, in: Peter Trudgill (ed.), 70-93.
Warantz, Elissa
1983 “The Bay Islands English of Honduras”, in: John Holm (ed.), 71-94.
Washabaugh, William
1983 “Creoles of the off-shore islands”, in: John Holm (ed.), 157-180.
Wells, John C.
1982 Accents of English. (3 vols.) Cambridge: Cambridge University Press.
Welmers, William E.
1973 African language structures. Berkeley: University of California Press.
Williams, Jeffrey P.
1985 “Preliminaries to the study of the dialects of white West Indian English”,
Nieuwe West-Indische Gids (New West Indian Guide) 59: 27-44.
1987 Anglo-Caribbean English: a study of its sociolinguistic history and the develop-
ment of its aspectual markers. [PhD dissertation, The University of Texas at
Austin.]
in press “White Saban English: a social history”, in: Michael Aceto – Jeffrey P. Williams
(eds.).
Wolfram, Walt – Erik Thomas
2002 The development of African American English. Malden – Oxford: Blackwell.
Wolfram, Walt – Nona Clark (eds.)
1971 Black-white speech relationships. Washington: Center for Applied Linguistics.
Wright, Joseph
1905 The English dialect grammar. Oxford: Froude.
 233

Wurm, Stephen A. – Peter Mühlhäusler – Darrell T. Tryon (eds.)

1996 Atlas of languages of intercultural communication in the Pacific, Asia, and the
Americas. Berlin: Mouton de Gruyter.
Wyld, Henry C.
1956 A history of modern colloquial English. (3rd edition.) Oxford: Basil Blackwell.
Yasuda, Ayako
1968 The structure of the Penrhyn phrase. [MA thesis, University of Hawaii.]
Zettersten, Arne
1969 The English of Tristan da Cunha (Lund Studies in English 37.) Lund: Gleerup.

View publication stats