‘Proper’ pro-nun-ſha- ſhun in eighteenth-century English: ECEP as a new tool for the study of historical phonology and dialectology
This is a repository copy of ‘Proper’ pro-nun-ſha- ſhun in eighteenth-century English: ECEP
as a new tool for the study of historical phonology and dialectology.
White Rose Research Online URL for this paper:
http://eprints.whiterose.ac.uk/110307/
Version: Accepted Version
Article:
Yanez-Bouza, N., Beal, J., Sen, R. orcid.org/0000-0002-1558-7938 et al. (1 more author)
(2017) ‘Proper’ pro-nun-ſha- ſhun in eighteenth-century English: ECEP as a new tool for
the study of historical phonology and dialectology. Digital Scholarship in the Humanities.
ISSN 2055-7671
https://doi.org/10.1093/llc/fqx003
This is a pre-copyedited, author-produced version of an article accepted for publication in
Digital Scholarship in the Humanities following peer review. The version of record
Yáñez-Bouza, N. et al; ‘Proper’ pro-nun- ha- hun in Eighteenth-Century English: ECEP as ʃ ʃ
a New Tool for the Study of Historical Phonology and Dialectology, Digital Scholarship in
the Humanities, fqx003 is available online at: https://doi.org/10.1093/llc/fqx003
eprints@whiterose.ac.uk
https://eprints.whiterose.ac.uk/
Reuse
Items deposited in White Rose Research Online are protected by copyright, with all rights reserved unless
indicated otherwise. They may be downloaded and/or printed for private study, or other acts as permitted by
national copyright laws. The publisher or other rights holders may allow further reproduction and re-use of
the full text version. This is indicated by the licence information on the White Rose Research Online record
for the item.
Takedown
If you consider content in White Rose Research Online to be in breach of UK law, please notify us by
emailing eprints@whiterose.ac.uk including the URL of the record and the reason for the withdrawal request.
mailto:eprints@whiterose.ac.uk
https://eprints.whiterose.ac.uk/
╅Proper╆ pro-nun-ſha- ſhun1 in Eighteenth-century
English: ECEP as a New Tool for the Study of
Historical Phonology and Dialectology
1. Introduction
In recent years English historical linguists have voiced complaints about the scholarly neglect
of the Late Modern English period (1700に1900). While grammar and the prescriptive
grammatical tradition have received increasing attention over the last couple of decades (Beal
et al., 2008; Tieken-Boon van Ostade, 2008), there is still relatively little research on the
phonology of Late Modern English, and of the eighteenth century in particular; as Beal (1999:
13) points out, け[w]here interest is shown in the eighteenth century, phonology is neglected,
and where interest is shown in the history of English phonology, the eighteenth century is
neglectedげ. There remains an urgent need for new studies of historical phonology in general
and of eighteenth-century phonology in particular. One reason for this lack of research could
be the idiosyncratic notation systems used by eighteenth-century authors, which make it
difficult to search and interpret phonological evidence. Yet the value of pronouncing
dictionaries as rich and reliable evidence of lexical diffusion as well as of sound variation and
change in eighteenth-century pronunciation has been observed in studies such as Beal (1999)
and Jones (2006).
With this in mind, we have constructed a new electronic, searchable database of
Eighteenth-Century English Phonology (ECEP). The purpose of this paper is to present ECEP as a
tool to facilitate research on the social, regional and lexical distribution of phonological
variants in eighteenth-century English, thereby meeting the demands of the growing research
community in Late Modern English generally (Mugglestone, 2003; Hickey, 2010) and in
historical phonology in particular (Honeybone and Salmons, 2015). ʍ;ニキミェ WWノノゲげ ふヱΓΒヲ: 127に
68) lexical sets for comparing the vowel systems of present-day varieties of English as its
reference, the database provides unicode IPA transcriptions for the relevant segment of each
word given as an example of its lexical set or subset in WWノノゲげ account of standard lexical sets,
as documented in eighteenth-century pronouncing dictionaries (e.gく ʍエラマ;ゲ SエWヴキS;ミげゲ A
General Dictionary of the English Language, 1780). WWノノゲげ ゲ┌HゲWデゲ ヮヴラ┗キSW キマヮラヴデ;ミデ ヮラキミデゲ ラa
comparison where varieties of English differ as to the distribution of variants within the lexical
set. For example, whilst the lexical set BATH キゲ SWaキミWS ;ゲ けIラマヮヴキゲキミェ デエラゲW ┘ラヴSゲ ┘エラゲW
Iキデ;デキラミ aラヴマ Iラミデ;キミゲ デエW ゲデヴWゲゲWS ┗ラ┘Wノ っFっ キミ ɑWミAマが H┌デ っ嵯衾っ キミ ‘Pげ ふWells 1982: 133),
subset (59Hぶ Iラミゲキゲデゲ ラa ┘ラヴSゲ デエ;デ ;ヴW ゲラマWデキマWゲ ヮヴラミラ┌ミIWS ┘キデエ っFっ けキミ ;IIWミデゲ ┘エキIエ
otherwise have broad BATH (Australian, West Indian)げ (1982: 134に5), whilst those in subset
(59c) けデ┞ヮキI;ノノ┞ エ;┗W デエW PALM vowel in the otherwise flat-BATH ;IIWミデゲ ラa デエW ミラヴデエ ラa Eミェノ;ミSげ
(1982: 135). We have retained the subsets in the structure of ECEP in order to determine
whether these major distinctions between accent systems in present-day English have
precursors in the variation present across our eighteenth-century data sources. We have
included all of the examples provided by Wells which occur in these sources in order to reveal
;ミ┞ a┌ヴデエWヴ ヮ;デデWヴミゲ ラa ノW┝キI;ノ SキゲデヴキH┌デキラミく SキミIW WWノノゲげ ノW┝キI;ノ ゲWデゲ ;ヴW SWゲキェミWS ラミノ┞ デラ
investigate vowel systems, and some consonantal changes are also of interest in the study of
Late Modern English phonology, we have constructed consonantal sets of phonological
interest and extracted ヴWノW┗;ミデ S;デ; キミ デエW ゲ;マW ┘;┞ ;ゲ aラヴ WWノノゲげ ┗ラ┘Wノ ゲWデゲく
1
SエWヴキS;ミげゲ ふヱΑΒヰぶ デヴ;ミゲIヴキヮデキラミ ふSキ;IヴキデキIゲ ラマキデデWSぶ aラヴ デエW ┘ラヴS けヮヴラミ┌ミIキ;デキラミげが ゲエラ┘キミェ ヮ;ノ;デ;ノキ┣WS
っ崎っ キミゲデW;S ラa っゲキっ ;デ デエW ゲデ;ヴデ ラa デエW デエキヴS ;ミS aラ┌ヴデエ ゲ┞ノノ;HノWゲく
This paper describes the structure and contents of the ECEP database, and reports on the
three-step method of compilation: (i) the selection of primary sources (Section 2); (ii) the
process of data input and annotation (Section 3); and (iii) the design of the web-based
interface (Section 4). The context for the development of this new tool is provided in Section 2,
which gives an overview of the phonology of eighteenth-century English and of the value of
pronouncing dictionaries as evidence for eighteenth-century pronunciation. Section 5
describes two pilot case studies which demonstrate the value of ECEP for the study of English
historical phonology.
2. Background
2.1 The Phonology of Eighteenth-century English
Since Charles Jones described the eighteenth and nineteenth centuries as the けCキミSWヴWノノ;ゲ ラa
English hisデラヴキI;ノ ノキミェ┌キゲデキI ゲデ┌S┞げ ふヱΓΒΓぎ ヲΑ9), there has been considerable progress in Late
Modern English
2
studies. Much of this progress has been made possible by the availability of
corpora such as ARCHER (A Representative Corpus of Historical English Registers), which have
enabled searches across large datasets for the complex patterns of variation and change which
characterize this period. However, despite the monographs by Beal (1999) and Jones (2006),
research on the phonology of this period has been less prolific than that in other areas such as
morphosyntax, pragmatics and language ideology. One reason for this relative neglect of
eighteenth-century phonology is the lack of accessible primary source material: as argued by
Beal (2012b), the corpus revolution which has energised other areas of Late Modern English
studies has so far had little effect on phonology. The ECEP project aims to redress this.
Some scholars have actually suggested that the phonology of eighteenth- and nineteenth-
century English is not worthy of their attention. Bloomfield and Newmark (1963: 288), for
example, state that changes in the language between the eighteenth century and the present
day are けdue to matters of style and rhetoric [...] rather than to differences in phonology,
grammar or vocabularyげ. Bloomfield and Newmark go on to state that historical linguists are
less interested in style and rhetoric, a statement which no longer rings true given the recent
development of historical pragmatics and historical sociolinguistics. Strang notes デエ;デ けsome
short histories of English give the impression that change in pronunciation stopped dead in the
eighteenth c[entury], a development which would be quite inexplicable for a language in
everyday useげ ふヱΓΑヰぎ ΑΒぶく MacMahon (1998), after summarising the views of earlier scholars
who claimed that there had been little change in this period, states that けデエWヴW is other
evidence to show that the pronunciation of English more than 150 years ago was noticeably
different, for reasons mainly of phonotactics (structure and lexical incidence) from what it is
デラS;┞げ (1998: 374, original italics). Whilst both Strang and McMahon assert that changes in
pronunciation have taken place since 1700, both make the point that these more recent
changes are less systemic than those occurring in earlier periods. Beal argues that this
opposition is to some extent け;ミ キノノ┌ゲキラミ IヴW;デWS by the different types of evidence available
aラヴ デエW W;ヴノキWヴ ;ミS ノ;デWヴ ヮWヴキラSゲげ ;ミS ェラWゲ ラミ デラ キミ┗ラニW デエW ゲ;┞キミェ けI;ミげデ ゲWW デエW ┘ララS aラヴ デエW
デヴWWゲげ: in her opinion, け[i]t is a matter of perspective: at a distance, a forest appears as a
monolithic block, but, the closer you get to the forest, the more you notice the variation
HWデ┘WWミ キミSキ┗キS┌;ノ デヴWWゲげ ふBeal, 2004: 125). Not only are we closer in time to the eighteenth
century than to the Middle or Early Modern English periods, but the amount of detailed
2
L;デW MラSWヴミ Eミェノキゲエ キゲ ェWミWヴ;ノノ┞ ;ェヴWWS デラ Iラ┗Wヴ ┘エ;デ エキゲデラヴキ;ミゲ ┘ラ┌ノS デWヴマ デエW けノラミェげ WキェエデWWミデエ ;ミS
nineteenth centuries. See Beal (2004; 2012a), Tieken-Boon van Ostade (2009) for more detailed
definitions.
information on the pronunciation of more recent English makes us more aware of the range of
variation. Moreover, some of the changes occurring in this period are still ongoing and/or are
reflected in variation between varieties of English today. Examples of these changes are the
SキゲデヴキH┌デキラミ ラa けノラミェげ っ嵯准っが っ査准っ ;ミS けゲエラヴデげ っ;っが っ左っ variants in the BATH and CLOTH lexical sets
respectively (see Beal and Condorelli, ヲヰヱヴ aラヴ ;ミ ;IIラ┌ミデ ラa デエW ノ;デデWヴぶき デエW けNラヴデエ-South
Sキ┗キSWげ キミ デエW ;HゲWミIW ラr presence of the phoneme /朔/ (Beal, 2012c); and the ongoing
palatalization of alveolar consonants preceding earlier /ju准/ (see Section 5 below).
Since many of the phonological changes taking place in the eighteenth century involve
shifts in lexical incidence, sources of evidence used to investigate these changes need to be
lexically rich. As we shall see in the next section, the sources chosen for inclusion in ECEP are
ideal for these purposes, as they provide evidence for the entire lexicon.
2.2 Phonology Sources in ECEP
Written evidence for the historical pronunciation of English can be divided into direct and
indirect types. Evidence that is indirect involves sources whose authors were not overtly
commenting on or describing pronunciation, but which give clues about it. Typical sources of
indirect evidence are rhymes, puns and non-standard spellings. Direct evidence, on the other
hand, comes from authors who deliberately set out to describe (or prescribe) the
pronunciation of their time. In reconstructing the pronunciation of earlier periods of English,
we have to rely mainly on indirect evidence, but from the sixteenth century onwards, direct
evidence becomes increasingly available as interest in spelling reform and in phonetics
increases. Texts such as Christopher CララヮWヴげゲ (1687) The English Teacher provide detailed and
sophisticated descriptions of the sounds of English, lists of homophones and near-
homophones and even metalinguistic comments on the social and/or geographic distribution
of variants, but exemplify their descriptions with a very restricted number of lexical tokens.
However, from the middle of the eighteenth century dictionaries are published in which the
pronunciation of every word is described, and these provide the source material for ECEP.
To illusデヴ;デW デエW ケ┌;ミデキデ;デキ┗W SキaaWヴWミIW HWデ┘WWミ ラヴデエラWヮキゲデキI ┘ラヴニゲ ゲ┌Iエ ;ゲ CララヮWヴげゲ
(1687) and eighteenth-century pronouncing dictionaries, let us consider the distribution of
long and short variants of ME 鍋 in the BATH and START sets. Cooper provides important early
evidence for the lengthening of the vowel in these sets, which allows us to identify some of the
ヮエラミWデキI Wミ┗キヴラミマWミデゲ キミ ┘エキIエ デエW Iエ;ミェW aキヴゲデ ラII┌ヴゲく Fキヴゲデ エW デWノノゲ ┌ゲ ラa けデエW ┗ラ┘Wノ ; ノキミェ┌;ノげ
that けキミ デエWゲW can, pass by, a is short; in cast, past, for passedが キデ キゲ ノラミェげ ふヱヶΒΑぎ ヴぶく ʍエWミ エW ェラWゲ
ラミ デラ SキゲI┌ゲゲ デエW IラミデW┝デゲ キミ ┘エキIエ デエW ┗ラ┘Wノ けキゲ ヮヴラミラ┌ミIWS ノラミェ キミ キデゲ ラ┘ミ ゲラ┌ミSげ ふデエ;デ キゲが
/a准っ rather than /e准っぶが デエWゲW HWキミェ けHWaラヴW nch and s when another Consonant follows, and
before r unless sh folノラ┘ゲげ ふヱヶΒΑぎ ンヴぶく CララヮWヴ ヮヴラ┗キSWゲ ; ノキゲデ ラa ┘ラヴSゲ キミ ┘エキIエ デエキゲ ノWミェデエWミWS
a occurs: barge, blast, carking, carp, cast, dart, flasket, gasp, grant, lance, mask, path, tart.
These words have been chosen to provide the same pre-vocalic environments as words which
W┝Wマヮノキa┞ け; ゲエラヴデげ ふっ;っぶ ;ミS け; ゲノWミSWヴげ (/e准/): thus bar with short /a/ is contrasted with barge
pronounced with /a准/ and bare with /e准っく Fヴラマ CララヮWヴげゲ W┗キSWミIW ┘W I;ミ ヮキWIW デラェWデエWヴ ;ミ
account of the environments in which early lengthening occurs, but we have no way of
knowing whether the examples chosen represent all the words in which orthographic
occurs in the given environments. For instance, Cooper provides path as an example of a word
with /a准/, but does not specify whether the same vowel would be used in other words with this
post-vocalic environment, such as bath, lath, etc. By contrast, ECEP contains 127 words from
the BATH set and 28 words from the START set. This will allow users to trace variation between
/a/and /a准/ across a much larger subset of the lexicon and to identify differences in the
transcriptions of authors from different places writing at different times within the eighteenth
century (see Beal, 1999: 105–18 for further discussion of this sound change).
The qualitative value of evidence from eighteenth-century pronouncing dictionaries has
been disputed in the past. Dobson, albeit writing at a time when many of the sources used in
the Eighteenth Century Collections Online
3
were unknown or inaccessible, staデWS デエ;デ けデエW
eighteenth century produced no writers to compare with the spelling reformers who are our
main source up to 1644 (Hodges) or with the phoneticians who, beginning with Robinson
ふヱヶヱΑぶ I;ヴヴ┞ ┌ゲ ラミ aヴラマ ヱヶヵン ふW;ノノキゲぶ デラ ヱヶΒΑ ふCララヮWヴげゲ English Teacher)げ (Dobson, 1957: 311).
Others have taken issue with the prescriptivism of eighteenth-century authors. John Walker,
the most successful and influential of these, is often singled out for criticism on this account.
Sheldon (1947: 146) ┘ヴキデWゲ デエ;デ けW;ノニWヴ ゲ;デキゲaキWゲ デエW デWマヮWヴ ラa エキゲ デキマW ぷぐへ ;ミS キデゲ SWマ;ミS aラヴ
ノキミェ┌キゲデキI ヴWェ┌ノ;デキラミ ;ミS ヴWaラヴマげが ┘エキノゲデ HラノマHWヴェ (1964: 41) accuses Walker of being
けキミaノ┌WミIWS H┞ デエW ゲヮWノノキミェげく Iデ キゲ デヴ┌W デエ;デ ;ノノ デエW ヮヴラミラ┌ミIキミェ SキIデキラミ;ヴキWゲ ┌ゲWS aラヴ ECEP ┘WヴW
written with the aim of providing their readers with a guide to what the authors considered
けIラヴヴWIデげ ヮヴラミ┌ミIキ;デキラミが H┌デ デエW ゲ;マW Iラ┌ノS HW ゲ;キS ラa デエW マ;ミ┞ デ┘WミデキWデエ- and twenty-first
century dictionaries which transcribe the pronunciation of English words in RP and/or General
American. Recent scholars such as Agha (2003), Beal (2003), Ranson (2012) and Trapateau
(2016ぶ エ;┗W ヴWエ;Hキノキデ;デWS W;ノニWヴげゲ ヴWヮ┌デ;デキラミ ;ゲ ; ヮエラミWデキIキ;ミ H┞ デ;ニキミェ エキゲ ┘ラヴニ ラミ キデゲ ラ┘ミ
terms as an important and highly informative source of information on the prestigious
マWデヴラヮラノキデ;ミ ヮヴラミ┌ミIキ;デキラミ ┘エキIエ ┘;ゲ デエW ヮヴWI┌ヴゲラヴ ラa ‘Pく W;ノニWヴげゲ (1791) Critical
Pronouncing Dictionary is the major source of metalinguistic comments in ECEP, many of which
provide valuable sociolinguistic information (see Section 3.2 below for further discussion of
metalinguistic comments). Other sources used in the compilation of ECEP provide accounts of
┘エ;デ ┘;ゲ IラミゲキSWヴWS けIラヴヴWIデげ ヮヴラミ┌ミIキ;デキラミ キミ デエW ヮヴラ┗キミIWゲく
The sources included in ECEP include the earliest available editions of all the accessible
pronouncing dictionaries of English printed in eighteenth-century Britain.
4
At the time of
writing, these are as follows:
Buchanan (1757) Linguae Britannicae Vera Pronuntiatio.
Johnston (1764) A Pronouncing and Spelling Dictionary.
Kenrick (1773) A New Dictionary of the English Language.
Perry (1775) The Royal Standard English Dictionary.
Spence (1775) The Grand Repository of the English Language.
Sheridan (1780) A General Dictionary of the English Language.
Burn (1786) A Pronouncing Dictionary of the English Language.
Scott (1786) A New Spelling, Pronouncing and Explanatory Dictionary of the English
Language.
Walker (1791) A Critical Pronouncing Dictionary and Expositor of the English Language.
Jones (1797, 1798) Sheridan Improved. A General Pronouncing and Explanatory
Dictionary of the English Language. 2nd and 3rd editions.
Buchanan (1757) is the first true pronouncing dictionary of English, in the sense that every
word is transcribed. It was decided to include two editions of JonWゲげゲ SキIデキラミ;ヴ┞ HWI;┌ゲW デエW
third edition demonstrates significant changes in which Jones distances himself from Sheridan,
most noticeably in recognising a distinction between long and short vowels in the BATH and
START sets. In future, we intend to augment ECEP with data from later editions and from other
sources, but those listed above provide evidence across the second half of the eighteenth
century from authors of varying geographical provenance に one Irishman (Sheridan), four
Scotsmen (Buchanan, Perry, Burn, Scott), one northern author from Newcastle (Spence), three
authors from the London area (Kenrick, Jones, Walker), and one author of uncertain origin but
3
http://gale.cengage.co.uk/product-highlights/history/eighteenth-century-collections-online.aspx
4
We intend to include early American pronouncing dictionaries in later versions of ECEP.
who lived and worked in the south-east county of Kent (Johnston). It is important to state that
ECEP is not intended to be a database of dialectal pronunciation, but it does reflect the
┗;ヴキ;デキラミ HWデ┘WWミ デエW けヴWIWキ┗WSげ ゲヮWWIエ ラa LラミSラミ ;ミS ラa デエW Wケ┌キ┗;ノWミデ キミ ヮヴラ┗キミIキ;ノ IWミデヴWゲ
such as Edinburgh and Newcastle, as well as providing evidence for change over the course of
the eighteenth century.
3. Data Annotation
Once the pronouncing dictionaries had been selected, the next step in the compilation of ECEP
was the process of data input and annotation. This section reports on the design and contents
of the database, including the methodological principles adopted.
3.1 Database Design
ECEP has been built in MS Access format as a relational database constructed with a variety of
integrated tables. The data have been systematically annotated and thematically grouped in
three major categories: phonology data, source metadata and author metadata. Details for
each category are set out in Table 1.
Table 1 Design of the ECEP database
(Meta)Data Fields
Phonology Lexical set, Lexical subset, Keyword, IPA, IPA variants, Example word
frequency, Metalinguistic comments, Metalinguistic attitude,
MWデ;ノキミェ┌キゲデキI ノ;HWノが CラマヮキノWヴゲげ ミラデWゲ
Source Type of work, Title, Year of publication (of the edition consulted), Edition,
Place of publication, Imprint (printers, booksellers), Price, Physical
description, Paratext, Audience (age, gender, social class, instruction,
ゲヮWIキaキI ヮ┌ヴヮラゲWぶが ‘WaWヴWミIWゲ Iラミゲ┌ノデWSが CラマヮキノWヴゲげ ミラデWゲ
Author Name, Life dates, Gender, Social class, Place of birth, Places of residence,
Occupation, Other biographical details, Works by this author in ECEP
The metadata for the dictionaries have been drawn from the original sources, such as the
title-pages and prefaces to works, and also from the literature (e.g. Alston, 1966; Beal, 1999).
The metadata for the authors come principally from the Oxford Dictionary of National
Biography.
Regarding the phonology data, the starting point for drawing up the list of words for ECEP
was John WWノノゲげ ふヱΓΒヲぶ Accents of English, in particular his list of Standard Lexical Sets for the
vowel system in varieties of present-day English (1982: 119に20, 127に68). Our aim was for ECEP
to incorporate data from the selected pronouncing dictionaries in the form of IPA
transcriptions so that the historical data documented in the database could be easily
compared to present-day studies; this was necessary because, as mentioned above, the
notation systems used by eighteenth-century authors were often idiosyncratic and difficult to
interpret (see Section 3.2 and Appendix III). The use of WWノノゲげ lexical sets and their associated
example words is standard practice in studies of variation and change in present-day English.
Including the full range of example words allows for differences in lexical distribution between
the primary sources, and also between these and the contemporary accents described by
Wells. For instance, a scholar interested in the distribution of words related to the STRUT-FOOT
split would be able to find how each of the words provided as examples for WWノノゲげ sets is
transcribed in each of the eighteenth-century sources documented in the database, and how
phonological variants are perceived at the time in the context of the standardization of English
(e.g. correct, vulgar, improper, etc.).
Wells (1982: 119にヲヰぶ W┝ヮノ;キミゲ デエ;デ けぷデへエW ┌ゲW ラa ラミW ┗ラ┘Wノ ラヴ ;ミラデエWヴ キミ ヮ;ヴデキI┌ノ;ヴ ┘ラヴSゲ
(lexical items) can be illustrated H┞ デ;H┌ノ;デキミェ デエWキヴ ラII┌ヴヴWミIWげ キミ デエW ゲWデ ラa keywords
presented in Table 2 in small caps, so that each of them けゲデ;ミSゲ aラヴ ; ノ;ヴェW ミ┌マHWヴ ラa ┘ラヴSゲ
┘エキIエ HWエ;┗W デエW ゲ;マW ┘;┞ キミ ヴWゲヮWIデ ラa デエW キミIキSWミIW ラa ┗ラ┘Wノゲ キミ SキaaWヴWミデ ;IIWミデゲげ; the
latter are referred to in this paper as example words. Overall his list contains twenty-four
lexical sets for stressed vowels and three sets for unstressed vowels; this makes 1,737 example
words in total, distributed in sixty-one subsets. The sets KIT, DRESS, TRAP, LOT, STRUT, FOOT, CLOTH,
concern short vowels; the sets BATH, NURSE, FLEECE, PALM, THOUGHT, GOOSE, START, NORTH, FORCE
refer to long vowels;
5
the sets FACE, GOAT, PRICE, CHOICE, MOUTH, NEAR, SQUARE, CURE include
diphthongs; and the sets happY, lettER, commA represent unstressed vowels.
6
Table 2 WWノノゲげ ふヱΓΒヲぎ 127に68) lexical sets in ECEP (sorted as in Wells)
SET SUBSET EXAMPLE WORD
Short Vowels
KIT -- bit, drink
DRESS -- bed, deaf
TRAP -- back, thank
LOT -- box, sock
STRUT -- blood, cut
FOOT -- bush, full
Long Vowels and Diphthongs
BATH BATH_a ask, castle
BATH_b branch, enhance
BATH_c banana, calf
BATH_f blasphemy, plastic
CLOTH CLOTH_a broth, cough
CLOTH_b coffee_1, moth
CLOTH_c coroner, florin
NURSE -- birth, nerve
FLEECE FLEECE_a agree, cheese
FLEECE_b bead, deceive
FLEECE_c machine, police
FACE FACE_a age, safe
FACE_b day, faith
FACE_c break, great
PALM PALM_a calm, father
PALM_b bravado , inamorato
PALM_f almond, sultana
THOUGHT THOUGHT_a fall, sought
THOUGHT_b false, fault
5
The sets are categorized as long- or short-vowel sets according to their pronunciation in RP.
6
Pヴ;IデキI;ノ ミラデWゲく ふ;ぶ ʍエW IラSWゲ ぱ;が ぱH WデIく キミ WWノノゲげ ノW┝キI;ノ ゲ┌HゲWデゲ ;ヴW ヮヴWゲWヴ┗WS ;ゲ キミ エキゲ ラヴキェキミ;ノ ノキゲデが
W┝IWヮデ aラヴ ぱaが ┘エキIエ WWノノゲ IラSWゲ ┘キデエ ;ミ ;ヮラゲデヴラヮエW ;ミS ラaデWミ ヴWaWヴゲ デラ ;ゲ ;ミ け;ヮヮWミSキ┝げ デラ デエW ラヴキェキミ;ノ
set. (b) The codes _1 and _2 in some of Weノノゲげ W┝;マヮノW ┘ラヴSゲ ;ヴW ┌ゲWS ┘エWミ デエW ゲ;マW ┘ラヴS ;ヮヮW;ヴゲ キミ
more than one subset; the number indicates the syllable that is relevant in each particular case, as in
coffee_1 for CLOTH_B and coffee_2 for happY_b.
GOAT GOAT_a boat, holy
GOAT_b grow, know
GOOSE GOOSE_a choose, shoot
GOOSE_b blue, few
PRICE PRICE_a arrive, try
PRICE_b fight, high
CHOICE CHOICE_a boy, noise
CHOICE_b join, spoil
CHOICE_c groin, hoist
MOUTH -- down, mountain
NEAR NEAR_a beer, near
NEAR_b beard, fierce
NEAR_c hero, period
NEAR_f idea, real
SQUARE SQUARE_a air, pear
SQUARE_b scarce
SQUARE_c dairy, rarity
START START_a far, start
START_b bark, party
START_c tiara
NORTH NORTH_a for, war
NORTH_b assort, mortal
NORTH_c aura, Taurus
FORCE FORCE_a adore, door
FORCE_bi deport, forth
FORCE_bii coarse, fourth
FORCE_c aurora, glorious
CURE CURE_ai amour, tour
CURE_aii endure_vw, pure
CURE_b gourd, tournament
CURE_ci boorish
CURE_cii bureau, curious
Weak Vowels
happY happY_a baby, city
happY_b coffee_2, vanity
lettER -- better, razor
commA -- opera, saliva
To these sets for the study of the vowel system in general we have added five
supplementary sets for the study of the consonant system in eighteenth-century English,
including ten subsets and a total of 204 example words.
7
The sets DEUCE, FEATURE and SURE
address the process of palatalization, dealing with stress patterns (subsets _a for stressed
syllable, _b for post-stress syllable, _c for pre-stress syllable), and the pre-/j/ phoneme (/t, d, s,
z/ in each set). The set HEIR relates to the presence or absence of initial /h/, and the set WHALE
to the pronunciation ラa け┘エげく See Table 3 and Appendix I for details.8
7
More consonant sets may be added in due course.
8
Practical notes. (a) When the same example word appears in a vowel set and in a consonant set, the
former is coded with _vw and the latter with_cn, for instance heir_vw for SQUARE_a and heir_cn for HEIR.
Table 3 Consonant lexical sets in ECEP
SET SUBSET EXAMPLE WORD
DEUCE DEUCE_a /t/ Tuesday
/d/ due
/s/ suit
/z/ resume
DEUCE_b /t/ altitude
/d/ module
/s/ issue
/z/ visual
DEUCE_c /t/ tumultuous
/d/ adulation
/s/ superior
/z/ --
FEATURE
9
-- /t/ creature
/d/ procedure
/s/ pressure_cn
/z/ pleasure
SURE SURE_a /t/ mature
/d/ during_cn
/s/ surety
/z/ c(a)esura_cn
SURE_b /t/ century
/d/ verdure
/s/ censure
/z/ closure
SURE_c /t/ maturation
/d/ duration
/s/ mensuration
/z/ --
HEIR -- honour, humble
WHALE WHALE_a when, whine
WHALE_b elsewhere, somewhat
Each of the eighteenth-century pronouncing dictionaries in ECEP was examined in order to find
WWノノゲげ example words for vowels and consonants, and the data were entered according to the
following principles:
(b) If the same example word appears in more than one of the consonant sets, _deu stands for DEUCE,
_ture for FEATURE, and _sure for SURE, for instance fissure_ture and fissure_sure.
9
This set consists of words which have schwa in the post-stress syllable in present-day English according
to the Oxford English Dictionary (as opposed to a full vowel in SURE_b), following a palatalized
consonant in at least one pronunciation variant. These forms presumably arose from pronunciations
┘キデエ っテ鎖っが ┘エキIエ ;ヮヮW;ヴ デラ エ;┗W HWIラマW マラヴW ┘キSWゲヮヴW;S キミ デエW WキェエデWWミデエ Ientury. These in turn
ラヴキェキミ;デWS キミ aラヴマゲ ┘キデエ ┗;ヴキ;デキラミ HWデ┘WWミ ぷ┞准へ ;ミS ぷキ┌へ キミ デエW aキミ;ノ ゲ┞ノノ;HノW キミ MキSSノW Eミェノキゲエく WエWミ デエW
aキミ;ノ ゲ┞ノノ;HノW HWI;マW ┌ミゲデヴWゲゲWSが デエWヴW ┘;ゲ ┗;ヴキ;デキラミ HWデ┘WWミ けa┌ノノげ aラヴマゲ ┘キデエ っキ┌っ ;ミS ヴWS┌IWS aラヴマゲ
┘キデエ っ畷っく ʍエW ┗;ヴキ;ミデゲ ┘ith /iu/ could then develop to /ju/ with the subsequent possibility of palatalizing
デエW ヮヴWIWSキミェ Iラミゲラミ;ミデが ┘エWヴW;ゲ デエラゲW ┘キデエ っ畷っ SキS ミラデ ノW;S デラ ヮ;ノ;デ;ノキ┣;デキラミく S┌HゲWケ┌Wミデ ヴWゲデラヴ;デキラミ
of /j/ in the schwa-forms (with possible palatalization) combined with the reduction of /u/ to schwa in
デエW a┌ノノ aラヴマゲ ヴWゲ┌ノデゲ キミ デエW ヴWマ;ヴニ;HノW ┗;ヴキ;デキラミ ┘W ゲWW キミ ECEP HWデ┘WWミ Wくェく っデテ┌准っが 三┌准っが っ三テ┌准っが っデテ鎖っが
っ三鎖っが っ三テ鎖っが ;ミS っデ畷っ キミ デエキゲ ゲWデ.
a) WWノノゲげ ノW┝キI;ノ ゲWデゲ ;ヴW SWゲキェミWS aラヴ デエW ;ミ;ノ┞ゲキゲ ラa ヮヴWゲWミデ-day English. Naturally, the sets
include example words that were introduced into the English language in recent times.
Given that the scope of ECEP is limited to the phonology of the eighteenth century, we
have excluded from the database those lexical items created or borrowed after 1800
(source: Oxford English Dictionary, January 2015).
b) WWノノゲげ example words that are not documented in any of the pronouncing dictionaries
examined have been excluded.
c) Proper names and cliticised spellings of the type ĚŽŶ͛ƚ, ĐĂŶ͛ƚ have been excluded on the
grounds that they are unlikely to be considered headwords in dictionaries. Country
names appear occasionally in lists, as in Johnston (1764), but some did not exist at the
time.
10
d) Example words that are documented in at least one pronouncing dictionary are included
in the database, and the dictionaries in which an example word does not appear are
IラSWS けNIDげ ふキくWく けNラデ Iミ デエキゲ DキIデキラミ;ヴ┞げぶく For instance, macaroni (set happY_a) is missing
in all but Perry (1775) and Scott (1786), and whorl (set NURSE) appears only in Johnston
(1764).
e) If an example word is listed in the dictionary but no pronunciation is provided, it is
IラSWS ;ゲ NラP ふキくWく けNラ Pヴラミ┌ミIキ;デキラミげぶ, such as cup (set FOOT) in Kenrick (1773).
f) At times pronouncing dictionaries do not list the precise example word, but they do list
or make reference to a related word. In such cases we take note of the latter and add an
explanatory note for users. For instance, for awn (set THOUGHT_a) we have taken awning
as the reference in five of the six dictionaries in which it is documented; and for
honourable and honesty (set HEIR) we have taken honour and honest as reference
example wordゲ キミ KWミヴキIニげゲ ふヱΑΑン) dictionary.
g) Example words for which the notation system in the original source is unclear or
ambiguous have been coded as Unclear.
Following the above method, ECEP currently lists 1,599 example words for each pronouncing
dictionary: 1,395 example words in the vowel sets in 61 subsets, and 204 example words in
consonant sets across 10 subsets. This leads to a total of 17,589 items annotated for the study
of eighteenth-century English phonology. A summary of the contents of ECEP is set out in
Table 4. Appendix II ノキゲデゲ WWノノゲげ example words that have been excluded from ECEP according
to principles a)-c).
Table 4 ECEP contents
Lexical Sets Subsets Example
words
Vowels に Wells (1982)
KIT, DRESS, TRAP, LOT, STRUT, FOOT; BATH, CLOTH,
NURSE, FLEECE, PALM, THOUGHT, GOOSE, START, NORTH,
FORCE; FACE, GOAT, PRICE, CHOICE, MOUTH, NEAR,
SQUARE, CURE; happY, lettER, commA
27 61 1,395
Consonants に Supplementary list
DEUCE, FEATURE, SURE; HEIR; WHALE
5 10 204
Total in each pronouncing dictionary 32 71 1,599
Total in all pronouncing dictionaries 17,589
10
The exception to country names is England. Note that Alexander, Charles, George and Morris are
included in ECEP because the dictionary entries refer to derivations which are no longer proper names
as such; for instance, Alexander refers to the name of the herb.
3.2 Database Annotation
The database is designed to address research questions concerning the chronological, social,
geographical and phonological distribution of variants such as /hw/~/w/~/h/ in the WHALE set,
BATH broadening or the STRUT-FOOT split, all of which are of interest to sociolinguists,
dialectologists and historical phonologists. To this purpose ECEP has been compiled to reflect
the inventory of categorically distinct sounds in the way that the eighteenth-century
pronouncing dictionaries document them; we avoid second-guessing issues of phonology here.
As Beal (1999) has rightly argued with respect to notations for orthographic :
the systems of notation provided in these pronouncing dictionaries tell us about the
phonemic inventory of the recommended accent に that is, how many phonemes there
are (we can, for instance, easily tell that Sheridan has three soundsねwhilst Spence and
Walker have four) whilst we can find out about the incidence of those phonemes from
the dictionary entries themselves. What we cannot tell from a dictionary such as The
Grand Repository is the phonetic nature of those phonemes: how do we know that the
sound in father was ぷ嵯准] rather than [æ准] ラヴ W┗Wミ ぷ坐准]? (Beal, 1999: 52)
Bringing together the information from all the pronouncing dictionaries, as we aim to do in
ECEP, will help us address BW;ノげゲ ケ┌Wゲデキラミく Our method has thus been to translate the
idiosyncratic notation systems of the dictionaries into unicode IPA transcriptions, based on the
descriptions provided by the authors in the preface or introduction to their works. According
to Bert EマゲノW┞げゲ categories of pronouncing dictionaries, eighteenth-century sources are
けtypicallyげ diacritic, so that diacritic marks indicate quality as well as quantity of sounds (cited in
Beal, 1999: 80). They all tended to use different types of diacritic marks, though, and SヮWミIWげゲ
Grand Repository in fact けstands apart from all the others both in its purpose and in the means
of executing that purposeげ (Beal, 1999: 80) in that it uses a truly phonemic system of notation
in which any one symbol always represents the same phoneme and vice versa. For instance, in
A New Dictionary of the English Language Kenrick (1773) used a notation system based on
numbers placed over each syllable, a method ┘エキIエ エW ;Iニミラ┘ノWSェWゲ ┘;ゲ キミゲヮキヴWS H┞ けデエW
IWノWHヴ;デWS Mヴ SエWヴキS;ミげ ふBW;ノ, 1999: 74). In the introduction to the work he gives readers
けSキヴWIデキラミゲ aラヴ Iラミゲ┌ノデキミェ デエW aラノノラ┘キミェ SキIデキラミ;ヴ┞げ (1773: 1に8) and then elaborates on the
description of the sounds in the けRhetorical Grammarげ ヮヴWaキ┝WS デラ it (1773: 1に57). He first
provides a table of English sounds for vowels and another for consonants, taking note of
spelling variation for the same sound, as shown in Table 5.
Table 5 KWミヴキIニ ふヱΑΑンぎ ┗ぶ ラミ けデエW ノラミェ ;ミS ゲエラヴデ マラSWゲ ラa ┌デデWヴキミェ ラ┌ヴ aキ┗W ┗ラ┘Wノゲげ
A. ďĂƌƌ͛Ě͘ bard.
E. met. mate.
I. short in hit. long in heat.
O. not. naught.
U. pull. pool.
He goes on to explain the notation system with the word fascination as an illustrative example:
(1) The word is next printed, as it is divided into syllables according to a right pronunciation,
with figures placed over each syllable, to determine its exact sound, as the figures
correspond with those of the above table of sounds: thus FA
11
S-CI
15
-NA
12
-TI
1
ON.]
Now, by referring to the table, we find that the several syllables are to be
pronounced like the words placed over against the numbers 11, 15, 12, 1; by which the
quality of the sound, or the power of all the vowels, is exactly determined.
By shewing farther that the consonant C in the second syllable is printed in Italicks, it
is known, by the table of consonants, that it is here pronounced soft like an S. Again, the
letters TI in the last syllable being printed also in Italics, it is plain from the same table
that they have the usual power of sh; so that the word must be pronounced as if it had
been printed FA
11
S-SI
15
-NA
12
-SHO
1
N.
(Kenrick, 1773: vii)
KWミヴキIニげゲ ゲ┞ゲデWマ is キデゲWノa ; ヴWaWヴWミIW aラヴ PWヴヴ┞げゲ ふヱΑΑヵぶ dictionary, which also takes inspiration
from Jラエミゲデラミげゲ ふヱΑヶヴぶ method, and in turn is found in Sheridan (1780) in combination with
B┌Iエ;ミ;ミげゲ ふヱΑヵΑぶ ヴWゲヮWノノキミェ notations (Beal, 1999: 75, 78). The system in Walker (1791) is
け┗キヴデ┌;ノノ┞ キSWミデキI;ノ デラ デエ;デ SW┗キゲWS H┞ Sheridanげ ふBW;ノ, 1999: 78に9). Walker argues that
SエWヴキS;ミげゲ けmethod of conveying the sound of words, by spelling them as they are
ヮヴラミラ┌ミIWSが キゲ エキェエノ┞ ヴ;デキラミ;ノ ;ミS ┌ゲWa┌ノげが ;ミS デエWヴWaラヴW キデ けゲWWマWS デラ IラマヮノWデW デエW キSW;げ ラa
W;ノニWヴげゲ own dictionary (Walker, 1791: iii). Fig. 1 shows a summary of SエWヴキS;ミげゲ notation
system, where vowels are categorized けH┞ デエW デキデノWゲ ラa Fキヴゲデが SWIラミSが ;ミS ʍエキヴS ゲラ┌ミSゲが
;IIラヴSキミェ デラ デエW ラヴSWヴ キミ ┘エキIエ デエW┞ ノキWが ;ミS ;ゲ デエW┞ ;ヴW マ;ヴニWS H┞ デエラゲW aキェ┌ヴWゲげ ふ1780: 4),
and where consonants are preceded by a vowel (first row) or H┞ けゲラ┌ミSキミェげ デエW characters so
デエ;デ けデエWキヴ ミ;デ┌ヴW ;ミS ヮラ┘Wヴゲ ┘キノノ HW W┝ヮヴWゲゲWS キミ デエWキヴ ミ;マWゲげ ふヱΑΒヰぎ ヵぶく As an illustrative
example from the dictionary entries (see (2)), the example word whisker is documented by
Sheridan with the consonant cluster hw in the first syllable (set WHALE_a) and with the vowel u
1
in the second syllable (set lettER), that is IPA /朔ヴ/.
Figure 1 SエWヴキS;ミげゲ ふヱΑΒヰぎ ヵぶ notation system for vowels and consonants11
(2) WHISKER, hwi
1ゲげ-ku1r. s. The hair growing on the cheek unshaven, the mustachio.
(Sheridan, 1780: s.v. whisker)
Once the IラヴヴWゲヮラミSWミIW HWデ┘WWミ デエW SキIデキラミ;ヴキWゲげ ゲ┞ゲデWマゲ ;ミS デエW IPA Iラミ┗Wミデキラミゲ was
established (see Appendix III for a sample of two dictionaries), the relevant segment of each
example word was transcribed using IPA symbols in an individual entry in the database. The
11
Here, the symbols and refer to the semivowels /w/ and /j/.
following methodological principles were followed for the interpretation of all pronouncing
dictionaries. First, the symbol っ嵯准/, which would be used for the vowel produced by RP
speakers in the BATH, PALM, and START sets, has not been used in our IPA transcriptions; rather,
we have consistently used /a准/ in line with the general view by historical phonologists that the
H;Iニキミェ デラ っ嵯准/ was a later process (e.g. Lass, 1999: 104). This concerns the sets BATH, PALM,
START, and variants in FACE, LOT, SQUARE, THOUGHT, TRAP. Second, all the eighteenth-century
dictionaries examined describe and/or prescribe a rhotic pronunciation. Since it is therefore a
given that orthographic r is pronounced in all contexts, we have included post-consonantal /r/
in our transcriptions only when rhoticity is relevant to the pronunciation of the vowel in the
example word, namely in the sets CURE, FORCE, lettER, NEAR, NORTH, NURSE, SQUARE, START. In these
sets, historical changes in the pronunciation of the vowels are connected to the presence or
loss of rhoticity. The exceptions are the subsets CURE_ci, CURE_cii, FORCE_c, NEAR_c, NEAR_f,
NORTH_c, SQUARE_c, START_c because the example words in these subsets all have the vowel
before /r/ followed by another vowel, as in boorish, curious, and therefore rhoticity is not an
issue.
12
In the sets SURE, FEATURE, HEIR post-consonantal /r/ has not been included in the
transcription either, because the relevant segment in these sets is the prevocalic consonant,
not the vowel. Third, where current transcription conventions vary, we have chosen the one
that most closely corresponds with the descriptions provided by our eighteenth-century
sources. For example, in transcribing the vowel of the lexical set FLEECE we have chosen /i衾/
rather than /i/ because thW マ;テラヴキデ┞ ラa ラ┌ヴ ゲラ┌ヴIWゲ SWゲIヴキHW デエキゲ ;ゲ ; けノラミェげ ┗ラ┘Wノく
Authors typically provide a single pronunciation; if they comment on variation in the
pronunciation of a particular word we document that in a separate column, as shown in Table
6.
Table 6 Illustrative examples of example words with IPA variants
Lexical Set Subset Example word IPA IPA variant Dictionary
BATH BATH_a plant æ a准 Walker 1791
CURE CURE_ai your jo准r j朔ヴ Jones 31798
FACE FACE_a great e准 i准 Sheridan 1780
FOOT FOOT bosom u 朔 Scott 1786
SURE SURE_a sure_cn sju准 崎テ┌准 Johnston 1764
SQUARE SQUARE_a bear e准r i准r Buchanan 1757
WHALE WHALE_a whist hw w Kenrick 1773
In addition, if authors elaborate further on a context in which there is variation, the passage is
recorded in the field Metalinguistic Comments. An example of this is the need to explain that a
difference in pronunciation implies a difference in meaning, as noted by Buchanan (1757) for
the lexical item bear (set SQUARE_a)ぎ ;ゲ ; ミラ┌ミ マW;ミキミェ けA ┘キノS HW;ゲデげ キデ キゲ ヮヴラミラ┌ミIWS bĤar
(IPA /i准r/), ┘エキノW ;ゲ ; ┗WヴH マW;ミキミェ けʍラ I;ヴヴ┞げ デエW ヮヴラミ┌ミIキ;デキラミ キゲ beĈr (IPA /e准r/). If the
remarks convey prescriptive attitudes towards either variant, this is further annotated in the
fields for Attitudes (i.e. positive, negative, neutral) and Labels (e.g. vulgar, improper).
13
Criticism is usually related to pronunciations considered けv┌ノェ;ヴげ, whether in the sense けIラ;ヴゲWが
┌ミヴWaキミWSげ ふOED s.v. vulgar II.13.d) and けマW;ミき ノラ┘げ (Johnson, 1755: s.v. vulgar, sense 2), or in
12
There is a peculiar case in which post-consonantal /r/ stands in variation with /l/, namely in colonel
(set NURSE) with IPA variants /朔ノっ ;ミS っ朔ヴっく Jラエミゲデラミ ふヱΑヶヴぶ ゲキマヮノ┞ ノキゲデゲ デエW デ┘ラ ┗;ヴキ;ミデゲ ┘キデエラ┌デ a┌ヴデエWヴ
IラママWミデぎ けI境ノラミWノが I┎ヴミWノげが ┘エキノW KWミヴキIニ ふヱΑΑンぶ マ;ニWゲ デエW aラノノラ┘キミェ ヴWマ;ヴニぎ けIt is now generally
sounded with only two distinct syllables, ĐŽů͛ŶĞů, and vulgarly as if written cur-nelげ, that is IPA /左ノ/ and
っ朔ヴっ ヴWゲヮWIデキ┗Wノ┞く HWヴW ┘W エ;┗W ヮヴWゲWヴ┗WS デエW っヴっ キミ デエW デヴ;ミゲIヴキヮデキラミく
13
Beal (1999: 48に58ぶ SキゲI┌ゲゲWゲ ┘エWデエWヴ ;┌デエラヴゲ ラa ヮヴラミラ┌ミIキミェ SキIデキラミ;ヴキWゲ ┘WヴW けェララSげ ヮエラミWデキIキ;ミゲ
ラヴ ミラデが ;ミS エラ┘ けSWゲIヴキヮデキ┗Wげ ラヴ けヮヴWゲIヴキヮデキ┗Wげ デエWキヴ ヴWマ;ヴニゲ ┘WヴWく
デエW ゲWミゲW けIラママラミノ┞ ラヴ I┌ゲデラマ;ヴキノ┞ ┌ゲWS H┞ デエW ヮWラヮノW ラa ; Iラ┌ミデヴ┞き ラヴSキミ;ヴ┞が ┗Wヴミ;I┌ノ;ヴげ
(OED s.v. vulgar I.3.a)が ラaデWミ キミ ヮエヴ;ゲWゲ ゲ┌Iエ ;ゲ けデエW ┗┌ノェ;ヴ ゲ;┞げ ラヴ け;マラミェ デエW ┗┌ノェ;ヴげ ふゲWW ;ノゲラ
Sundby et al., 1991: 40に2, 52に3). W;ノニWヴげゲ Wミデヴ┞ aラヴ plant (set BATH_b) in passage (3) provides
an illustrative example of this. For his part, the Irish author Sheridan often comments on
variation between English and Irish pronunciation, as in デエW ゲWIデキラミ ラミ け‘┌ノWゲ デラ HW ラHゲWヴ┗WS
H┞ デエW N;デキ┗Wゲ ラa IヴWノ;ミS キミ ラヴSWヴ デラ ;デデ;キミ ; テ┌ゲデ Pヴラミ┌ミIキ;デキラミ ラa Eミェノキゲエげ ふヱΑΒヰぎ ヵΓに62). See,
for instance, his passage in (4) about lexical items such as great (set FACE_a), where he warns
けデエW ェWミデノWマWミ ラa IヴWノ;ミSげ to avoid the mistaken pronunciation /i准/ for the けテ┌ゲデげ ヮヴラミ┌ミIキ;デキラミ
/e准/ in English. Sheridan emphasizWゲ デエ;デ けぷ;へ ゲデヴキIデ ラHゲWヴ┗;デキラミ ラa デエWゲW aW┘ ヴ┌ノWゲ ぷくくくへ ┘キノノ
enable the well-educated natives of Ireland to pronounce their words exactly in the same way
as the more polished part of the inhabitants of England Sラげ ふヱΑΒヰぎ ヶヰぶく
(3) PLANT, pla
4
nt. [IPA /æ/]
There is a coarse pronunciation of this word, chiefly among the vulgar, which rhymes
it with aunt [i.e. a
2
nt, IPA /a准/]. This pronunciation seems a remnant of that broad sound
which was probably given to the a before two consonants in all words, but which has
been gradually wearing away, and which is now, except in a few words, become a mark
of vulgarity. (Walker, 1791: s.v. plant; s.v. aunt)
(4) The second vowel, e, is for the most part sounded ee by the English [IPA /i准/], when the
accent is upon it; whilst the Irish in most words give it the sound of second a
2
, as in hate
[IPA /e准/]. This sound of e3 [ee] is marked by different combinations of vowels, such as,
ea, ei, e final mute, ee, and ie. [...] The English constantly give this sound [i.e. /i准/] to ea,
whenever the accent is on the vowel e, except in the following words, gre
2
at, a pe
2
ar, a
be
2
ar, to be
2
ar, to forbe
2
ar, to swe
2
ar, to te
2
ar, to we
2
ar. In all which the e has its second
sound [e
2
, IPA /e准/]. For want of knowing these exceptions, the gentlemen of Ireland,
after some time of residence in London, are apt to fall into the general rule, and
pronounce these words as if spelt, greet, beer, sweer, &c. (Sheridan, 1780: 59)
Finally, since word frequency may be an influential factor in the choice of variants or in the
development of sound changes such as those arising through lexical diffusion, we have
compiled a frequency list with an estimated frequency rate of the lexical item in eighteenth-
century British English, based on the data available in the multi-genre historical corpus ARCHER
1650–1999, version 3.2 (535,767 words).
4. Web-based Interface
The ECEP database will be made available to users via a web-based application hosted on the
website of the Humanities Research Institute, University of Sheffield. Access to ECEP will be
free for any user registering at the website. The reference line for citation is as follows:
ECEP = Eighteenth-Century English Phonology database, 2015. Compiled by Joan C. Beal,
Nuria Yáñez-Bouza, Ranjan Sen and Christine Wallis. The University of Sheffield and
Universidade de Vigo. Published by: University of Sheffield.
http://hridigital.shef.ac.uk/eighteenth-century-english-phonology
The online interface has been developed using client-side HTML and Javascript and server-
side PHP and MySQL. It displays two layouts に Browse, Search に and offers a download
function in CVS file format. The design aims to replicate the MS Access format, and therefore it
offers three main blocks of data: the lexical sets plus metadata for works and for authors (see
Table 1). Fig. 2 shows the homepage, from which each of these sections can be accessed (see
top row), and from which users can go directly to the pronouncing dictionary they are
interested in (see Buchanan 1757 and Burn 1786 in the image). Fig. 3 and Fig. 4 are
screenshots of the Browse layouts for Works and Authors, respectively. The Search tool allows
users to search in one field or in a combination of fields. Fields which contain a predefined list
of values (e.g. lexical sets, example words, authorげゲ name) offer an automatic drop-down list
menu to facilitate selection, as in the field IPA in Fig. 5. Lexical sets and example words can be
searched in the entire database or within a particular work; for instance, Fig. 6 displays a
sample of the set BATH, where users can compare the occurrence of the variants /æ/, /a准/ and
っ査准/.
Figure 2 ECEP online interface に homepage
Figure 3 ECEP online interface に Works in Browse layout
Figure 4 ECEP online interface に Authors in Browse layout
Figure 5 ECEP online interface に Lexical Sets in Search layout
Figure 6 ECEP online interface に Lexical Sets in Browse layout
5. Case Studies
In this section we report on two case studies that demonstrate the value of evidence that can
be systematically extracted from this database for the analysis of segmental and
suprasegmental phonology, in their regional and chronological settings. The results constitute
a valuable distillation of the conditioning factors to look out for in a wider range of eighteenth-
century evidence, hence a point of departure for further investigation. In this light, these
results must be interpreted as indications of patterns rather than definitive analyses; that is, if
a sub-set of dictionary writers display a pattern in their choices, it is worth exploring that
pattern using all the available evidence to establish whether a conditioning factor in the sound
change indeed underlies it.
The first of these studies examined variation in the pronunciation of け┘エげ (/hw/~/w/~/h/) in
example words of the consonantal set WHALE (Beal and Sen, 2014a; 2014b). In present-day RP,
words such as whale, what, where begin with the sound /w/, whilst who, whole have initial
/h/. Eighteenth-century pronouncing dictionaries present evidence, through their orthographic
systems, of variation between /hw/ and /w/ for the first set, hence a preserved versus
unpreserved contrast in where/wear. The fifty example words in this consonantal set were
selected on the basis of their occurrence in as many of the sources as possible, and to
represent three phonological contexts: (1) thirty-nine example words beginning with the
ゲヮWノノキミェ け┘エげ ┘エキIエ ;ヴW ヮヴラミラ┌ミIWS ┘キデエ っ┘っ キミ ヮヴWゲWミデ-day RP, (2) six example words with
キミキデキ;ノ け┘エげ ┘エキIエ ;ヴW ミラ┘ ヮヴラミラ┌nced with initial /h/, and (3) five example words ┘キデエ け┘エげ
word internally, which are now all pronounced with internal /w/ (e.g. somewhere). The
デヴ;ミゲIヴキヮデキラミゲ ラa デエW け┘エげ ゲWェマWミデ キミ W;Iエ W┝;マヮノW ┘ラヴS aラ┌ミS キミ nine of the eleven
pronouncing dictionaries compiled in ECEP are displayed in Table 7.
Table Α Tヴ;ミゲIヴキヮデキラミゲ ラa デエW け┘エげ ゲWェマWミデ in the WHALE lexical set
WHALE set Bu57 Joh64 Ke73 Pe75 Sp75 Sh80 Bur86 Wa91 Jo97
whale hw hw w w hw hw w hw hw
wharf hw w w w hw hw w hw hw
what hw hw w w hw hw w hw hw
wheat hw hw w w hw hw w hw hw
wheedle hw hw w w hw hw w hw hw
wheel hw hw w w hw hw w hw hw
wheeze hw hw w w hw hw w hw hw
whelm hw hw hw hw hw hw w hw hw
whelp hw hw w w hw hw w hw hw
when NID hw w w hw hw w hw hw
whence NID hw w w hw hw w hw hw
where_cn NID hw w w hw hw w hw/w hw
wherry hw hw hw w hw hw w hw hw
whet hw hw w w hw hw w hw/w hw
whether hw hw hw w hw hw w hw hw
whey_cn hw hw w w hw hw w hw hw
which NID hw w w hw hw w hw hw
whiff hw hw w w hw hw w hw hw
whiffle hw hw w w hw hw w hw hw
whig hw hw w w hw hw w hw hw
while NID hw w w hw hw w hw/w hw
whim hw hw w w hw hw w hw hw
whimper hw hw w w hw hw w hw hw
whin hw NID w w hw hw w hw hw
whine hw hw w w hw hw w hw hw
whip hw hw w w hw hw w hw hw
whirl hw hw w w hw hw w hw hw
whisk hw hw hw hw hw hw w hw hw
whisker_cn hw hw hw hw hw hw w hw hw
whisper hw hw hw hw hw hw w hw hw
whist hw hw hw/w w/hw hw hw w hw hw
whistle hw hw hw w hw hw w hw hw
whit hw hw w w hw hw w hw hw
white hw hw w w hw hw w hw hw
whither hw hw w w hw hw w hw hw
whitlow hw hw hw hw hw hw hw hw hw
whitsuntide hw hw hw w hw hw hw hw hw
whiz hw hw hw hw hw hw hw hw hw
who_cn NID h h h hw h h h h
whole h h h h hw h h h h
whom NID NID h h hw h h h h
whoop hw h h h hw h h h h
whore_cn h h h h h h h h h
whose_cn NID h h h hw h h h h
why NID hw w w hw hw hw hw hw
elsewhere NID NID NID hw hw hw w hw hw
nowhere NID NID NID w hw hw w hw hw
overwhelm hw hw NID hw hw hw hw hw hw
somewhat NID NID NID w hw hw w hw hw
somewhere NID hw NID w hw hw w hw hw
This systematic data collection even on such a small scale enabled us to identify patterns in the
evidence, along dimensions commonly under investigation in sociolinguistic, historical and
phonological research, namely geography, chronology, phonology, lexical factors, and social
Iノ;ゲゲく F┌ヴデエWヴマラヴWが デエW ミ;デ┌ヴW ラa デエW S;デ; ;ノゲラ Wミ;HノWS ┌ゲ デラ ェノW;ミ けSキヴWIデげ W┗キSWミIW キミ デエW
form of contemporary commentary on the choices made by the authors. A notable example is
that Walker (1791) presents the loss of the /hw ~ ┘っ Iラミデヴ;ゲデ ;ゲ ; ゲヮWIキ;ノ I;ゲW ラa けエ-Sヴラヮヮキミェげ
in lower-class London English, which was just beginning to attract social stigma in the middle
of the eighteenth century (Beal, 1999: 176に8).
Three main patterns emerged from the data based on geographical and chronological
SキゲデヴキH┌デキラミく Fキヴゲデノ┞が デエW LラミSラミ ;┌デエラヴゲ ヮヴWaWヴ っエ┘っ デラ っ┘っ デラ ;┗ラキS デエW ヮヴラゲIヴキHWS けエ-
Sヴラヮヮキミェげ ;ゲ SキゲI┌ゲゲWS H┞ W;ノニWヴが ┘キデエ デエW W┝IWヮデキラミ ラa KWミヴキIニ ふヱΑΑンぶが one of the earliest of
the group, presumably because the stigmatization of /w/ had not yet fully taken effect by this
time. Secondly, two out of the three Scottish authors prefer /w/ (Perry, 1775; Burn, 1786),
whereas the earliest, Buchanan (1757), prefers /hw/. Perry and Burn appear to be advising a
more London-like pronunciation to avoid the Scottish /hw/, stigmatized due to its regional
connotations (Douglas, 1991 [1779]: 141). The /w/ pronunciation could therefore be analysed
as a hypercorrect Anglicism, one which is particularly remarkable in the light of the
contemporaneous oppラゲキデW デヴWミS キミ LラミSラミ ┘エWヴW っエ┘っ ┘;ゲ ヮヴラゲIヴキHWS S┌W デラ けエ-Sヴラヮヮキミェげく
Arguably, this trend was only taking hold in London at the time and had not yet reached the
consciousness of the Scottish authors. Thirdly, Spence (1775) from Newcastle in north-east
England has near-consistent /hw/, even in words containing a following back, rounded vowel,
where other authors have delabialized /h/ e.g. who. Along with the fact that Spence is the only
;┌デエラヴ デラ ┌ゲW ; ゲヮWIキ;ノ ゲ┞マHラノ aラヴ デエW け┘エげ ゲラ┌ミSが デエキゲ Iラ┌ノS HW キミデWヴヮヴeted as evidence in
SヮWミIWげゲ Sキ;ノWIデ aラヴ マラミラゲWェマWミデ;ノ っ柵っが ;ミS ミラデ ; Iノ┌ゲデWヴ っエ┘っが デエW ┗ラキIWノWゲゲ Iラ┌ミデWヴヮ;ヴデ ラa
voiced /w/ which also retained its labial element before back, rounded vowels, e.g. wound,
womb, wool, wood.
Two lexically based patterns emerged. The first, homophone avoidance, as shown by
B┌Iエ;ミ;ミげゲ ふヱΑヵΑぶ っエ┘っ aラヴ whoop け; Iヴ┞げが H┌デ っ┘っ aラヴ whoop け; HキヴSげ ;ミS B┌ヴミげゲ ふヱΑΒヶぶ ;ミS
PWヴヴ┞げゲ ふヱΑΑヵぶ Whitsuntide with /hw/ and whit with /w/, is evidence that sensitivity to the
contrast remained to a sufficient degree to construct minimal pairs in some regions, notably
Scotland where the contrast survives to the present day. The second, onomatopoeia as
illustrated by /hw/ in whisk, whisper in Kenrick (1773) and Perry (1775), could also be
interpreted as evidence for an increased chance of /hw/-preservation (perhaps enhanced by
considerations of sound symbolism) before a front vowel in precisely these two authors, e.g.
whelm, and often with a following /s/, e.g. whisk, whiskers, whisper.
Aside from this partial pattern, two main explanations based on phonological context
emerged. The first is the unambiguous delabialization to /h/ before any vowel that is higher
;ミS マラヴW ヴラ┌ミS デエ;ミ っ査っ ふデエWヴW キゲ ミラ っエっ キミ wharf in any of the dictionaries), e.g. who.
Secondly, the realization of word-キミデWヴミ;ノ け┘エげ キミ PWヴヴ┞ ふヱΑΑヵぶ appears to be conditioned by
stress, as marked by the author himself, thus stressed-syllable onset /hw/ in overwhélm,
elsewhére, but unstressed-syllable onset /w/ in sómewhere, sómewhat, nówhere.
We therefore repeatedly found that by systematically collating the different types of direct
evidence afforded by the eighteenth-century pronouncing dictionaries (sounds and stress,
contemporary commentary, geographical and chronological spread), and analysing them in the
light of acknowledged influences in sound change, we were able to posit accounts for many of
the patterns in a way that only such an orderly approach to the data permitted.
The second case study explored palatalization in eighteenth-century English, i.e. where a
postalveolar fricative /崎 桜/ or affricate /三 鮫/ arose from the sequence alveolar /t d s z/ + /j/ +
/u衾/, as in the word tune (Beal and Sen, 2015). The palatalization of alveolar consonants before
late Middle English /u准/ is still variable and is diffusing in present-day English. The OED gives
several pronunciations for mature ふWくェく っマ恭ろ三搾恭 れ マ恭デテ搾恭っぶが H┌デ ヮヴラ┗キSWゲ ラミノ┞ ┌ミヮ;ノ;デ;ノキ┣WS
(/dj tj/) transcriptions for endure, tune, and duke, despite the common occurrence of
palatalized (and yod-dropped) variants in many varieties of British English. Extensive variability
is not recent in origin, and we can already detect relevant patterns in the eighteenth century
from the evidence of a range of pronouncing dictionaries; for instance, Beal (1996; 1999) notes
a tendency for northern English and Scottish authors to be more conservative. She concludes
デエ;デ ┘W ヴWケ┌キヴW け; IラマヮヴWエWミゲキ┗W ゲ┌ヴ┗W┞ ラa デエW マ;ミ┞ ヮヴラミラ┌ミIキミェ SキIデキラミ;ヴキWゲ ;ミS ラデエWヴ
┘ラヴニゲ ラミ ヮヴラミ┌ミIキ;デキラミげ ふヱΓΓヶぎ ンΑΓぶ デラ ェ;キミ マラre insight into the historical variation patterns
underlying present-S;┞ Eミェノキゲエく ʍエキゲ ゲデ┌S┞ ヮヴWゲWミデWS ヴWゲ┌ノデゲ aヴラマ ゲ┌Iエ ; けIラマヮヴWエWミゲキ┗W
ゲ┌ヴ┗W┞げ H;ゲW ラミ デエW S;デ; IラマヮキノWS キミ ECEPく
The data were divided into two main consonantal lexical sets: DEUCE where there was no /r/
following the vowel, and SURE where an /r/ followed. A third set was FEATURE, where the vowel
following the palatalized sequence is schwa in present-day English, and /r/ originally followed
the vowel. This division was made after preliminary examination demonstrated a clear
difference in the behaviour of the consonantal sequences in these contexts. We were then
able to further clarify the nature of the divergence after constructing the database with
information from ten dictionaries, and with word-frequency information for the period 1700に
1799 from ARCHER 3.2. As mode of illustration, Table 8 displays transcriptions of example
words in some of the DEUCE subsets, and Table 9 transcriptions of some of the SURE and FEATURE
subsets.
Table 8 Transcriptions of the DEUCE lexical set in ECEP: subsets DEUCE_a /t/, DEUCE_b /t/,
DEUCE_c /t/, DEUCE_b /s/
DEUCE_a /t/ Bu57 Joh64 Ke73 Pe75 Sp75 Sh80 Sc86 Wa91 Jo97 Jo98
opportunity tju: tju: tju: tju: tju: tju: tju: tju: tju: tju:
Tuesday tju: tju: tju: tju: tju: デ崎┌ぎ tju: tju: tju: tju:
tumour tju: tju: to: tju: tju: デ崎┌ぎ tju: tju: tju: tju:
tube tju: tju: tju: tju: tju: デ崎┌ぎ tju: tju: tju: tju:
tutor tju: tju: tju: tju: tju: デ崎┌ぎ tju: tju: tju: tju:
tune_cn tju: tju: tju: tju: tju: デ崎┌ぎ tju: tju: tju: tju:
obtuse_cn tju: tju: tju: tju: tju: tju: tju: tju: tju: tju:
tulip tju: tju: tju: tju: tju: デ崎┌ぎ tju: tju: tju: tju:
tumult tju: tju: tju: tju: tju: デ崎┌ぎ tju: tju: tju: tju:
tubular NID NID tju: tju: NID デ崎┌ぎ tju: tju: tju: tju:
contusion tju: tju: tju: tju: tju: tju: tju: tju: デ朔 unclear
tumid tju: tju: tju: tju: tju: デ崎┌ぎ tju: tju: tju: tju:
tuberous tju: tju: tju: tju: NID デ崎┌ぎ tju: tju: tju: tju:
tunic tju: NID tju: tju: NID デ崎┌ぎ tju: tju: tju: tju:
opportune_a tju: tju: tju: tju: tju: tju: tju: tju: tju: tju:
attune NID NID tju: tju: tju: tju: tju: tju: tju: tju:
DEUCE_b /t/ Bu57 Joh64 Ke73 Pe75 Sp75 Sh80 Sc86 Wa91 Jo97 Jo98
latitude tju: tju: tju: tju: tju: デ崎┌ぎ tju: tju: tju: tju:
amplitude tju: tju: tju: tju: tju: tju: tju: tju: tju: tju:
longitude tju: tju: tju: tju: tju: デ崎┌ぎ tju: tju: tju: tju:
altitude tju: tju: tju: tju: tju: tju: tju: tju: tju: tju:
magnitude tju: tju: tju: tju: tju: デ崎┌ぎ tju: tju: tju: tju:
fortitude tju: tju: tju: tju: tju: tju: tju: tju: tju: tju:
punctual tju: tju: tju: NoP tju: デ崎搾 tju: デ崎テ┌ぎ デ崎テ┌ぎ tju:
solitude tju: tju: tju: tju: tju: tju: tju: tju: デ崎テ┌ぎ tju:
attitude tju: tju: tju: tju: tju: tju: tju: tju: unclear uncl
aptitude tju: tju: tju: tju: tju: tju: tju: tju: unclear tju:
sanctuary tju: tju: tju: tju: tju: デ崎搾 tju: デ崎テ┌ぎ デ崎テ┌ぎ tju:
mortuary_deu tju: tju: tju: tju: tju: tju: tju: デ崎テ┌ぎ tju: tju:
actuary_ deu tju: NID NID tju: NID tju: NID デ崎テ┌ぎ tju: tju:
opportune_b tju: tju: tju: tju: tju: tju: tju: tju: tju: tju:
bitumen tju: tju: tju: tju: tju: 俊デテ┌ぎ tju: 俊デテ┌ぎ tju: tju:
DEUCE_c /t/ Bu57 Joh64 Ke73 Pe75 Sp75 Sh80 Sc86 Wa91 Jo97 Jo98
tumultuous tju: tju: tju: tju: NID デ崎┌ぎ tju: tju: tju: tju:
tutorial NID NID NID NID NID NID NID デ崎テ┌ぎ NID NID
DEUCE_b /s/ Bu57 Joh64 Ke73 Pe75 Sp75 Sh80 Sc86 Wa91 Jo97 Jo98
issue sju: sju: sju: 崎テ┌ぎ sju: 崎搾 sju: 崎テ┌ぎ 崎┌ぎ 崎テ┌ぎ
consular NID sju: sju: NoP sju: 崎搾 sju: 崎テ┌ぎ 崎搾 unclear
consummate ゲ朔 ゲ朔 ゲ朔 ゲ朔 ゲ搾 NID ゲ朔 ゲ朔 ゲ朔 unclear
tissue sju: sju: NID 崎テ┌ぎ sju: 崎搾 sju: 崎テ┌ぎ 崎テ┌ぎ 崎テ┌ぎ
insulate NID NID sju: NID NID sju: NID 崎テ┌ぎ NID NID
Table 9 Transcriptions of the SURE and FEATURE lexical sets in ECEP: subsets SURE_a /t/,
SURE_c/t/, SURE_a /s/, SURE_b /z/, FEATURE /z/
SURE_a /t/ Bu57 Joh64 Ke73 Pe75 Sp75 Sh80 Sc86 Wa91 Jo97 Jo98
futurity_cn tju: tju: tju: tju: tju: デ崎┌ぎ tju: tju: unclear tju:
centurion_cn tju: tju: デ査 tju: tju: tju: tju: NID tju: tju:
mature_cn tju: tju: tju: tju: tju: tju: tju: tju: tju: tju:
maturity_cn tju: tju: tju: tju: tju: tju: tju: tju: tju: tju:
SURE_c/t/ Bu57 Joh64 Ke73 Pe75 Sp75 Sh80 Sc86 Wa91 Jo97 Jo98
maturation tju: NID tju: tju: NID tju: tju: デ崎テ┌ぎ tju: デ崎テ┌ぎ
SURE_a /s/ Bu57 Joh64 Ke73 Pe75 Sp75 Sh80 Sc86 Wa91 Jo97 Jo98
sure_cn sju: sju: 崎テ┌ぎ 崎テ┌ぎ 崎┌ぎ 崎┌ぎ sju: 崎テ┌ぎ 崎┌ぎ 崎テ┌ぎ
assure_cn sju: sju: 崎テ┌ぎ sju: 崎┌ぎ 崎┌ぎ sju: 崎テ┌ぎ 崎┌ぎ 崎テ┌ぎ
assurance_cn sju: sju: 崎テ┌ぎ sju: 崎┌ぎ 崎┌ぎ sju: 崎テ┌ぎ 崎┌ぎ sju:
insure_cn NID sju: NID NID NID NID sju: NID NID NID
ensure_cn NID NID sju: 崎テ┌ぎ NID NID NID 崎テ┌ぎ sju: 崎テ┌ぎ
surety sju: sju: 崎テ┌ぎ 崎テ┌ぎ 崎┌ぎ 崎┌ぎ sju: 崎テ┌ぎ 崎┌ぎ 崎テ┌ぎ
insurance_cn sju: sju: NID NID NID sju: sju: NID NID NID
unsure NID sju: 崎テ┌ぎ 崎テ┌ぎ NID 崎┌ぎ sju: 崎テ┌ぎ NID 崎テ┌ぎ
SURE_b /z/ Bu57 Joh64 Ke73 Pe75 Sp75 Sh80 Sc86 Wa91 Jo97 Jo98
composure zju: ┣朔 zju: ┣朔 桜搾 桜朔 ゲテ朔 桜テ┌ぎ 桜朔 桜テ┌ぎ
seizure zju: ┣朔 ┣朔 NID zju: 桜朔 ゲテ朔 桜テ┌ぎ 桜朔 桜朔
azure_SURE zju: ┣朔 ┣査 ┣朔 桜搾 桜朔 ┣テ朔 桜テ┌ぎ 桜朔 桜テ┌ぎ
closure NID NID ┣朔 ┣朔 桜朔 桜朔 NID 桜テ┌ぎ 桜朔 桜朔
FEATURE /z/ Bu57 Joh64 Ke73 Pe75 Sp75 Sh80 Sc86 Wa91 Jo97 Jo98
pleasure zju: ┣朔 桜朔 桜朔 桜搾 桜朔 ┣テ朔 桜テ┌ぎ 桜朔 桜朔
measure_cn zju: ┣朔 桜朔 桜朔 桜搾 桜朔 ┣テ朔 桜テ┌ぎ 桜朔 桜朔
treasure zju: ┣朔 桜朔 ┣朔 桜搾 桜朔 ┣テ朔 桜テ┌ぎ 桜朔 桜テ┌ぎ
leisure zju: NID 桜朔 ┣朔 桜搾 桜朔 ┣テ朔 桜テ┌ぎ 桜朔 桜朔
azure_FEAT zju: ┣朔 ┣査 ┣朔 桜搾 桜朔 ┣テ朔 桜テ┌ぎ 桜朔 桜テ┌ぎ
rasure sju: ゲ朔 ゲ朔 崎朔 zju: 崎朔 ┣テ朔 zju: 崎朔 桜テ┌ぎ
All the pronouncing dictionaries are consistently rhotic, i.e. they report syllable-final /r/ in
forms such as sure. It was found that there is significantly more palatalization when /r/ follows
(SURE/FEATURE) than when it did not (DEUCE), particularly in a post-stress syllable, thus even the
resistant Spence (1775) has /炸/ in closure, pleasure. The more frequent of these palatalized
forms (e.g. nature) seem to be the words which have become lexicalized in present-day
English.
The nature of the palatalizing phoneme was also relevant. Palatalization occurred in /sj/ in
particular, thus it is near-regular in post-stress DEUCE in Perry (1775), Sheridan (1780), Walker
(1791), and Jones (1797, 1798), e.g. /灑/ in issue. This is arguably because the high tongue
position of palatal /j/ shapes frication noise, producing post-alveolar percepts. Furthermore,
/sj/ is the only context which palatalizes in a stressed syllable with any regularity, particularly
when in a rhotic context, thus sure, surety with /灑/ even in Kenrick (1773), Perry (1775), and
Spence (1775). Stress therefore also appears to have been a conditioning factor, with
palatalization generally resisted in the onset of a stressed syllable, as noted explicitly by
Walker (1791), and more common in post-stress syllables. Pre-stress syllables also show some
palatalization, yielding interesting alternations such as /tj/útor H┌デ っ三っutórial and ma/tj/úre
but maっ三っurátion in Walker (1791).
Two other contexts proved to be more conducive to palatalization: word-initial position,
デエ┌ゲ SエWヴキS;ミ ふヱΑΒヰぶ っ三っ キミ tune, but /tj/ in attuneが ;ミS HWaラヴW ┗ラ┘Wノ エキ;デ┌ゲが デエ┌ゲ っ三っ キミ
punctual, sanctuary in Sheridan (1780), Walker (1791), and Jones (1797; 1798), but mainly /tj/
elsewhere.
Aゲ ┘キデエ デエW け┘エげ ゲデ┌S┞が Iエヴラミラノラェ┞が ェWラェヴ;ヮエ┞が ;ミS ゲデキェマ;デキ┣;デキラミ ;ノゲラ ヮヴラ┗WS デラ HW
relevant factors in accounting for the variation. Palatalization appears to have become
increasingly more common over the course of the eighteenth century: there is little in Kenrick
(1773), but Sheridan (1780; late in career) is the arch-ヮ;ノ;デ;ノキ┣Wヴく Hラ┘W┗Wヴが デエW ノ;デデWヴげゲ
dictionary was repeatedly singled out for criticism later in the century, as such pronunciations
came to be stigmatized (e.g. Jones, 1798: iv). Palatalization consequently became much less
common at the end of the century; it is less widespread, but stress-based in Walker (1791; see
his principles 376, 450, 459にヶヴぶが ;ミS ヮヴラェヴWゲゲキ┗Wノ┞ W┗Wミ ノWゲゲ Iラママラミ aヴラマ JラミWゲげ ゲWIラミS
edition ふヱΑΓΑぶ デラ エキゲ デエキヴS ふヱΑΓΒぶく Iミ デWヴマゲ ラa ェWラェヴ;ヮエ┞が SエWヴキS;ミげゲ ヮ;ノ;デ;ノキ┣キミェ デWミSWミIキWゲ
were attributed at the time to his Irish origin; this contemporary explanation requires further
scrutiny as there is little evidence that palatalization was common in the Irish English of the
time. There is little palatalization in the Scottish sources, with Buchanan (1757; early source),
and Scott (1786) notably having no palatalized forms whatsoever. Spence (1775) from
Newcastle also has little palatalization. Palatalization in the late-middle part of the eighteenth
century may have increased due to the earlier restitution of post-consonantal yod in earlier
yod-dropped forms, as in the London-H;ゲWS けマWデヴラヮラノキデ;ミ ヮヴラミ┌ミIキ;デキラミげ IヴキデキIキ┣WS H┞ KWミヴキIニ
(1773). For example, the earlier sources almost all have yod-dropped /t/ in creature (Johnston,
1764; Kenrick, 1773; Perry, ヱΑΑヵぶが H┌デ デエW ノ;デWヴ ラミWゲ エ;┗W っ三っ ふSエWヴキS;ミ, 1780; Walker, 1791;
Jones, 1797; 1798). Furthermore, this observation forms part of a further pattern revealed by
the database: there were two chronologically and phonologically distinct yod-droppings. The
first, mentioned above, notably occurred in the earlier sources after all phonemes /t d s z/ in
unstressed syllables before /r/. The second yod-dropping occurred in the later sources in a
different context: after any phoneme in a stressed syllable. Sheridan (1780) is the earliest to do
this in the single example dual; Scott (1786) is the most frequent omitter of stressed yod,
mostly in fricative and only in the most frequent words, e.g. duty.
6. Conclusion
In this paper we have presented a new digital resource for the study of English historical
phonology: the Eighteenth-Century English Phonology database (ECEP). The database provides
IPA デヴ;ミゲIヴキヮデキラミゲ aラヴ デエW ヴWノW┗;ミデ ゲWェマWミデ ラa W;Iエ W┝;マヮノW ┘ラヴS キミ WWノノゲげ ふヱΓΒヲぶ ノW┝キI;ノ ゲWデゲ
for the vowel system of present-day English, and some complementary consonant sets, as
documented in a selection of eighteenth-century pronouncing dictionaries. We have described
the structure and content of ECEP, while reporting on the methodology of compilation: source
selection, data input and annotation, and the web-based interface for users. ECEP is already
available, but work will continue with a view to enlarging the database gradually.
Originally designed as a sister to the Eighteenth-Century English Grammars database (ECEG,
2010), on the practical side ECEP will help to promote the use of databases as research
resources in historical linguistics, beyond or alongside largely available text corpora. In terms
of content, ECEP will contribute to English historical phonology, dialectology and
sociolinguistics, with a focus on the eighteenth century, but will also be of use for comparative
studies with nineteenth-century English or present-day English.
The two case studies outlined in Section 5 demonstrate the potential of ECEP as a resource
for investigating the historical phonology of Late Modern English. The database has also been
used in studies of the CLOTH lexical set (Beal and Condorelli, 2014) and of the use of labels in
the enregisterment of non-standard pronunciation (Beal and Trapateau, in prep.). The
availability of this resource will ensure that in the future historical phonology will no longer be
デエW けヮララヴ ヴWノ;デキラミげ ラa L;デW MラSWヴミ Eミェノキゲエ ゲデ┌SキWゲ (Beal, 2012b: 27).
Funding
This work was supported by the British Academy / Leverhulme Trust [SGに132806] and the
Santander Research Mobility Scheme (calls 2012に13 and 2014に15). We are also grateful to the
Humanities Research Institute at the University of Sheffield for technical support, in particular
Michael Pidd (Digital Director) and Ryan Bloor (Developer). Nuria Yáñez-Bouza would like to
thank the Spanish Ministry of Economy, Ramón y Cajal Scheme [RYC-2011-07863] and the
research group Language Variation and Textual Categorisation at the University of Vigo, the
European Regional Development Fund [FFI2013-44065-P], and the Autonomous Government
of Galicia [GPC2014/060].
References
Primary Sources
Buchanan, J. (1757). Linguae Britannicae vera pronunciatio: or, A new English dictionary.
London.
Burn, J. (1786). A pronouncing dictionary of the English language. Glasgow.
Johnston, W. (1764). A pronouncing and spelling dictionary. London.
Jones, S. (1797). Genuine edition. Sheridan improved: A general pronouncing and explanatory
dictionary of the English language, 2nd edition. London
Jones, S. (1798). Sheridan improved: A general pronouncing and explanatory dictionary of the
English language, 3rd edition. London
Kenrick, W. (1773). A new dictionary of the English language. London.
Perry, W. (1775). The royal standard English dictionary. Edinburgh.
Scott, W. (1786). A new spelling, pronouncing, and explanatory dictionary of the English
language. Edinburgh.
Sheridan, T. (1780). A general dictionary of the English language. London.
Spence, T. ( 1775). The grand repository of the English language. Newcastle.
Walker, J. (1791). A critical pronouncing dictionary and expositor of the English language.
London.
Secondary Sources
Agha, A. (2003). The social life of cultural value. Language and Communication, 23: 231–73.
Alston, R. C. (1966). A bibliography of the English language from the invention of printing to the
year 1800. Vol. V, The English dictionary. Leeds: Arnold and Son.
ARCHER 3.2. A Representative Corpus of Historical English Registers version 3.2. 1990に
1993/2002/2007/2010/2013. Originally compiled under the supervision of Douglas Biber
and Edward Finegan at Northern Arizona University and University of Southern California;
modified and expanded by subsequent members of a consortium of universities. Current
member universities are Northern Arizona, Southern California, Freiburg, Heidelberg,
Helsinki, Uppsala, Michigan, Manchester, Lancaster, Bamberg, Zurich, Trier, Santiago de
Compostela and Leicester. www.manchester.ac.uk/archer (accessed January 2015).
Beal, J. C. (1996). The Jocks and the Geordies: Modified standards in eighteenth-century
pronouncing dictionaries. In Britton, D. (ed.), English historical linguistics 1994. Papers from
the 8th International Conference on English Historical Linguistics, Edinburgh, 19ʹ23
September 1994. Amsterdam: John Benjamins, pp. 363ʹ82.
Beal, J. C. (1999). English pronunciation in the eighteenth century: Thomas Spence͛s Grand
Repository of the English Language (1775). Oxford: Clarendon Press.
Beal, J. C. (2003). John Walker: Prescriptivist or linguistic innovator? In Dossena, M. and Jones,
C. (eds.), Insights into Late Modern English. Bern: Peter Lang, pp. 83–105.
Beal, J. C. (2004). English in Modern times 1700–1945. London: Hodder Arnold.
Beal, J. C. (2012a). Late Modern English. In Brinton, L. and Bergs, A. (eds.), Historical linguistics
of English: An international handbook. Berlin: Mouton de Gruyter, pp. 63–78.
Beal, J. C. (2012b). けCaミげデ ゲWW デエW ┘ララS aラヴ デエW デヴWWゲいげ Corpora and the study of Late Modern
English. In Markus, M., Iyeiri, Y., Heuberger, R. and Chamson, E. (eds.), Middle and Modern
English corpus linguistics: A multi-dimensional approach. Amsterdam: Benjamins, pp. 13–
30.
Beal, J. C. (2012c). けBy those ヮヴラ┗キミIキ;ノゲ マキゲヮヴラミラ┌ミIWSげ: The STRUT vowel in eighteenth-
century pronouncing dictionaries. Language and History, 55(1): 5–17.
Beal, J. C. & Condorelli, M. (2014). Cut from the same CLOTH? Variation and change in the CLOTH
lexical set. Token: A Journal of English Linguistics, 3: 15–36.
Beal, J. C., Nocera, C. and Sturiale, M., eds. (2008). Perspectives on prescriptivism. Bern: Peter
Lang.
Beal, J. C. and Sen, R. (2014a). Towards a corpus of eighteenth-century English phonology. In
Davidse, K., Gentens, C., Kimps, D. and Vandelanotte, L. (eds), Recent advances in corpus
linguistics: Developing and exploiting corpora. Amsterdam: Rodopi, pp. 31に53.
Beal, J. C. and Sen, R. (2014b). W(h)o, w(h)en, w(h)ere, and w(h)at?. The eighteenth-century
pronunciation of けwhげ. Paper presented at the 18th International Conference on English
Historical Linguistics (ICEHL 18), Leuven 14に18 July 2014.
Beal, J. C. and Sen, R. (2015). EミぷSテへ┌ヴキミェ ぷ三へ┌ミWゲ ラヴ マ;ぷデテへ┌ヴW ぷ鮫へ┌ニWゲい Palatalization in
eighteenth-century English: Evidence from the Eighteenth-Century English Phonology
Database. Paper presented at the 9th Studies in the History of the English Language
Conference (SHEL-9), Vancouver 5に7 June 2015.
Beal, J. C. and Trapateau, N. (In preparation). けNamed and Shamedげ: Labels for non-standard
pronunciation in 18
th
-century pronouncing dictionaries. In Brownlees, N., Iamartino, G. and
Sturiale, M. (eds.), Labelling English, English labelled: From the 9th century to Late Modern
times.
Bloomfield, M. W. and Newmark, L. (1963). A linguistic introduction to the history of English.
New York: Knopf.
Cooper, C. (1687). The English teacher, or, The discovery of the art of teaching and learning the
English tongue. London.
Dobson, E. J. (1957). English pronunciation 1500–1700. Oxford: Clarendon Press.
Douglas, S. (1991 [1779]). A treatise on the provincial dialect of Scotland, edited by C. Jones.
Edinburgh: Edinburgh University Press.
Eighteenth-Century English Grammars database (ECEG). 2010. Compiled by Nuria Yáñez-Bouza
(University of Manchester) and María E. Rodríguez-Gil (University of Las Palmas de Gran
Canaria). www.manchester.ac.uk/eceg (accessed January 2015).
Hickey, R., ed. (2010). Eighteenth-century English. Ideology and change. Cambridge: Cambridge
University Press.
Holmberg, B. (1964). On the concept of Standard English and the history of Modern English
pronunciation. Lund: Gleerup.
Honeybone, P. and Salmons, J., eds. (2015). The Oxford handbook of historical phonology.
Oxford: Oxford University Press.
Johnson, S. (1755). A dictionary of the English language. London.
Jones, C. (1989). A history of English phonology. London: Longman.
Jones, C. (2006). English pronunciation in the eighteenth and nineteenth centuries.
Basingstoke: Palgrave Macmillan.
Lass, R. (1999). Phonology and morphology. In Lass, R. (ed.), The Cambridge history of the
English language, Vol. III: 1476–1776. Cambridge: Cambridge University Press, pp. 56–186.
MacMahon, M. K. C. (1998). Phonology. In Romaine, S. (ed.), The Cambridge history of the
English language, Vol. IV: 1776–1997. Cambridge: Cambridge University Press, pp. 373–535.
Mugglestone, L. (2003). 'Talking proper': The rise of accent as social symbol, 2nd edition.
Oxford: Clarendon Press.
Oxford Dictionary of National Biography (ODNB). www.odnb.com (accessed January 2015).
Oxford English Dictionary (OED). www.oed.com (accessed January 2015).
Ranson, R. (2012). Some aゲヮWIデゲ ラa W;ノニWヴげゲ methodology: A decisive step towards
standardization? Language and History, 55(1): 18–33.
Sheldon, E. K. (1947). W;ノニWヴげゲ influence of the pronunciation of English. Proceedings of the
Modern Language Association of America, 62: 130–46.
Strang, B. M. H. (1970). A history of English. London: Methuen.
Sundby, B., Bjørge, A. K. and Haugland, K. E. (1991). A dictionary of English normative grammar
1700に1800. Amsterdam: John Benjamins.
Tieken-Boon van Ostade, I., ed. (2008). Grammars, grammarians and grammar-writing in
eighteenth-century England. Berlin: Mouton de Gruyter.
Tieken-Boon van Ostade, I. (2009). An introduction to Late Modern English. Edinburgh:
Edinburgh University Press.
ʍヴ;ヮ;デW;┌が Nく ふヲヰヱヶぶく けPWS;ミデキIニげが けヮラノキデWげ ラヴ け┗┌ノェ;ヴげい A systematic analysis of eighteenth-
IWミデ┌ヴ┞ ミラヴマ;デキ┗W SキゲIラ┌ヴゲW ラミ ヮヴラミ┌ミIキ;デキラミ キミ Jラエミ W;ノニWヴげゲ dictionary (1791). Language
and History 59(1): 25–36.
Wells, J. C. (1982). Accents of English. Cambridge: Cambridge University Press.
Appendices
Appendix I. Consonant sets and example words
SET SUBSET EXAMPLE WORD
DEUCE DEUCE_a assume, attune, consume, contusion, deuce_cn, dual,
dubious, due, duel, duke_cn, duly, dupe_cn, duplicate,
duty_cn, exuberant, exude, fiducial, fiduciary, indubitable,
obtuse_cn, opportune_a, opportunity, presume, resume,
sudatory, sudorous, suicide, suit, suitable, suitor, supine,
suture_deu, tube, tuberous, tubular, Tuesday, tulip, tumid,
tumour, tumult, tune_cn, tunic, tutor, zeugma
DEUCE_b actuary_deu, altitude, amplitude, aptitude, arduous,
attitude, bitumen, casual, casualty, consular, consummate,
fortitude, fraudulent, glandulous, gradual, incredulous,
insulate, issue, latitude, longitude, magnitude, modulate,
module, mortuary_deu, opportune_b, punctual, sanctuary,
solitude, tissue, visual
DEUCE_c adulation, duplicity, insulation, modulation, sudation,
sudorific, superb, superior, superlative, supremacy,
supreme, tumultuous, tutorial
FEATURE FEATURE azure_ture, creature, feature_cn, fissure_ture, future,
leisure, measure_cn, nature_cn, ordure_ture, pleasure,
pressure_cn, procedure, rasure, suture_ture, torture_cn,
treasure
HEIR HEIR heir_cn, heiress, herb, herbage, honest_cn, honesty,
honour, honourable, hospital, hostler, hour, humble,
humorous, humour_cn, humoursome
SURE SURE_a assurance_cn, assure_cn, centurion_cn, cesura_caesura_cn,
durable, dure, during_cn, endure_cn, ensure_cn,
futurity_cn, insurance_cn, insure_cn, mature_cn,
maturity_cn, perdure/perdurable, sure_cn, surety, unsure
SURE_b actuary_sure, azure_sure, censure, century, closure,
composure, fissure_sure, mortuary_sure, ordure_sure,
seizure, suture_sure, tonsure, verdure
SURE_c duration, duress, induration, maturation, mensuration
WHALE WHALE_a whale, wharf, what, wheat, wheedle, wheel, wheeze,
whelm, whelp, when, whence, where_cn, wherry, whet,
whether, whey_cn, which, whiff, whiffle, whig, while, whim,
whimper, whin, whine, whip, whirl, whisk, whisker_cn,
whisper, whist, whistle, whit, white, whither, whitlow,
whitsuntide, whiz, who_cn, whole, whom, whoop,
whore_cn, whose_cn, why
WHALE_b elsewhere, nowhere, overwhelm, somewhat, somewhere
Appendix II. Example wordゲ W┝Iノ┌SWS aヴラマ WWノノげゲ ノW┝キI;ノ ゲWデゲ (alphabetic order by set)
SET SUBSET EXAMPLE WORD
BATH BATH_a giraffe, Shaftesbury
BATH_b commando, Flanders, France, Frances, Francis, ranch,
Sandra
BATH_c I;ミげデが Iラヴヴ;ノが Iヴ;ミが Iヴ;ケが マラヴ;ノWが ゲエ;ミげデが Sノ;┗が S┌S;ミ
BATH_f Basque, Cleopatra, contralto, Glasgow, graph, intransigent,
masturbate, plaque, stance, transept
CHOICE CHOICE_a --
CHOICE_b --
CHOICE_c --
CLOTH CLOTH_a Austen, Austin, Australia, Austria, doss, floss
CLOTH_b Boston, Gloucester, gong, joss, Ross
CLOTH_c Florida, horrify, Laurence_Lawrence, moribund, Norwich,
Oregon, tomorrow, Warwick
commA commA amoeba_ameba, arena, balsa, Bertha, catalpa, Cinderella,
dementia, neuralgia, panda_2, phobia, saga, visa_2, vodka
CURE CURE_ai dour, spoor
CURE_aii McClure
CURE_b Bourbon, bourse, gourmand, gourmet
CURE_ci houri, tourism, tourist
CURE_cii angostura, anthurium, bravura, Huron, Muriel, neural,
neuron_neurone, sulfuric_sulphuric, tellurium, thurible,
Truro, Ural, Uriel
DRESS DRESS fez, Leicester, rev, Thames
FACE FACE_a bouquet, fête
FACE_b aitch, beige, raid
FACE_c --
FLEECE FLEECE_a grebe, Keith, Peter, Sheila
FLEECE_b Aesop, anemic_anaemic, Caesar
FLEECE_c casino, chic, elite, prestige, ski, trio, unique, visa_1
FOOT FOOT ゲエラ┌ノSミげデ
FORCE FORCE_a chore, crore, galore
FORCE_bi Borneo
FORCE_bii --
FORCE_c angora, boron, Dora, euphoria, fedora, Gregorian,
moratorium, moron, Nora_Norah, thorium, torus,
Victoria_Victorian
GOAT GOAT_a Sラミげデが ェ;┌IエWが マ;┌┗W
GOAT_b Owen
GOOSE GOOSE_a ghoul, Moog, schooner, smooch, tarboosh, Vancouver
GOOSE_b flu, sewage, sleuth
happY happY_a birdie, boogie, breathy, budgie, calorie, chilli, corgi, edgy,
fluffy, fussy, hibachi, khaki_2, lassie, movie, Nazi_2,
prairie_2, salami, sari_2, scampi, sortie, spaghetti, strategy,
stymie, talkie, taxi
happY_b Chelsea, hockey, Swansea
KIT KIT Syria
lettER lettER indicator, liner, ogre, pallor, scorer, Tudor
LOT LOT bother, Tom, waffle
MOUTH MOUTH MacLeod
NEAR NEAR_a --
NEAR_b Deirdre
NEAR_c diphtheria, eerie, Madeira
NEAR_f Colosseum, Crimean, Galatea, Jacobean, Korea, Maria,
Sophia, TeDeum
NORTH NORTH_a Thor
NORTH_b cavort, corm, Dorking, Morgan, Mormon, morph,
morpheme, morphia, morphine, orchid, porn, quartz,
Thorpe, torque, torso, Warsaw, York
NORTH_c aural, Laura, Taurus
NURSE NURSE berth, Byrne, Earp, erg, liqueur, masseur, twerp, Worthing
PALM PALM_a blah, bra, ma, pa
PALM_b Afrikaans, Armagh, Bach, Bahai, baht, Botswana, Brahmin,
Brahms, candelabra, couvade, Dada, Dali, façade, guano,
Guatemala, guava, ha-ha, iguana, incommunicado, Java,
Kahn, Karachi, kava, kraal, laager, lager, legato, llama,
Lusaka, mafia, Mahal, maharajah_rajah, maharani_rani,
Mahdi, Malawi, Mali, marijuana, Mikado, pizzicato, Pooh-
Bah, raj, roulade, salaam, schwa, Shah, Somalia, staccato,
Sumatra, swami, Swazi, taj_1, Taj_2, Transvaal, Yokohama,
Zhivago
PALM_f aubade, bah, Bali, chorale, Colorado, enchilada, finale,
Ghana, khaki_1, khan, Koran, lava, locale, Nazi_1, Nevada,
nirvana, Pakistan, palaver, panorama, pasha, piranha, plaza,
pyjama_pajama, Shan, soprano
PRICE PRICE_a bicycle, chi, Christ, Cyprus, eider, Glynde, hi-fi, hybrid,
kaleidoscope, tried
PRICE_b --
SQUARE SQUARE_a Ayr, Eyre
SQUARE_b --
SQUARE_c aquarium, Dun Laoghaire, Eire, Mary, Pharaoh, prairie_1
START START_a bazaar, Saar
START_b aardvark
START_c aria, Bari, cascara, curare, Mata Hari, safari, Sahara, sari_1,
scenario
STRUT STRUT ɑ┌デエヴキWが マ┌ゲデミげデ
THOUGHT THOUGHT_a auk, Maugham, Paul, Raleigh, taut, Vaughan, Waugh
THOUGHT_b --
TRAP TRAP jazz, math_maths, panda_1
Appendix III. IPA デヴ;ミゲIヴキヮデキラミゲ aラヴ B┌Iエ;ミ;ミげゲ ふヱΑヵΑぶ ;ミS W;ノニWヴげゲ ふヱΑΓヱぶ ミラデ;デキラミ ゲ┞ゲデWマゲ
Buchanan 1757 IPA Walker 1791 IPA ̄ /e准/ a1 /e准/ ̆ /æ/ a2 /a准/
ai /e准/ a3 【濮准【┸ 【臑准/
au/aw 【濮准/ a4 /æ/
oi /ai/14 e1 /i准/ e違 /i准/ e2 【瀑【
ee /i/ i1 /ai/ ̆ 【瀑【 i2 【與【 ̄ /ai/ o1 /o准/ ̆ /i/ o2 /u准/ ̄【oa /o准/ o3 【濮准/ ̆ 【臑【 o4 【臑【
oo /u/ u1 /ju准/
ou 【濮u【 u2 【炬【 ̄ /ju准/ u3 【炯【 ̆ 【炬【 o3i2 【濮i【
o3u3 【a炯【
o2u2 【u濺【
14
oi and oy have a mixed sound which is never varied, and sounds like long (i) (Buchanan, 1757: 11).