A Survey Of English Spelling Carney Edward

download

https://ebookbell.com/product/a-survey-of-english-spelling-
carney-edward-5297150

Explore and download more ebooks at ebookbell.com

Here are some recommended products that we believe you will be
interested in. You can click the link to download.

A Survey Of Modern English 2nd Ed Ptzold Kurtmichael Gramley

https://ebookbell.com/product/a-survey-of-modern-english-2nd-ed-
ptzold-kurtmichael-gramley-21995518

A Survey Of Modern English 3 Ed Stephan Gramley Vivian Gramley

https://ebookbell.com/product/a-survey-of-modern-english-3-ed-stephan-
gramley-vivian-gramley-23664562

Old English And Its Closest Relatives A Survey Of The Earliest

Germanic Languages Orrin W Robinson

https://ebookbell.com/product/old-english-and-its-closest-relatives-a-
survey-of-the-earliest-germanic-languages-orrin-w-robinson-2216600

The Englishamerican A New Survey Of The West Indies 1648 1st Edition
Thomas Gage A P Newton Arthur Percival Newton

https://ebookbell.com/product/the-englishamerican-a-new-survey-of-the-
west-indies-1648-1st-edition-thomas-gage-a-p-newton-arthur-percival-
newton-51321268
Hambledon Hill Dorset England Excavation And Survey Of A Neolithic
Monument Complex And Its Surrounding Landscape Vol 12 Roger Mercer

https://ebookbell.com/product/hambledon-hill-dorset-england-
excavation-and-survey-of-a-neolithic-monument-complex-and-its-
surrounding-landscape-vol-12-roger-mercer-11937322

English Biography In The Seventeenth Century A Critical Survey 1st

Edition Allan Pritchard

https://ebookbell.com/product/english-biography-in-the-seventeenth-
century-a-critical-survey-1st-edition-allan-pritchard-51589178

A Survey Of Metaphysics E Jonathan Lowe

https://ebookbell.com/product/a-survey-of-metaphysics-e-jonathan-
lowe-47205224

A Survey Of Vinaya Literature Reprint Charles S Prebish

https://ebookbell.com/product/a-survey-of-vinaya-literature-reprint-
charles-s-prebish-49052514

A Survey Of Matrix Theory And Matrix Inequalities Revised Ed Marvin

Marcus

https://ebookbell.com/product/a-survey-of-matrix-theory-and-matrix-
inequalities-revised-ed-marvin-marcus-50124552
A Survey of English Spelling
A Survey of English Spelling

Edward Carney

London and New York

First published 1994
by Routledge
2 Park Square, Milton Park, Abingdon, Oxon, OX 14 4RN
Simultaneously published in the U S A and Canada
by Routledge Inc.
270 Madison Ave, New York N Y 10016
Transferred to Digital Printing 2006
© 1994 Edward Carney
Typeset in 10/12pt Times Linotronic 300 by
Florencetype Ltd, Kewstoke, Avon

A l l rights reserved. No part of this book may be reprinted or reproduced or

utilized in any form or by any electronic, mechanical, or other means, now known
or hereafter invented, including photocopying and recording, or in any information
storage or retrieval system, without permission in writing from the publishers.
British Library Catatoguing-in-Pubiication Data
A catalogue record for this book is available from the British Library
Library of Congress Cataloging-in-Publication Data
A catalog record for this book is available from the Library of Congress
ISBN 0^115-09270-1 (hbk)
Publisher's Note
The publisher has gone to great lengths to ensure the quality of this reprint
but points out that some imperfections in the original may be apparent
Contents

ACKNOWLEDGEMENTS xv

PREFACE xvi

CONVENTIONS, SYMBOLS A N D T E C H N I C A L T E R M S xx

1 ALTERNATIVE APPROACHES TO DESCRIBING

ENGLISH SPELLING 1
1.1 A philological approach 1
H o w did the present system come about? 1

1.2 A functional approach 3

H o w does the present system work? 3

2 LITERACY AND ENGLISH SPELLING: METHODS

AND PROBLEMS 5
2.1 Writing without speech 5
Can the written forms of English be described systematically
without linking them to spoken forms? 5

2.2 Linking speech to writing: sounds and letters 6

Points of view 6
Classroom units 7
Finding symbols for phonemes 8
Sounds and letters: the misuse of 'vowel' and 'consonant' 9
Making spelling an authority for pronunciation 11
'Learning the alphabet' 14

2.3 Divergence 15
Departures from one-to-one correspondence 15
The temptation to overstate irregularities 16
vi A Survey of English Spelling

2.4 Lexical spelling 18

Keeping the spelling of a morpheme constant 18
Processes and underlying forms 20
A n optimal orthography? The views of Venezky and Chomsky
and Halle 21
A n untidy alphabetic system? 24

2.5 Reading 26
What does a reader read? 26
Should spelling be taught deliberately or should it seep in? 28
Phonological awareness and taught spelling 29
What if you look, but cannot say? 30

2.6 Finding correspondences 32

Splitting the string of letters into units to match up with phonemes 32

2.6.1 Segmentation 32
Where to chop. The size of units 32

2.6.2 Simplicity 34
H o w many units of correspondence? 34
Complex phonemes 35

2.6.3 Exhaustiveness and discreteness 36

Remainders, overlaps and morpheme boundaries 36

2.6.4 Distinctiveness and appropriateness 38

Take it to the left, or take it to the right? 38

2.6.5 Auxiliary, inert and empty letters 40

A l l letters are 'silent', but some are more silent than others 40
Auxiliary letters: making do with the Roman alphabet 40
Inert letters: now you hear them, now you don't 41
Empty letters and special markers 42
Phonetic transparency and functional load 45
Transferability: using what you've got 46
Some hard cases 47

2.6.6 The punctuation of words 48

Spaces and hyphens 48
The flight from the apostrophe 50
Capital letters 51

2.7 A C C E N T A N D D I A L E C T : L I T E R A C Y IS H A R D E R
FOR SOME 52
 Contents vii

2.7.1 Differences between accents 52

Four types of difference between accents (Wells 1982) 52
Differences of phonetic realization 54
Differences of phonotactic distribution 55
Differences of phonological system 56
Differences of lexical distribution 58

2.7.2 Recognizing problems due to phonological interference 60

2.7.3 Phonological interference in Black American English 61

2.7.4 An orthography to cope with dialect variation 62

2.7.5 Stage dialect: spelling out the intentions of the playwright 63

2.8 RULES A N D ERRORS 66

2.8.1 Types of rule 66

Correspondence, adaptation, graphotactic and reference rules 66

2.8.2 Classroom spelling rules 69

Reference rules used by teachers 69

2.8.3 Syllabification rules: splitting words 76

Variation in usage 76
Phonetic syllable boundaries 77
General principles and usual practice 78

2.8.4 Types of spelling error 79

Social penalties 79
Competence errors and performance errors.
Variant errors. Slips 81
Lexical errors. Malapropisms 82
Analogy errors. Jumbling. Splits 84
Articulation or interference errors 84

2.8.5 'Phoneme-grapheme correspondences as cues to spelling

improvement' (Hanna et at. 1966) 86
The Hanna speech-to-text algorithm 86
Spelling rules based on the 'syllable' 87
Spelling pronunciations as a mnemonic device 88
Is English spelling '50% regular'? 89
The case for understatement: is it better than 50%? 89
The case for overstatement: is it worse than 50% ? 92
A n unjustified counterclaim? Simon and Simon (1973) 95
A futuristic algorithm for 'word recognition' 95
viii A Survey of English Spelling

2.9 SPELLING SUBSYSTEMS A N D L I T E R A C Y 96

2.9.1 Native and foreign 96

English spelling as a system of subsystems 96

2.9.2 Subsystem markers 97

Albrow's (1972) three-system model 97
Observable markers 100

2.9.3 Awareness 102

Phonological awareness, system awareness and lexical
awareness 102

3 SPEECH-TO-TEXT CORRESPONDENCES:
ENCODING 104
3.1 A corpus-based study 104

3.1.1 The database 104

3.1.2 Analysis of correspondences 107

3.1.3 Text frequency and lexical frequency 109

3.2 G E N E R A L F A C T O R S IN S P E E C H - T O - T E X T R U L E S 112
3.2.1 Common features of spelling correspondences 112

3.2.2 Consonant-letter doubling 112

Marking short vowels: matting-mating 114
§Latinate prefixes: approve, immerse, offend 119
Consonant letters that double 120
Marking stressed short vowels 121
<C>-doubling after vowels lacking primary stress 121
Absence of <C>-doubling before a word boundary 123
Absence of <C>-doubling in three-syllable words 123
Absence of <C>-doubling before certain endings 123
<C>-doubling and the constant morpheme shape 124
<C>-doubling before inverted <le> in §Basic words: mettle,
meddle 124
<C>-doubling after 'new' long vowels: /a:/, h-J and /a:/. 124
Graphotactic restrictions on <C>-doubling 124
A trial run: instances of <bb> 125
3.2.3 <e>-marking functions 129
Marking long vowels 129
Marking stem-final /s/, Izl and IQI 129
 Contents ix

Marking word-final stressed<-CCe> in French loan-words 129

3.2.4 Deletion of the final <e> marker in derived forms 130

3.2.5 The 'short word rule' 131

Lexical words usually have a minimum of three letters 131

3.3 Speech-to-text correspondences phoneme by phoneme 134

3.3.1 Short vowels: /i e a; u A D / 135

.1 / i / as in bit ('short <i>') 135
.2 Id as in bet ('short <e>') 141
.3 Ixl as in bat ('short <a>') 143
.4 fut as in put ('short <u>') 144
.5 / A / as in putt ('low short <u>') 147
.6 / D / as in pot ('short <o>') 149

3.3.2 Long phonological counterparts of the short vowels:

/ai i : ei auau/ 151
.1 /ai/ as in like ('long <i>') 151
.2 Iv.l as in leak (Hong <e>') 155
.3 lei/ as in lake ('long <a>') 164
.4 /au/ as in clown 169
.5 /au/ as in clone ('long <o>') 171

3.3.3 Long vowels associated with IMh fa: a: x ra es (j)i»/ 177

.1 /a:/ as in carf, ca/m 177
.2 /:>:/ as in court, caught 181
.3 / :/ as in curt
3 186
.4 /ra/ as infe<?r,«/ea 190
.5 /es/ as in lair 192
.6 /t»/ as in lure, /jua/ as in cwre 194

3.3.4 Other long vowels: /u! jot 3i/ 196

.1 /u;/ as in booty ('simple long <u>') 196
2/)\i:/ as in beauty ('complex long <u>') 200
.3 hil as in boy 202

3.3.5 hi and vowel reduction 203

3.3.6 Stop consonants:/p b t d k g tf (fc/ 210

.l/p/asinpan 210
.2/b/asinfcan 211
.3 III as in tame 212
.4 /d7 as in dame 214
.5 /k/ as in came 216
.6 /g/ as in game 223
x A Survey of English Spelling

.7 /|f/ as in chest 226

.8 /qV as in jest 227

3.3.7 Fricative consonants:/f v 6 0 s z J 3 h/ 228

.1/f/ as in /erry 228
.2 M as in very 231
.3 10/ as in /tag/i, and 151 as in /Ay 232
A Is/as in seal 232
.5/z/asinzea/ 238
.6 /JV as in s/iop 240
.7/3/as in measure 242
.&/hi as in head 243

3.3.8 Nasal consonants: /m n n/ 244

.1 /ml as in sum 244
.2 /n/ as in sun 246
,3/n/asinjwng 248

3.3.9 Liquids: / l r/ 249

.1 III as in /amp 249
.2 /r/ as in ramp 251

3.3.10 Semivowels: /w j / 253

.1 /w/ as in well 253
.2 /j/ as in yell 254

4 TEXT-TO-SPEECH CORRESPONDENCES:
DECODING 256
4.1 Scale and complexity in text-to-speech spelling rules 256
Text-to-speech technology 256

4.1.1 Modelling human cognitive ability in spelling 257

The M I T a l k system (Allen, J., Hunnicutt, M . S. and
Klatt, D . 1987) 257

4.1.2 Two early miniature systems with 'readable' rules 258

M c l l r o y (1974) 258
The smallest practicable text-to-speech algorithm?
Ainsworth (1973) 263

4.2 G E N E R A L F A C T O R S IN T E X T - T O - S P E E C H R U L E S 267

4.2.1 Letter-to-phoneme correspondences 267

4.2.2 Rule evaluation 270

4.2.3 Problems in correspondence matching 273

 Contents xi

4.2.4 Long values of single vowel letters 275

4.2.5 Suffix-marked short vowels 280

4.3 T E X T - T O - S P E E C H SPELLING R U L E S 280

<a>-rules 280
<t»-rules 300
<c>-rules 301
<d>-rules 306
<e>-rules 307
<f>-rules 322
<g>-rules 322
<h>-rules 328
<i>-rules 329
<j>-rules 337
<k>-rules 337
<l>-rules 338
<m>-rules 339
<n>-rules 339
<o>-rules 341
<p>-rules 356
<q>-ru!es 357
<r>-rules 358
<s>-rules 359
<t>-rules 364
<u>-rules 370
<v>-rules 376
<w>-rules 376
<x>-rules 377
<y>-rules 378
<z>-rules 380
Summary list of text-to-speech spelling rules 381

5 IDENTICAL FORMS: HOMOGRAPHS AND

HOMOPHONES 395
5.1 Homographs and homophones 395
Sharing a written or spoken shape or even both 395

5.2 Types of homograph 401

Sharing the same spelling: 'if a year is short, a minute is minute' 398

5.3 Types of homophone 401

Sharing the same pronunciation: rite, right and write 401
xii A Survey of English Spelling

5.4 New words from variant spellings 407

Many pairs of homophones developed out of variant spellings
of the same word: flower and flour 407
Some archaic specialized spellings 409
Variants which retain a semantic link 410
Variants where the semantic link has been eroded 413

5.5 Homophonous affixes 417

Identifying affixes 417
1 /-3n/, /-ffn/, /-Jn/ as in <-sion>, <-tion>, <-cion> 420
2 /-/as/, /-fas/ as in <-cious>, <-tious> 421
3 /-ant/, /-ans/ as in <-a/ent>, <-a/ence>, <-a/ency> 422
4 /-abal/ as <-able>, <-ible> in adjectives 424
5 /-a(r)/ as <-er>, <-or> in agentive nouns 426
6 /-a(r)/ as <-or>, <-our> , <-ure> in mass nouns 428
7 /-a(r)/as <-ar> in adjectives 429
8 /-ra(r)/ in <-eer>, <-ier> in agentive nouns 429
9 stressed /-et/ as <-ette>, <-et> 429
10 /-ei/ in French loans as <-e>, <-ee>, <-er>, <-et> 430
11 /-id/as <-ied>, <-id> 430
12 /-is/ as <-ice>, <-is> in nouns 430
13 / - i / ( A m E /U) in nouns as <-y>, <-ey>, <-ie> 431
14 /-an/ as <-ary>, <-ery>, <-ory>, <-ury> in nouns 431
15 Aim/ in nouns as <-ene>, <-ine> 432
16 /-i:z/ in nouns as <-ese>, <-ise> 433
17 /-aiz/ in verbs as <-ise>, <-ize> 433
18 /-al/ as <-le>, <-al>, <-el>, <-il>, <-ol> 433
19 /-am/ as <-am>, <-em>, <-om>, <-um> in nouns 438
20 variation in <e>/<i> spellings at suffix boundaries 438
21 miscellaneous suffix problems 440
22 miscellaneous prefix problems 441

6 CONVENTIONS USED IN THE SPELLING

OF NAMES 443
6.1 The spelling of names 443
Names have a wider range of spelling correspondences 443

6.2 Manipulated spellings in trade-names 446

'Klean Shyne' for 'clean shine', 'Tracc-Tac' for 'track tack' 446

6.3 Variants and totems 448

Allsopp, Alsopp, Allsop, or Alsopl 448
The prestige value of archaic spelling 449
The sensitivity factor 451
 Contents xiii

The phonetic erosion of place-names. Spelling pronunciations 452

6.4 Padding 454

Empty letters in names. Unusual final <-CC> and <-e> 454

6.5 Variability in vowel correspondences 458

,r
'How do you pronounce your name, Miss er Batho ! 458
<a> spellings 459
<e> spellings 461
<i> and <y> spellings 462
<o> spellings 463
<u> spellings 464

6.6 Stress placement and vowel reduction 464

'Is Dominica stressed on the <min> or on the <mc>? 464

6.7 Variability in consonant correspondences 465

465

7 STANDARDIZATION AND SPELLING REFORM 467

7.1 Sixteenth and seventeenth century standardizations 467

7.2 Etymology: familiarity and respectability 468

'The desire to avoid certain vulgar associations' (OED on 'coney') 468

7.3 Spelling reform in Britain and the U S A 473

7.3.1 American spelling: the influence of Webster 474
7.3.2 'New Spelling' Ripman and Archer (1948) 477
7.3.3 'Cut Spelling' Upward (1992) 479
7.3.4 'Regularized Ingush' Wijk (1959) 481

7.3.5 The Shaw alphabet and the Initial Teaching Alphabet 483

REFERENCES 489

G E N E R A L INDEX

S E L E C T I V E W O R D INDEX 512

INDEX O F SPELLING C O R R E S P O N D E N C E S 522

INDEX O F INITIAL A N D FINAL L E T T E R STRINGS IN WORDS 527

I N D E X O F SPELLING E R R O R S A N D RE-SPELLINGS 530

Acknowledgements

Here at Manchester, I am surrounded by colleagues who know more about

the English writing system than I do. William Haas, founder of the
Department of Linguistics at the University of Manchester, first interested
me in writing systems. O f the many other colleagues who have helped me
very generously with their time and patience, I wish to single out Martin
and Sue Barry, Dennis Bradley, Neville Collinge, Katharine Perera and the
present M o n t Follick professor, Nigel Vincent. If they and other long suf-
fering colleagues have failed to save me from error, the fault can only be
mine. I am also indebted to Eunice Baker and Irene Pickford for data
preparation and secretarial help in the book's early stages.
I had the privilege of presenting some of the theoretical sections of
the book in three of the annual Mont Follick Lectures at the University of
Manchester. I am grateful to the Mont Follick Trust for defraying some
of the data preparation costs of the research.
Preface

Page for page, most of what has been written about English spelling has had
a particular and often practical aim in view - to show how the writing system
has evolved over the centuries as an integral part of the history of the English
language, to advocate some reform of the writing system, to lay down a
framework for the teaching of literacy or to provide the foreign learner with
a guide to pronunciation. Few people have set out to describe the English
writing system in its present state as a working system. Yet there does seem
to be a need for more insight into how our writing system actually manages to
function. Unhappily, the spelling literature is beset with disagreements
based on ignorance and with controversies fuelled by prejudice.
If we take a radical reformist standpoint, the present English writing
system is simply not worth describing:
Our present spelling is just a chaotic concoction of oddities without
order and cohesion.
(Follick 1965: 1)
Present-day reformers are equally insistent:

proper analysis of the synchronic and diachronic evidence shows rather

that [English spelling] is unplanned, phonographically highly incon-
sistent, and historically, pragmatically and geographically fluid. Its lack
of coherent system and its unpredictable deviations from the spelling of
other languages are detrimental to its role as a medium of international
communication, while to native speakers of English it has proved a
serious obstacle to the acquisition of literacy.
(Upward 1988: 3)
Such a view has been frequently stated. Ever since English spelling settled
down in the seventeenth and eighteenth centuries, the consensus seems to
have been that the conventions we have inherited are ill-suited:

Such indeed is the state of our written language, that the darkest
hierogliphics, or most difficult cyphers which the art of man has hitherto
 Preface xvii

invented, were not better calculated to conceal the sentiments of those

who used them from all who had not the key, than the state of our
spelling is to conceal the true pronunciation of our words, from all except
a few well educated natives.
(Sheridan 1780: 13)
Happily, rather more than a few well-educated natives seem to cope
with the present system, though after a heavy investment of time and
effort. Linguists today are even prepared to concede that, in spite of its
imperfections, the English writing system has some virtue:

our orthography is possibly not the least valuable of the institutions our
ancestors have bequeathed to us.
(Sampson 1985: 213)
The extreme statement of this point of view is the provocative declaration
by Noam Chomsky and Morris Halle that:

English orthography turns out to be rather close to an optimal system for

spelling English.
(1968: 184)
N o r does a linguistic analysis of the writing system necessarily end in a
pessimistic outlook for the educational task of teaching literacy:

most English speakers learn our writing system, to one degree or other,
without much explicit analysis of it to guide them. They do not do this
by learning each item separately, but by making some sort of analysis
themselves. If they can do so well without much explicit description, the
system need not be beyond anyone when understood and presented
systematically.
(Albrow 1972: 51)
It is part of the recently established national curriculum for British schools
that 'the rules of spelling' in traditional orthography should, and presum-
ably can, be learnt. What these rules might be and how they should be
taught is another matter.
Public concern about falling standards of literacy is reinforced by the
national press. In the educational debate, spelling mistakes provide
powerful ammunition:

She had spent 'houres' over her essay but she had no 'apptitude' and no
'flare' for spelling. Even after a 'brake' for lunch, it was still 'suprisingly'
bad, though you could see what she 'ment\ A n 'independant' girl, she
did not find it 'forefilling'. Despite 'baring' a good 'refrence' and a
respectable 'adress', her hopes of college 'enterance' were dashed
because she was not 'apreciated'.
( A l l those spelling mistakes were taken from an essay written by an
xviii A survey of English spelling

18-year-old student of English literature and reassembled in this concen-

trated form by Peter Wilby as copy for an article in the Sunday Times
3 February 1985.)
Faced with frequent statements that there are no rules for traditional
English spelling and with curricula which require that the rules should
be learnt, we may conclude that there must still be scope for neutral ex-
plorative studies of the English writing system. W e still do not seem to
know enough about it. This present study tries to deal equally with the
problems of the writer and the reader, which are rather different.
Computer analysis has provided some statistical insights, especially on
the interrelationship of text frequency and lexical frequency. The present
account is also different in that it covers in some detail the spelling of
names. I have not made excursions into the general theory of writing
systems, but I have made some comparisons with the Swedish writing sys-
tem to show the workings of'spelling reform in a more 'managed' system
and a different approach to the spelling of loan-words. I have drawn
attention to differences between British and American spelling. I have
taken Southern British Standard pronunciation as the basis for analysing
spelling correspondences with selective cross-reference to American
English and other accents
I am shy of calling this study 'a survey', since some aspects of the English
writing system deserve a far more detailed treatment than I have been able
to give them. It is, however, something of a survey in the sense that I have
trawled through a fairly large database to find what regularities I could in
the English writing system and to see how they might best be described.
This is a strictly functional approach. M y main object has been to see how
traditional orthography works, or fails to work. I have not been concerned
with tracing the development of the present system over the centuries.
Only occasionally have I referred to the past to explain the present.
T o make the book more accessible as a work of reference, I have
provided a detailed list of contents and five different indexes. I have also
tended to repeat brief glosses of technical terms in the text to save the
reader from interruptions. There is also some repetition of data, where
variant pronunciations are logged under different phonemes and spellings.
In quotation, I have preserved the writer's idiosyncratic spellings, such as
Shaw's spelling of <Shakespear>. This does not, of course, imply any
recommendation.
Occasionally, I have strayed from my descriptive brief to make pres-
criptive comments on features of traditional orthography that seem
inconsistent or undesirable. Spelling reform as it has been implemented in
the United States and as it has been variously planned in Britain has been
dealt with selectively in §7.3. Detailed proposals for spelling reform in
Britain have been discussed for many years through the publications of the
Simplified Spelling Society and it is to them, or to their American and
 Preface xix

Canadian counterparts, that anyone interested in that aspect of literacy

should turn.
The current addresses of some active associations, to which any
enquiries should be made, are as follows:
1 The Simplified Spelling Society, 39 Chepstow Rise, Croydon CRO 5LX, England
2 American Literacy Council, 106 Morningside Drive, New York City 10027, U S A
3 BEtSS, 24034 Bingham Pointe Drive, Birmingham, Michigan 48010, USA (BEtSS
= 'Better Education thru Simplified Spelling')
4 The Internasional Union For The Kanadian Langwaje, 94 Glenholm Avenue,
Toronto ONT M6H 3B1, Canada
I have tried to leave a few warning signs on entrenched heresies that still
thrive in works on reading, writing and spelling. In §2 particularly, I have
sought out some examples of how not to describe English spelling. M y
excuse for this missionary zeal is that all too often the literature of literacy
is taken up by non-arguments about non-problems.
Conventions, symbols and
technical terms

SBS stands for 'Southern British Standard' and refers to the speech of
educated English speakers from London and the surrounding counties and
across southern Britain generally (Wells 1982: 117). This is the accent of
standard British English used here as the basis of the description of spelling
correspondences. Other accents are referred to when there are differences
that affect spelling.
I have used the term 'Southern British Standard' in preference to 'RP'
(or 'Received Pronunciation') for several good reasons. It is not merely
that the term 'Received Pronunciation' has a Victorian stuffiness. SBS is a
much wider concept. I wish to prevent an assumption that all the phonetic
detail of the pronunciation of R P , or of any other accent, is relevant in
tracing the correspondences between spellings and phonemes. What is
important for literacy is the number of contrasting phonemes in the accent
and their distribution, not the minutiae of how they are pronounced (their
phonetic realization).
However, R P is simply a socially defined subsystem within SBS. Readers
used to the term ' R P ' may, for our purposes, regard them as equivalent.
The two terms refer to speakers who have the same number of phonemes
with essentially the same distribution. To refer restrictively to R P in
describing English spelling, rather than the wider notion of SBS, would be
to accept a narrow social irrelevance:

Socially, [RP] is characteristic of the upper and upper middle class . . .

Occupations perhaps most typically associated with R P are barrister,
stockbroker, and diplomat. . . . Typically [RP speakers] belong to
families whose menfolk were or are pupils at one of the 'public schools'
(exclusive private schools standing outside the state education system)
(Wells 1982: 117)

A m E stands somewhat naively for 'American English'. I have used this as

a loose cover term for general features which distinguish the speech of
most Americans from Southern British Standard in phoneme contrasts and
in the distribution of phonemes in words. These are the main factors that
 Conventions, symbols and technical terms xxi

affect the description of spelling correspondences. It would have been

unwise to use the term 'General American', claiming it to be a standard
American accent, since its definition, and indeed existence, is a matter of
some controversy. What is here taken to be 'American' may well not be
valid for Eastern or Southern accents of American English. For instance,
A m E is here taken to be 'rhotic' speech which has not lost llrli before
a consonant in words such as farm. So, A m E does not here refer to
any strictly definable accent of American English, but to features of
pronunciation shared by many Americans.
The bearing of accent and dialect on spelling is discussed in §2.7.
IPA stands for The International Phonetic Association, particularly with
reference to their system of phonetic symbols.

Phoneme symbols

The phoneme symbols here used for representing Southern British

Standard are the I P A symbols used in Gimson's Pronunciation of English
(Cruttenden 1994). They are shown in table 1 below. Alongside I have
given the spelling-based letter symbols o f Cummings American English
Spelling (1988), which are derived from the symbols used in the Webster
dictionaries ( W 3 N I D ) .

Table 1 Vowel phoneme symbols

Keyword JPA Cummings symbol

Short vowels
1 bit i i {'short / ' )
2 bet e e ('short e ' )
3 bat EE a ('short a')
4 full u u (high short u')
5 dull A u ('low short u')
6 bomb D a' ('low short o ' )

Long counterparts of the short vowels

7 bite ai 1 ('long / ' )
8 beet i: e ('long e ' )
9 bait ei a ('long a')
10 bout au au
11 boat 3il 0 ('long o ' )

Long vowels and diphthongs wholly or partly associated with mil

12 bard a: (see note 2 below)
13 board o: or (see note 3 below)
14 bird 3: ur
15 beard 13 er
16 bear £3 ar
17 boor ur
18 fire aid Ir
xxii A survey of English spelling

Keyword IPA Cummings symbol

19 flour aur

Other long vowels and diphthongs

20 booty u: u ('simple long u')
21 beauty ju: yu ('complex long u')
22 boy 3i 6i
Reduced vowel
23 about 3 3 ('schwa')

Notes on the vowel phonemes and their symbols

1 The group of short vowels, (1) to (6) in Table 1, are sometimes called
'checked' vowels, because of their distribution. They do not occur in final
open syllables, so Ikstl cat is a possible English word, but there can be no
English word */ks/. Long vowels and diphthongs have no such restric-
tion and are in consequence 'free' (/kaul/ cowl, /kau/ cow). Since the
length difference between 'short* and 'long' is less marked in some
accents of English, the terms 'lax' and 'tense' may also be used. The short
vowels vary with their long counterparts in some morphemes:

/ai/ - hi sign - signal mime - mimic line - linear

li-J - Id redeem - redemption plenary - replenish serene - serenity
leil - M vain - vanity mania - manic inflame - inflammatory
/au/ - / A / renounce - renunciation South - Southern
abound - abundance
teul-lul tone-tonic omen-ominous know - knowledge.

2 In S B S the vowel /a:/ occurs not only before l/rll, but also in words such
as after, bath, cast, dance, where many other speakers, including A m E ,
would have fx/. (See §3.3.3.1 pp. 177ff.) In A m E the vowel /ail also takes
in the vowel of bomb (= balm), box, dodge, stop, watch, which in SBS is
6 above, the short low back rounded /of.
3 In SBS, the vowel b-J occurs in both caught and court as Iko-Xf, since fix//
has been lost before a consonant. In rhotic accents, as in most A m E , they
will differ as /ko:t/ and /kort/. A halfway stage may keep caught and court
distinct as /koit/ and /ksat/. '
4 In SBS and some other British accents unstressed hi can end a word such
as city, happy, where other accents including A m E have IvJ.
5 Some writers on A m E merge the stressed vowel [A] of dull together with
the unstressed schwa [a] of about in a single phoneme, using Id as the
symbol for both, since they are phonetically very similar.
6 Cummings (1988), uses W 3 N I D symbols, slightly modified. This system
tries to help the reader by choosing phoneme symbols which mirror
the most common spelling. In the vowel symbols, there may be some
 Conventions, symbols and technical terms xxiii

advantage in using terms and symbols such as 'long f IV and 'short i' HI
for the /ai/ and / i / of mime and mimic. It draws attention to the phono-
logical relationship. It cannot, however, entirely free the reader from the
task of remembering different phoneme symbols. (See pp. 8f.)
7 There are minor differences of phonetic detail between SBS and other
accents which are reflected in the SBS phoneme symbols, but which have
little bearing on spelling. For instance, the feu/ vowel (11 above) starts
without lip rounding in SBS: coat is [ksut], but before a dark [I] SBS does
have rounding as in coal [kDo+]. In other accents lip rounding remains as
[ou], or as a pure vowel [o:]. Similarly the vowel length marker [:] for li-J,
/•:/, bil, 13-J and /(j)u:/ may be inappropriate for accents in which the dif-
ference between 'short' and 'long' vowels is more a matter of vowel qual-
ity than length. The quality differences are indicated by the vowel
symbol itself IvJ - /u7, I'll - lil, etc. However, I have kept Gimson's redun-
dant and sometimes misleading length marks for a practical reason: sim-
ply to make the symbols appear different to the rapid reader.

More detailed comments will be found under each vowel phoneme in §3.3.

Table 2 Consonant phoneme symbols

Stops p pan t ten k cap f choke

bban dden g gap dsjoke
Fricatives f ferry e thin s sat j ship
v very a then z zeal 3 measure
Nasals m met n net rj long
Liquids 1 late r rate
Glides h hat w wet j yet

Notes on the consonant phonemes and their symbols

Unlike the vowels, the system of consonant phonemes shows little differ-
ence across accents. SBS and A m E , for instance, are here identical. Usually
the symbol reflects a common spelling, as in the case of /p b t d k g f v s z h
mnlrw/.
The symbol /rj/ does not include a following /g/: in SBS finger and singer
differ as /finga/ and /'sirp/; in Northern British English they may both have
/-rjgs/.
The glide /h/ is traditionally classed as as ninth member of the set of
fricatives (as in §3.3.7), though unlike them it does not enter into a voice-
ing contrast {IV - hi, 1st - Izl, etc.). The glides /w/ and 1)1 are traditionally
referred to as 'semivowels' (as in §3.3.10).
Some alternative symbols in common use for English consonants are
shown below in Table 3. The standard I P A use of /j/ for the semivowel in
yet (not the affricate in jet) needs noting.
xxiv A survey of English spelling

Table 3 Alternatives to the IPA consonant symbols

IPA Non-IPA Cummingi

thin - then e a th i h
choke-joke ch j
ship - measure J 3 s z sh zh
yet y y

Other phonetic symbols used

[-] is sometimes used to show a syllable boundary, as /u:-i/ in fruition.

[?] is the symbol for a glottal stop sound.
['] placed before a symbol indicates that the following syllable is stressed, as
in /ts'mDrsu/.
• p l a c e d underneath a symbol indicates that a consonant is syllabic, as in
[litl].

Other phonetic symbols found occasionally in the text are briefly

explained where they occur.

Types of bracketing

[ ] - square brackets with phonetic symbols enclose sounds or strings of

sounds without necessarily assigning them to any particular English
phoneme (thus, [?] represents a glottal stop sound and [1] represents the
voiceless variant of the IV phoneme found after voiceless Ipl in /plot/).
For 'phoneme' see p. xxvii.
/ / - single diagonal slashes with phonetic symbols enclose phonemes or
strings of phonemes (thus, box ends in /ks/ and seraph ends in lil). These
are traditional 'surface' phonemes, which are directly represented by a
sound.
// // - double slashes enclose a more abstract 'underlying' phoneme. Thus
llxll does not necessarily refer to present-day SBS Ixl, but effectively to
the Ixl of early Modern English, which in S B S and many other dialects
does not now survive finally (far), or before a consonant (farm). In SBS
the w o r d / a r has an 'underlying' final llxll but no actual 'surface' Ixl unless
a vowel follows immediately (far away). In A m E and other 'rhotic'
accents llxll has not been lost in these contexts, so, as captured in the
spellings, far is /far/ and farm is /farm/.
<> - angled brackets enclose letters or strings of letters (thus, box ends in
<x> and seraph ends in <ph>).
/ feo - indicates a spelling correspondence between a string of one or
more phonemes and a string of one or more letters: at the end of box we
have /ks/=<x> and at the end of seraph we have /ffe<ph>. The order of
 Conventions, symbols and technical terms xxv

the two sides depends on the topic: speech-to-text /ks/s<x> or text-to-

speech <x>=/ks/. Often the order is not critical. These are often referred
to as 'phoneme-grapheme correspondences'.
{ J - curly brackets enclose morphemes (minimal units of word structure)
cited in ordinary spelling. Thus the word photograph contains the mor-
phemes (photo) and {graph), unreliably contains the morphemes (un-),
(rely), (able) and (-ly). {photo} and (graph) are free morphemes: they can
form a word on their own. (un-) and (-ly) are bound morphemes: they do
not appear on their own, but are always attached to other morphemes.
The phonetic form of a morpheme often varies from context to context:
(photo) varies as /fsutou/, /fato/, llsutsf in telephoto, photographer, pho-
tographic. The English writing system contains spelling correspondences
with whole morphemes such as {-ed}=<ed>, where (-ed) varies phoneti-
cally as /id/ (wanted), Idi (begged), or /t/ (washed).

Notational symbols

* 'is not equal to', 'contrasts with'.

* - an asterisk attached to a written form, may denote either a wrong,
unconventional, reformed, hypothetical or dialect spelling, such as
*<stoopid>, *<sed>, *<woz>. A n asterisk is also used in formulae to
indicate some specific restriction which is indicated in the following text
(e.g. '<C>*' may exclude some letters specified in a note).
« - zero, as in <h>=0for the initial spelling in hour, where <h> has no pho-
netic counterpart.
* - a word boundary, possibly followed by suffixes that can attach to free
forms (see pp. 269ff). So the context' -<e> #' would apply not only to
care, but also to carer, caring, careful, careless, carelessly, carelessness.
The text-to-speech rules of §4 include a 'compound-guesser', which tries
to find a boundary in compounds such as carefree, careworn, by using
possible letter sequences.

The following capital letters will be found in rule formulae with particu-
lar uses:

rVI any vowel phoneme.

<V> any vowel letter, any letter from the set: < a, e, i , o, u, y >.
ICI any consonant phoneme (including glides /h, j , w/ and liquids / l , r/).
<C> any consonant letter, any letter from the set: < b, c, d, f, g, h, j , k, 1, m,
n, p, q, r, s, t, v, w, x, y, z >.

The letter <y> belongs to both the <C> anti <V> sets.

<C > zero or more consonant letters.

0

<Cj> one or more consonant letters.

xxvi A survey of English spelling

<C>* the asterisk indicates some specific restriction which is indicated in

the explanation that follows.
<C>-doubling refers to the doubling of a consonant letter as in matting
compared with mat.
T F , L F refer to the text frequency and lexical frequency of words in the
database described in §3.1.1. The percentage figures quoted for word fre-
quency in §3 exclude grammatical words such as pronouns, auxiliaries,
articles, and only refer to the frequency of lexical words.
F+, F- are used in the description of speech-to-text correspondences in §3
to draw attention to a difference between the text frequency and lexical
frequency of a particular spelling. ' F - ' indicates a tendency to occur in
low frequency words, where the per cent share of lexical frequency for
a particular spelling is notably higher than the per cent share of text
frequency. *F+' shows a tendency to occur in high frequency words,
where the per cent share of text frequency for a particular spelling is
notably higher than the per cent share of lexical frequency.

Some technical terms

affix, prefix, suffix - these terms are normally applied to bound morphemes
added to other morphemes in the process of word-formation: the added
units have a distinct function or meaning. For example, the suffix <-ness>
does not occur on its own, but only when bound to a stem, as in good-
ness, where it turns the adjective into a noun. In describing English
spelling, it is sometimes convenient to refer to initial and final strings of
letters that do not have an add-on meaning or a clear marking function.
So, <-tion> in suggestion is often referred to as a unit, even though the
division is then <sugges>+<tion>, with the <t> arbitrarily separated from
<suggest>. The string <-ant> may be dealt with as a unit not only in
accountant, but also in covenant, elephant. I have used the general
terms 'beginning' and 'ending' when the letter strings referred to are not
strictly affixes.
auxiliary, inert, empty letters refer to different functional types of letter.
See §2.6.5 pp. 40ff.
bias, workload refer to the performance of text-to-speech rules in §4 and
are explained in §4.2.2 pp. 270ff.
consonant, vowel refer only to sounds, not letters. A statement such as: 'the
stressed vowel is followed by a single consonant' would apply to both
lemon and common. There is a double consonant letter in common, but
not a double consonant. See §2.2 pp. 9ff.
diphthong, digraph -'diphthong' is a purely phonetic term and refers to a
vowel glide, as distinct from a relatively 'pure' vowel, within a single
syllable. The words cycle, omen, mouse, mice, all contain diphthongs.
The words react, poet, (with two vowels) and head, brawn, (with a single
 Conventions, symbols and technical terms xxvii

'pure' vowel) do not contain diphthongs. 'Digraph' refers to a string of

two letters: in head the digraph <ea> represents the vowel Id; in react it
represents the vowels /i:ae/.
divergence - a lack of one-to-one correspondence between phonemes and
spellings shows divergence from the alphabetic principle of one symbol
per phoneme and one phoneme per symbol (Haas 1970: 51). There may
be divergence on the phonetic side of a correspondence: <th>=/8/ and
<tn>=/5/; or on the graphic side: /f/=<f>, /f/=<ff> and /f/=<ph>. More
often than not we find divergence on both sides: /i:fe<ea>, /i:/s<ee>, and
<ea>=/ii/, <ea>=/e/, etc.
grapheme - this term has a number of different meanings in the study of
writing systems. It is not used here as an abstraction of 'letter' (the set of
different written shapes of 'the same' letter), but for any minimal letter
string used in correspondences. So, <ea> in head may be referred to as a
single (but complex) grapheme.
phoneme - this term is used in the traditional sense for contrasting units
of sound. The words exit and seraph are each pronounced with three
consonant phonemes and two vowel phonemes. Readers unused to
phonetics will probably find that their notion of 'speech sound' is effec-
tively the same as 'phoneme'. A phoneme may be realized by a range of
slightly different sounds with different speakers and in different con-
texts.
lexeme - a word defined semantically. Gaol and jail represent the same lex-
eme.
long, short - vowels are referred to as 'long' and 'short', rather than 'tense'
and 'lax'. The shortening of the vowel in words such as sanity (from
sane), is here referred to as 'third-syllable shortening'. It is otherwise
known as 'trisyllabic laxing'.
rhotic, non-rhotic - a rhotic accent is one, such as A m E , Scottish or Irish
English, in which underlying l/r/l survives in all contexts, so a firm offer is
pronounced with two instances of Ixl. A non-rhotic accent is one, such
as SBS, in which IMI is lost before a consonant and in final position, so
a firm offer has no instances of Id in S B S .
§Basic, §Greek, §Latinate, etc. refer to subsystems of spelling conventions
in the English writing system, but not directly or by definition to the his-
torical origin of words. The symbol ' § ' attached to such a label indicates
'subsystem'. The word deacon (of Greek origin) has the characteristics of
a §Basic word as does beacon (of Germanic origin). See §2.9.
Chapter 1

Alternative approaches to describing

English spelling

1.1 A PHILOLOGICAL APPROACH

How did the present system come about?

English spelling, like most of our institutions, has a history. So, in matters
of spelling, the past can help us to explain the present. We can see that
loan-words acquired from particular languages at particular times have
brought with them their own spelling conventions. The words village
and entourage are both borrowed into English from French. Village was
borrowed in the Middle Ages, but entourage seems to have been first used
in written English in the 1830s. The final French stress has shifted to the
front in the early borrowings, making them more like other English words
and the <-age> of village has become an unstressed /-ids/. In a late borrow-
ing such as entourage the <-age> is usually unchanged as /-'a: /. Studying
3

the changing relationships between written and spoken English through

time is the business of philology.
Philology studies language as part of cultural history. Its viewpoint is
diachronic, looking at language evolution, showing how the present-day
state of the language has developed over the centuries. A study of English
spelling that is primarily philological, such as Cummings (1988), will con-
sequently focus on the writing system as a transmission of culture. This will
include material which, though interesting in itself, need have no relevance
at all to the purely synchronic problems of achieving literacy in the
language as it is today.
The philological approach can dig out an interesting antiquarian diver-
sity often unsuspected by the common reader. Cummings (1988: 332), for
instance, points out that the double <bb> spellings of chubby and shabby
come before a word boundary: 'apparently the adjective came from chub,
a fattish fish'; 'shab reflects an obsolete word used to refer to a skin disease
of sheep and to a low fellow'. But for all practical spelling purposes,
chubby and shabby are to be treated as simple forms on a par with the
nouns hobby, lobby, tabby or the adjectives silly, merry, happy. Similarly,
2 A survey of English spelling

it has no bearing on the problems of literacy to know that : 'The only

known Romance instance of lul = oo is the monosyllabic rook "chess
piece'". The word did indeed come into English via French. Before that,
along with the game of chess itself, it came into French by way of Arabic
and beyond that from Persian. Yet, in spite of this chequered history, it is
spelt by §Basic spelling conventions and is identical to rook 'type of crow',
a purely Germanic word.
O n the other hand, it would be useful to keep apart (unlike Cummings
1988: 356) instances of < c o ^ k / with §Latinate prefixes in words such as
accumulate, occasion, occupy, from the <cc> of §Italianate words such as
peccadillo, staccato, toccata. The <cc> of accumulate has to be linked to the
double letters found with other prefixes, as in affiliate, alleviate, annihilate,
appreciate. The sfaccaro-type words have the common feature of penulti-
mate stress and a final vowel. These two groups of words with the spelling
<cc>s/k/ belong to different subsystems for practical spelling.
In a philological approach, the concept of 'rule' may be rather different
from the kind of rules (or generalizations) that might underlie competence
in spelling (see §2.8.1). For instance, Cummings has a 'Stress Frontshift
R u l e ' (ibid.: 127), which simply reflects the fact that the final stress of
French loan-words, as we saw in village, tended to shift to the first syllable.
This has happened in cover, honest, gravel, lemon, model, refuge, river,
scholar. The importance for present-day spellers in this group of words is
that a short vowel followed by a single consonant then became stressed in
a spelling context where you would expect a long vowel - cf. over, navel,
demon, modal, polar. This stress shift was a 'rule' in that the historical
process affected most of the early French loan-words. It is not a rule in any
synchronic sense, since there are no criteria for identifying these early loan-
words by their present-day pronunciation or any other marker of
'Frenchness'. They are simply a group of exceptions. The consequences,
too, are different for reader and writer: the writer must avoid putting a
<CC> spelling after the short vowel (*<lemmon>); the reader must avoid
'saying' a long vowel "V'lhman/.
Framing such rules as active processes may be slightly misleading in a
philological framework. This is so of the rules which account for the short
vowel before the suffix <-ity> (ibid.: 112) and before the suffix <-ic>
(ibid.: 115). The underlying form is taken to be the vowel of the base form.
So the long vowels of sane and mime are 'shortened' in the derived forms
sanity and mimic. The rule is explicitly framed as a process: 'instances of
f\i = i in V C V strings whose head vowels have been shortened (sic) by the
Suffix -ic Rule', 'head vowels have been shortened by the Suffix -ity Rule'
(ibid.: 224). Many of the words given as examples of the 'shortening'
process did not have a long vowel to shorten, either in their etymology or
in related forms in English: critic, monolithic, prolific; acidity, humidify,
ability; logic, symbolic, topic. The process of 'shortening' means, of course,
 Alternative approaches to describing English spelling 3

that the vowel 'ends up as' short. So, here the historical difference between
the stressed vowels of mimic and topic is lost sight of by the descriptive
device.

1.2 A FUNCTIONAL A P P R O A C H

How does the present system work?

A philological approach is very much concerned with asking: H o w did
these spellings come about? The present study is primarily concerned with
the question: H o w does our spelling system work? The two approaches are
not at odds; they are complementary.
The difference between the two approaches can be shown by comparing
what they would say about one particular spelling problem: the difference
between final <-o> in words such as bingo, fresco, limbo, and words with
final <-ow> such as billow, farrow, follow. The philological approach would
tell you that halo is one of some old adoptions from French and Latin, that
photo is one of a number of back-formations or 'clippings', that motto is
one of several similar loans from Italian, that mango is from Portuguese,
and that there are several words, such as jumbo, of obscure origin. The
<-ow> words, such as follow, got their <-ow> spelling in Middle English as
a regularization of a number of O l d English word endings: <we>, <rg>,
<lg>, <rh> and <u>.
On the other hand, the present study is concerned with how spellers can
distinguish between the two groups of words as they are today, so that they
do not make errors such as *<sallo>, *<mangow> for sallow and mango. If
we look at the structure of the two groups of words, we find a pattern which
could be exploited to provide teaching material. None of the disyllabic
<-ow> words has stress on the second syllable as does hello, or a long
vowel in the first syllable as does photo, or medial consonants outside the
range id, n(d), 1, r/, as does mango (see §3.3.2.5, pp.l73f.). This approach
to the description of spelling conventions is best described as functional.
Since exceptions do invite comment, it may be of occasional interest to
mention the historical reason for an exception even in a strictly functional
approach. For instance, pickerel has an exceptional 'doubled' <ck> in the
third syllable from the end (see #D6 p.123). The word means 'a young
pike', so there is a hidden boundary after the first syllable, which makes the
word rather like shrubbery in structure, where the <CC> is regular.
Frippery, gallery and scullery are similar exceptions to the doubling rules,
compared with celery, misery, but to add that frippery comes from O l d
French fripe, meaning 'rag', would not help. The <frip(p)> has no claim to
be an English morpheme. The reverse is the case in shabby and chubby
(see p. 1), where there is a hidden boundary that need not be known to
explain the doubling.
4 A survey of English spelling

If we are to focus on the problems of literacy in English, we are not

primarily concerned with how the writing system came to be in its present
state. We shall try to look at the system as it functions today for a normal
literate adult and uncover those regularities that appear to be exploitable
by a competent speller. These will not usually be in the form of recallable
and explicit 'spelling rules' to be chanted in moments of insecurity. They
will often be letter patterns which correlate well with speech patterns, with
types of word-formation or with the various subsystems of foreign loan-
words. In trying to describe the system by exploring the awareness of an
average literate adult, we shall not assume knowledge outside the adult's
competence as a speaker and reader of English.
Chapter 2

Literacy and English spelling:

methods and problems

2.1 WRITING W I T H O U T S P E E C H

Can the written forms of English be described systematically without

linking them to spoken forms?
It would be quite feasible, though far from easy, to describe the English
writing system without referring to speech. Egyptian hieroglyphs and
Chinese characters are examples of writing systems that were originally
designed to encode meanings directly, without indicating what the spoken
forms were. Even when a writing system is based on an alphabet, as is
English, the written form of the language can be studied as an independent
system of purely graphic signs. This is obviously true of those signs in
the English writing system which are not made up of letters: written signs
such as <$>, <%>, <+>. The ampersand <&> was originally a manuscript
ligature of Latin <et> 'and', but for present English users, it is just a single
squiggle. For ordinary silent reading, we do not need to know a way of
saying these signs as /'dob/, /pa'sent/, /pks/ or /<end/. Indeed, in the case
of <@> we probably do not know a way of 'saying' it. These signs, like
numbers and scientific symbols, are international and independent of
speech.
The writing system also includes punctuation signs such as < . , : ; - ! ?>
and indeed the space between words. These relate to speech, but only in-
directly. Their main function is to show grammatical structure and its
boundaries, and to some extent the attitude of the writer. We know that
'Doesn't she enjoy a game of bridge!' is not expressing doubt, as it would
if <?> were used. W e know that '. . . bridge!' will probably have a falling
tone as an expression of certainty. But <!> can associate directly with the
certainty; there is no need to imagine the speech.
Usually, of course, the signs of the English writing system are made up of
letters strung together and we can make more or less regular connections
between the letters of the spelling and the phonemes of the pronunciation.
But we do not have to make this connection. Quite early in their schooling,
6 A survey of English spelling

competent readers acquire the skill of skimming over the written text with-
out referring to the corresponding spoken forms. Indeed it comes as a
surprise to learn later on, if they ever take a course in phonetics, that the
single written form of a morpheme such as {-edj or {photo} has several
different contextual pronunciations in their speech.
In describing such complex written signs we are quite free to ignore
speech altogether and simply say which letters go with which and, con-
versely, which fail to go with which. The letter <q> is always followed by
<u>, except in foreign names such as <Iraq>. That a similar-sounding <w>
never follows <q> is, however, no more interesting in such a description
than the fact that <t> or <p> never follows <q>. We can relate a final <-ed>
to [past tense} without being bothered by its different phonetic correlates:
/id/ (waited), Idt (begged) and tit (watched). We would note that <h> occurs
as the second member of some common clusters which seem to have the
status of units: <ch>, <gh>, <ph>, <rh>, <sh>, <th>, <wh> as in church,
ghost, physics, rhetoric, show, thin, when. Other clusters with <h> might
only be found at morpheme boundaries: <nh> in inhuman, <yh> in boy-
hood, <lh> in girlhood, etc. It would not be part of such a description to say
that in some accents of English the graphic contrast <wh> - <w>, as in
whine - wine, correlates with a phonemic contrast /hw/ - /w/, but not in all
accents. Nor would it be necessary to point out that <th> has the phonetic
correlate 151 initially in demonstratives (this, there, then, those,. . . ) and a
different phonetic correlate, IB/, initially in lexical words (thin, theorem, the-
sis, . . .). These would be just a few of the consequences of describing the
English writing system as if it were simply a system of graphic signs. There
are indeed several practical reasons why one might profitably study the way
letters and strings of letters are distributed: for instance, in studying the
purely visual aspect of reading, or in the technology of text processing.

2.2 LINKING S P E E C H T O WRITING: SOUNDS A N D L E T T E R S

Points of view
The more common alternative, however, is to describe English spelling by
exploring the relationship between spoken and written forms. We then
show how spelling relates to sound: how letters and strings of letters map
on to phonemes. This approach serves a wide range of practical needs and
interests: in the pursuit of spelling reform; in the technology of man-
machine communication; in theories of perception and in the teaching of
literacy, both to native English speakers and where English is a foreign
language. For foreign learners the spelling of an unfamiliar word is, for
better or worse, a reference from which to attempt its pronunciation. But
this wide range of practical applications brings problems. It is scarcely
surprising that when people with a particular interest set out to describe
 Literacy and English spelling 7

English spelling from their particular angle, the picture that emerges may
not be a fully objective record. Perspective may so easily shade off into
distortion. Since the aim of this present survey is to bring out whatever
regularities we can find in the English writing system, we ought to begin by
noting that the search for regularities and the business of describing them
is far from straightforward.

Classroom units
Many descriptions of English spelling conventions are bound up very
intimately with the teaching of reading to young children. The question
usually asked is: how is this string of letters pronounced and how is it
associated with a meaning? It is a very different question to look from the
other direction and ask: how is this string of phonemes (and the chunk
of meaning it associates with) represented in writing? The emphasis on
reading is natural enough. Literacy teaching is more often concerned with
the receiving of written messages than with sending them. Systematic
teaching of spelling can only be undertaken when pupils are well-launched
into reading. So, the educational process is naturally rather more con-
cerned with reading than with writing since reading is the basic means of
acquiring information. Consequently, if we want to take a fresh look at the
regularities of English orthography we need only approach it from the less
usual point of view: from speech to spelling. Parts of this study will explore
this less familiar approach.
The close ties between published descriptions of the English writing
system and the teaching of reading have had some unhappy consequences.
It might be expected that the units and rules which one would set up to
describe a system of correspondences between sounds and spellings would
naturally form a basis for the structural units to use in teaching - subject
to a grading for difficulty, productivity and other educational considera-
tions. The converse is certainly not true. It does not follow that units and
rules used by teachers and set up for classroom convenience will serve as
an adequate description. W e can take a few examples from a classroom
primer intended for foreigners.
Old-fashioned spelling primers often use units of correspondence
in their spelling drills which are incompatible with any consistent and
economical linguistic description. Morpheme boundaries may be ignored,
even when a complex spelling is clearly made up of two simple spellings
split by a morpheme boundary. There is no need to consider the <ayi> of
playing, saying, as a unit of spelling (as H i l l and U r e 1962) because it comes
about by the straightforward addition of <ing> to <play> and <say>.
Nonetheless they have <ayi> and similar complex strings as single units of
description: <eei> in agreeing, seeing; <eea> in agreeable; and even <ewa>
in chewable, sewable, beware, homewards, reward, prewar. Unlike the
8 A survey of English spelling

previous strings, which were at least unambiguous, <ewa> has a whole

range of different phonetic counterparts. But the apparent variety of these
counterparts is illusory. Once the morpheme boundaries in these words are
recognized, the variability is accounted for. The intention in such drills is
probably to warn the reader to look out for morpheme boundaries.
It is clear that units such as <ewa> would never figure in a properly
systematic description of how English spelling relates to speech. They are
mentioned here simply to show that there are no generally accepted and
explicit criteria for establishing the units of description. W e are unlikely to
find such criteria by hunting through the spelling books. If <ewa> is a unit,
why not <opi> or <ene>? Units such as these could only crop up in spelling
books destined for use by pupils who are already readers. Classroom-
oriented accounts of English spelling tend to highlight pitfalls for the
unwary reader, rather than to present a systematic description of the
regularities of English spelling.

Finding symbols for phonemes

Perhaps the greatest practical problem in dealing with spelling is that
people do not have a familiar and generally accepted way of tackling the
phonetic side of correspondences. There is plenty of evidence that native
speakers are passively aware of the phonemes of their particular language:
English speakers are aware that the commutation list pat, bat, mat, cat, hat,
rat, sat, etc. involves minimal distinctive changes. They know that what is
substituted here is a minimal distinctive unit of sound. Equally well they
know that what is substituted in flat, brat, slat, etc., is more complex and
involves a cluster of two such units. Yet there is no easy way for them to
refer to these units phonetically: no way of dealing with phonemes rather
than letters. There are, of course, phonetic symbols readily available which
can be used with a minimal knowledge of speech production, but few
people have access to them. Phonetics does not seem to play a significant
role in the training of primary-school teachers and many writers on spelling
deliberately avoid all reference to phonetic symbols and categories. It is
commonly held that to use I P A phonetic symbols for English phonemes
when writing about spelling is to add unnecessary technical complexity.
Cummings decided that since he was writing for:

teachers and others who are deeply engaged with English spelling
without necessarily being trained linguists, it seemed better to use a
system of pronunciation symbols that was less technical and perhaps less
intimidating.
(Cummings 1988: xxxi)

Taking this view, Cummings borrows the letter symbols used in W 3 N I D ,

slightly modified. This simply shows a preference for phoneme symbols
 Literacy and English spelling 9

more or less isographic with the spelling. There may well be some minimal
advantage in using 'a', 'a','a' rather than /«/, /ei/, /a:/ to woo the reader. But
surely anyone 'deeply engaged with English spelling' would be well-
advised to learn a little phonetics. Krapp (1925: 333) thought that Webster-
type symbols were a handicap in distinguishing between speech and
writing. He comments: 'This was the characteristic contribution of the
Webster dictionaries to the popular understanding of phonetics in America
and remains today the chief obstacle in the way of securing a treatment of
the subject in elementary books which is based on scientific principles'.
Certainly, it can be shown that this use of spellings as phoneme symbols
caused problems for the Hanna (1966) study of phoneme-grapheme corre-
spondences (see §2.8.5).
Since speech is often referred to by using common regular spellings as a
substitute for standard phoneme symbols, it follows that correspondences
are not usually stated in the direction /p/=<p>, but only in the direction
<p>=/p/. Writers usually refer the reader to a phoneme by quoting a letter
in a given keyword:

'th as in then\ 'th as in think', 'o as in hop', 'o as in hope'.

Such a formula takes the roundabout route: the phonente represented

by the spelling <th> in the word then. The lack of a simpler"technique for
dealing with phonemes helps to ensure that the usual direction of descrip-
tion is from writing to speech.

Sounds and letters: the misuse of 'vowel' and 'consonant'

(Note: in this section some of the references are drawn on simply to illustrate misleading
use of terminology and inadequate methods. Literacy studies are a minefield of misunder-
standings.)
The published literature on spelling is bedevilled by failure to distinguish
between speech and writing, between sounds and phonemes on the one
hand and letters on the other. ' V o w e l ' and 'consonant' are generally used
to refer to letters rather than sounds. This is bad practice because it is
bound to lead to misunderstanding. In many contexts the writer gets away
with it because the statement is ambiguous and could apply equally to
sound or spelling, but sooner or later the practice ends in confusion. Not
every statement that is true of vowel and consonant letters in English will
also be true of vowel and consonant sounds:

Monosyllables ending in b double it before suffixes beginning with

vowels if the sound (sic) preceding it is a single vowel (a, e, i , o, u, or y),
but not if it is a diphthong or a vowel plus r: cabby, webbed, glibbest,
bobbed, shrubbery; but daubed, barbed.
(Fowler 1926: 44)
10 A survey of English spelling

In the 2nd edition, Gowers quite rightly revised this to 'if the sound pre-
ceding it is a short vowel, but not if it is a long vowel or a vowel and r\
(Fowler 1965: 52).
But if we are indeed talking about sounds and not specifically about
letters, the vowel h-J of daubed is just as much a single vowel as Ixl or hi
and in phonetic terms it is not a diphthong. So, the <u> in quick is a
vowel letter representing a consonant /w/; the <w> in jaw is a consonant
letter used as part of a vowel spelling.
In adult literacy primers, which are naturally very careful to distinguish
statements involving sounds and those involving letters, we can find
l
oddities like: q is always followed by u and at least one other vowel, e.g.
quit' - [why 'one other vowel', when <u> here represents a consonant
/w/?]; 'the (f, 1, s, z) sounds are usually doubled when spelled at the end of
one-syllable words' - [the 'sounds' are not doubled]. It is admittedly more
cumbersome to be explicit. Where Fowler, in the first edition of his Modern
English Usage, writes about 'double and single consonants' (Fowler 1926:
554), Gowers alters the text to 'double and single letters for consonantal
sounds' (1965: 575), showing a willingness to pay the price of accuracy.
Similarly, Fowler calls those vowels after which <1> is not doubled before
a suffix, such as the /ei/=<ai> in failed (cf. equalled ) 'a compound vowel
sound, as ai, ea, ee, oi, ow, ur'. One might expect a 'compound vowel
sound' to be a diphthong, but what 'compound' refers to here is evidently
not the sounds but the spellings. Gowers in the 2nd edition alters this to 'a
long vowel sound made up either of two vowels or a vowel and a con-
sonant', specifically mentioning 'sound', but he is still using 'vowel' and
'consonant' to refer to letters (ibid.: 340). Gowers clearly felt that some-
thing was wrong in the original text, but did not quite make a clear dis-
tinction between letters and sounds.
Recognized authorities may use 'vowel' and 'consonant' to mean both
letters and sounds in the belief that this makes life easier for their readers,
but it docs make for some awkward statements. Cummings (1988: 212), for
instance, refers to 'the letter w, a consonant in wed but a vowel in dew' and
'the letter h . . . is a vowel only when it is the second letter in a vowel
digraph, as in ah, oh, uh, John, ohm, dahlia'. This is not perhaps so serious,
but taken at face value it means that there are two 'vowels' (as letters)
in wed but only one 'vowel' (as phoneme). This may well confuse some
readers. The problem surfaces again when Cummings classifies the vowel
of bite along with those of beat, bait, boat, boot, beaut-, as a 'long vowel
sound' and the vowels of oil and owl as diphthongs (ibid.: 203). Admittedly,
he is using the word 'sound' here 'for stylistic ease' to cover a more abstract
unit, which he calls a 'morphophone', but the vowel phoneme in bite is as
much a diphthong phonetically as those in oil and owl. The difference is
one of letters not of sounds. The 'long vowels' have single-letter spellings
and alternate with short vowels in the pronunciation of morphemes. So we
 Literacy and English spelling 11

have the pairs /ai/ - hi {define, definitive), lv.l - Id (serene, serenity), /ei/ -
Iztl (mania, manic), /so/ - hi (sole, solitude). The 'diphthongs' /au/ and /oil
cannot be spelt with single letters. If 'vowel' has to bridge both sound and
spelling, a solution might be to call the alternating vowels 'basic vowels'
with long and short values.
In this study 'consonant' and 'vowel' are used only to refer to sounds
or phonemes. When talking about letters, I have used 'consonant letter' or
'<C>' to refer to letters in the set <b, c, d, f, g, h, j , k, 1, m, n, p, q, r, s, t, v,
w, x, y, z> and 'vowel letter' or '<V>' to refer to letters in the set <a, e, i ,
o, u, y>. The letter <y> is a member of both sets: it is a consonant letter in
youth and a vowel letter in myth. Some consonant letters may be used as
auxiliary letters (see §2.6.5) in making up complex vowel spellings such as
<igh> sigh, <ow> now.
In the rest of this section we can profitably examine some of the dangers
inherent in the traditional misuse of 'vowel' and 'consonant'. The refer-
ences used are not necessarily important in the field of literacy studies.
They are chosen to highlight a problem that is endemic.

Making spelling an authority for pronunciation

Failure to realize that the relationship between letters and sounds is

purely conventional may have quite bizarre consequences. Because of the
'authority' of written language, people are tempted to use spelling
conventions as evidence of what goes on in speech production. Consider
the following speculation as to why the consonant letter is doubled after a
short vowel in latter but not after a long vowel in later.

Vowel sounds are of two species, long and short, i.e. those which, like the
a in 'father', can be prolonged, and those which, like the e in 'let',
cannot. The reason of this seems to be that in forming the short vowels
the throat is in such a position as to emit a large quantity of air, so that the
lungs are immediately emptied of wind; hence it is necessary to close or
partly close the mouth in order to lessen the expense of wind, if speech is
to continue; in other words, such vowels must always be followed
instantly by a consonant; from this cause is derived the rule that short
vowels are followed by double consonants, e.g. 'letter', and long vowels
by single consonants, e.g. 'later'. After a short vowel, as in 'letter', the
shutting of the mouth is done with some force and quickness, and is
clearly audible and distinguishable from the noise of the subsequent
opening. After a long vowel, however, as in 'later', the closing of the
mouth is clearly heard; the two sounds of shutting and opening after the
short vowel are represented by the double consonant. Hence the double
consonant has come to indicate the shortness of the vowel preceding it.
We have seen that after a short vowel the consonant follows instantly,
12 A survey of English spelling

and is made audible by means of this vowel preceding, e.g. 'nod'; but that
after a long vowel the consonant is not so made audible, therefore it
needs a vowel to follow it, e.g. 'no-de'. Hence it has come about that the
e following a consonant indicates that the preceding vowel is long.
These facts are important, because the practice of indicating the short
and long vowels in English in this way has been called irrational.
(Willis 1919: 34)

This piece of evident nonsense will serve as a lead towards recognizing less
obvious misunderstandings. The 'immediate emptying of the lungs', or the
existence of a final vowel in node are not of course real observations of
natural speech. Y e t , at the time Willis wrote, classical phonetics was in its
heyday and the real nature of speech production was well understood.
Even today people have problems in keeping apart the written and the
spoken medium. The heresy of thinking that mere spelling conventions are
a reliable encapsulation of phonetic facts still thrives.
This is the main reason why phonetics is wrapped in mystery for ordinary
people. Anyone who has an accent that does not fit the provisions of the
writing system will risk being told that their speech is 'wrong' or even that
their hearing is defective. Someone who does not contrast /hw/ and /w/ in
which and witch, may be told that they have 'lost' a consonant ' h \ or more
confusingly that they are 'not hearing' a consonant 'h'.
Those who practise phonetic methods in the teaching of reading might
even find it useful to be aware of phonetic detail below the level of phone-
mic contrasts. Some of their operations used in taking words apart and
putting them back together may have unforeseen phonetic consequences.
The idea of 'blends' and 'blending', where clusters are assembled from
their independently taught members, must take the subtleties of phonetic
context into account. If you associate a voiceless [p] with <pay> and a
voiced [1] with <lay>, it does not follow that the co-articulation of ipi and
Hi in <play> is phonetically straightforward, since the whole /pi/ cluster in
<play> is voiceless.
The failure to distinguish between sounds and letters may produce non-
problems and non-arguments which are ultimately used as evidence of the
complexity or irregularity of the English writing system. The following
example is all the more interesting in that it occurs in a book which tries to
include some explanation of what phonemes and allophones are, and which
has a section devoted to 'the surface structure of language' (Smith 1971:
30). In discussing 'phonics' as a method of teaching reading Smith says:

Here is a phonic rule that would appear to have impeccable antecedents:

final e following a single consonant [sic = letter] indicates that the pre-
ceding vowel [sic = phoneme] should be tense, as in hat and hate, or hop
and hope. A n d here is an instant exception: axe has a single consonant
but a lax (short) /a/ (while ache, which has a double consonant, takes a
 Literacy and English spelling 13

tense /a/). W e have the choice of admitting that our traditional rule is not
impervious to exceptions, or else we have to make a rule for the excep-
tions. One explanation that might be offered is that x is really a double
consonant, ks (and that ch is really a single consonant, it). But then
we are in the rather peculiar position of changing the notion of what
constitutes a single letter simply because we have a rule that does not fit
all cases. If we start to say that the definition of what constitutes a letter
depends on the pronunciation of a word, how can we say that the
pronunciation of a word can be predicted from its letters?
(ibid.: 166)
This has all the outward appearance of a rational argument, but it depends
purely on confusion between 'sound' and 'letter'. The message is that one
cannot rely on 'phonic rules' because the one discussed is apparently
flawed. The flaw is only in Smith's use of 'consonant' to refer to a conso-
nant letter. There is no problem with either axe or ache if 'consonant' is
used phonetically in the rule. The only exception in §Basic vocabulary
is the use of <-e> as a marker after <v> as in love, have. (See p. 121; see
also p. 220 for ache and, in the context of the 'short word rule', p. 133
for axe.)
There may be some advantage to be gained by this exploitation of the
confusion between sounds and letters. Smith finds it necessary to challenge
all phonographic regularities in the interests of his own theory of lexical
access in reading, which is based on sampling the text for visual features.
A 'regularity' is what you find; an 'irregularity' may be a regularity that
would be inconvenient for your purposes to find. This could be the reason
why Smith (1971) and Venezky (1967), writing about the very same data,
can come to quite different conclusions:

(a) : Some aspects of spelling are simply unpredictable, certainly to a

reader with a limited knowledge of word derivations, no matter how one
tries to define a spelling unit. A n example of a completely unpredictable
spelling to sound correspondence is th, which is pronounced in one way
at the beginning of words like this, than, that, those, them, then, these but
in another way at the beginning of think, thank, thatch, thong, theme, and
so on. There is only one way to tell whether th should be pronounced as
in this or as in think, and that is to remember every instance.
(Smith 1971: 167)

(b) : A final example of where form class identity is necessary for correct
pronunciation is in initial th. Functors beginning with this cluster have
the voiced inter-dental spirant /S/: the, then, this, those, while contentives
have the voiceless spirant /9/: thesis, thin, thumb.
(Venezky 1967: 92)
14 A survey of English spelling

If we take Venezky's point to be valid, as indeed we must, Smith is clearly

quite wrong to refer to the <th> spellings as 'completely unpredictable'.
There is nothing very esoteric about their distribution. Primary-school
teachers certainly need not brandish terminology such as 'functor' or
'contentive' at children to make them aware of the difference.

'Learning the alphabet-

Much of the confusion between sounds and letters derives from the
business of 'learning the alphabet' as a preliminary to reading. Alphabet
books for small children need careful design and material to do what they
are intended for: to bring out an awareness of letter-sound correspon-
dences. However good the pictures, no child will get very far with ' O is
for owl', ' O is for orange', 'S is for sugar', 'S is for ship'. The whole long
business of formal education should not begin with the rote learning of
manifest inconsistencies. This important first step in literacy does not seem
to attract much attention from educationalists. ( A notable exception is the
critical survey of alphabet books written by Suria Perera for Knowsley
Council Remedial Services.) 'Knowing the alphabet' involves learning a
letter's shape(s), its usual phonetic counterpart(s) and its 'name'. A great
deal of mystique attaches to the name of the letter, because in some vowel
correspondences 'the letter says its name'. The letter <a> 'says its name' in
mate but not in mat. Problems arise if the names of letters are taken to be a
basis for describing the writing system in relation to speech. For instance,
Follick (1965: 98), a leading spelling reformer, in a scheme designed to
teach English spelling and pronunciation to foreigners, groups vowels into:

'pure vowels': as in mate, mete, mite, mope, tube;

'impure vowels': as in mat, met, bit, mop, tub;
'diphthongs': as in brawn, brown, rood, stood;
'continental a': as in father.

This use of the term 'pure' obviously derives from the correspondence of
the 'name ' of the letter with its 'value'. By comparison the short vowels
are 'impure' since name and value are out of step. There is an irony here
because, in Mont Follick's reformed spelling scheme, the simple vowel
letters have their short value and the long vowels require digraphs. The fact
that they 'say their name' is not exploited. What is wrong with these cate-
gories is not so much that they are in drastic conflict with the phonetic use
of the terms 'pure' and 'diphthong'. It is simply that they would form a very
confusing basis for teaching English pronunciation to foreigners.
Apart from how the letter 'says its name', the other criterion used in
this classification is whether the spelling consists of one letter or two. The
term diphthong properly refers to vowel sounds, not vowel spellings. A
diphthong is a vowel glide within a single syllable, that is, a vowel in which
 Literacy and English spelling 15

there is a perceptible change of quality during its production. What

is common to brawn, brown, rood, stood is that the vowels are represented
by digraphs: strings of two letters. The only one of the four to contain a
diphthong is brown. In later stages of Follick's scheme for reformed spelling
the 'pure'/'impure' categories are dropped but the confusion of sounds and
letters persists to some extent:

Diphthongs : ei ii ai ou iu aa au aw.
Digraphs: ch sh th dh kh zh.
(Follick 1965: 201)

Some of these 'diphthongs' do not, of course, represent diphthongs at all

but, in the phonetic sense, pure vowels. The spelling <aa> represents /a:/,
<au> usually represents b-J. What Mont Follick really has here are a set
of vowel digraphs and a set of consonant digraphs. The ingrained idea that
a diphthong is a vowel spelt with two letters persists throughout the liter-
ature on reading and spelling and is a prime source of confusion (cf. the
example quoted from Fowler at the beginning of this section). Most
people would be aware that the element <graph> has to do with writing,
but unfortunately the element <phthong> is so infrequent in English words
that it carries no association with 'sound'.

2.3 DIVERGENCE

Departures from one-to-one correspondence

The alphabetic principle, in theory at least, requires not only that a given
phoneme is represented by a constant symbol but also that the symbol
involved does not represent other phonemes. This requirement is referred
to in phonological theory as 'biuniqueness'. A s we have already seen, the
English writing system is by no means entirely based on the alphabetic
principle: to a very large extent it seeks to secure a constant spelling for
a given morpheme. Partly because of this, partly because of large-scale
borrowing from other languages, and partly because of the effect of sound-
changes, the phoneme-grapheme correspondences of English usually fall
far short of biuniqueness. But individual phonemes and individual
graphemes vary considerably in the extent of their divergence from
biuniqueness. This is fairly obvious to any speller: the spelling of /p/ is more
straightforward than the spelling of /k/, for instance.
We shall use the term 'divergence' (Haas 1970: 51) to cover this aspect
of spelling. We may find divergence on the phonic side of a correspon-
dence: <th>=/6/ and <th>=/87; or on the graphic side: /f/s<f>, /f/=<ff> and
/f/=<ph>. More often than not we find divergence on both sides: /i:/=<ea>,
/i:/=<ee>, and <ea>=/i:/, <ea>=/e/, etc.
16 A survey of English spelling

The terms conditioned variant' and 'competing variant' would seem to

be useful in differentiating between real and apparent divergence.
Consider the different spellings of Ipl and the different spellings of Ixl. The
<pp> spelling is largely predictable either as consonant-letter doubling to
mark a previous short vowel in §Basic words {apple, copper, supper) or as
prefix assimilation in §Latinate words (approve, apprehend, appreciate). If
we can state general rules for the distribution of two alternative spellings,
then they can be regarded as conditioned variants in complementary
distribution. If we look, however, at the spellings of ia-J in SBS, we do not
find that the distribution of the spellings can be described by rule to any
marked extent. The choice between, say, <ir> and <ur> as spellings of !s\i
cannot be explained by context. W e might tentatively call them competing
variants. A n y divergence requires more effort by the reader/writer, but
competing variants are clearly much more troublesome than conditioned
variants. They make the task of differentiating word-shapes more difficult,
especially for the writer. For the reader, the cost might well be justified if
it gives a different shape to homophones such as seen - scene (see §5.3).

The temptation to overstate irregularities

The divergence found in English spelling is often overstated. This is

largely because people fail to realize that some spellings are conditioned
variants which do not in effect compete with each other. Henderson and
Chard (1980: 111) in an article devoted, ironically, to the reader's implicit
knowledge of orthographic structure, provide a listing of what they
consider to be the 'mapping options' for the spellings and sounds of the
word fake (<fake>=/feik/) reproduced below as Table 4. This is supposed
to represent the choices open to the reader. The correspondences they use
derive from Hanna (1966).
Commenting on this table they note that the reader has 19 options in
putting together the string of phonemes and the writer has 27 options in
putting together the string of letters. They conclude that correspondences
are more ambiguous in the phoneme to grapheme direction. This may well
be true, but the assumption is falsely made. The mapping of letters on to
phonemes and vice versa is seen as a series of 'pathways'. They find that
there is also asymmetry in the relative frequency of a given pathway: Ikl is
the regular sounding of <k>, while <c> is the regular spelling of ikl. They
go on to admit that 'some of these asymmetries of choice can be reduced
by applying other constraints, ranging from position within the syllable up
to more complex linguistic considerations'. That is so very obviously the
case that it is difficult to imagine any valid use to which such a table might
be put.
 Literacy and English spelling 17

Table 4 'Mapping options' for fake <fake>=/feik/ (as in Henderson and Chard)

Grapheme-phoneme correspondences Phoneme-grapheme correspondences

Grapheme Phonemic Example Phoneme Graphemic Example

options options

<f> IV fake IV <f> fake

<ph> phone
<a> led fake <ff> buff
la'J arm <gh> rough
I'vJ eat
Id many /ei/ <a-e> fake
Iml aisle <a> angel
Isd coat <ai> aid
b:l all <ai-e> aide
hi canal <aigh> straight
/u:/ beauty <au-e> gauge
<ay> way
<k> IVJ fake <e> cafe
silent knee <ea> break
<ei> vein
<e> silent fake <eigh> weigh
led there <et> beret
I'vJ feel <ey> they
/e/ end
/ai/ eye Ikl <k> fake
Im'J sleuth <c> ore
/a/ angel <cc> occupy
<cch> saccharine
<ch> echo
<ck> back
<cq> acquire
<kh> khaki
<qu> bouquet
<sc> viscount

What is wrong here is the curious disregard for context. This takes two
forms. A 'grapheme' is taken to be a single letter, as if reading and spelling
were done with units of one letter at a time. This seems to entail that a
single letter <a> in the slot <f-ke> is somehow able to be confused with
<ea>=/i:/ in eat, <ai>=Wai/ in aisle, <oa>=/au7 in coat, and <eau>=/ju:/ in
beauty and so denies these letter complexes the status of units. It also
assumes that the <e>-marker is not read as part of the correspondence
<a-e>, though they allow this in the other direction, speech-to-text. This
brings us to the other way in which context is here disregarded. The
authors ignore the fact that correspondences may be conditioned variants.
The final single <e> of <fake> cannot in any way form a correspondence
with the vowels of there, feel, end, eye, sleuth or angel. The 'silent' <k> is
limited to the initial cluster <kn>. The same point can be made for most of
18 A survey of English spelling

the correspondences cited for the speech-to-text direction. For instance,

the <ff> and <ck> options would only occur after a short vowel. The <gh>
option for III does not occur initially. Here we are on the familiar ground
of Shaw's comic suggestion that fish may equally well be spelt *<ghoti>
(see p. 21). In fact, in English the spelling <fake> can only be a spelling of
/feik/. O n the other hand, there are a few marginal ways in which the utter-
ance /feik/ might otherwise be spelt:

1 as *<phache> (cf. phase, and the discussion of ache, p. 472;

2 as *<feak> which is only feasible on the remote analogy of break);
3 as *<faik> (though<ai> does not seem to occur before <k>, except in a
few names such as Aiken, Blaikley, Craik, Raikes).

A l l in all, a better example of the difference between apparent and real

divergence could hardly have been given.
Disregard for context is a recurring problem in literacy studies. It lies at
the heart of Smith's treatment oiaxe (p. 12), <th-> (p. 13), |-ed) (p. 20) and
<ho-> (p. 33).

2.4 L E X I C A L SPELLING

Keeping the spelling of a morpheme constant

The English writing system is not simply concerned with mapping
phonemes on to letters. T o a large extent it tries to offer the reader
a constant spelling for a morpheme in spite of the varying pronunciation
of the morpheme in different contexts. Children are taught both a long
and a short phonemic value for the simple vowel letters. In spite of the
very considerable present-day phonetic differences in pronunciation,
the pairs /ei/ - Ixl (sane, sanity), /ai/ - hi (mime, mimic), /au/ - M (cone,
conical), I'vJ - Id (diabetes, diabetic) have a constant vowel letter. Only
in the pair /au/ - / A / (pronounce, pronunciation) do the vowel letters
used in the spelling (<ou> - <u>) sometimes vary to reflect the surface
difference. But even here there are long-short differences such as
South, Southern where the morpheme keeps its spelling identity. Similarly
the different correspondences /s/=<c>, /k/=<c>, and /J7=<ci> allow
a constant morphemic spelling in words such as electricity, electrical,
electrician. The past tense and participle morpheme {-ed} in regular
verbs (ignoring forms such as spelt, spent, dreamt) has a constant <-ed>
spelling in spite of the phonemic variation /id/ (waited), Idl (warned) and
lil (watched).
There is very similar allomorphic variation in the noun plural and 3rd
person singular present tense of the verb with its variants lizl (watches),
Izl (warns) and Is/ (waits). The two sets of phonetic variants are often
cited together as equivalent examples of morphemic spelling. If we look
 Literacy and English spelling 19

closely at the details, however we find in fact that they are dealt with rather
differently in the writing system (table 5).

Table 5 English plural and past tense inflections

{-ed} ending {-(e)s) ending

1 /-id/ matted 1-izJ masses

2 /-id/ mated l-\zl maces
3 /-V hopped l-sl (Ho'hoppes)
4 /-V hoped /-s/ hops, hopes
5 /-d/ planned 1-2] [Ho'plannes)
6 /-d/ planed 1-21 plans, planes
7 1-61 hurried 1-21 hurries
8 1-6/ radioed l-zl radios
9 /-d/ vetoed l-i] vetoes or vetos

Table 5 gives examples of each suffix in its three phonemic variants. The
words provided as contexts include a preceding long vowel followed by a
single consonant, a preceding short vowel followed by a single consonant,
and the spelling variation found after <o>. The past tense has a constant
<ed> spelling across all these contexts which is only varied by merging with
a final <-e> in the long vowel words (e.g. hope - hoped). The marking of
the long vowels is also done indirectly by using <C>-doubling to mark a
previous short vowel - absence of doubling marks the long vowel. This is
not done in the noun plural suffix, so there are no equivalents in rows 3 and
5 to <hopped> and <planned>, that is *<hoppes> and *<plannes>. Nor is
there a variant spelling *<vetod> to spoil the uniformity of <-ed>. The <e>
of <-oes> and oies> is best treated as a marker of the long vowel.
Grammarians refer to the two inflections as the -s form and the -ed form
and this is in accordance with the traditional spelling rule that one adds
<-s> and <-ed> respectively to the base form.
If we try to use a consistent <-ed> morphemic spelling in chopping up
words into correspondences without any overlap, this lack of symmetry
becomes apparent. In scribbles we have hV=<le> and in scribbled we have
hl/=<\>; in hopes we have <e>-marking of the vowel as /au/=<o..e>,
/p/s<p>, /sfe<s>, but in hoped we have bul=<o>, /p/=<p>, /t/=<ed>,
while the long vowel is marked by absence of <C>-doubling. In a process
description, and indeed for any human speller, this is catered for by
'dovetailing' so that <hope>+<ed> becomes <hoped>, <eight>+<ty>
becomes <eighty>, etc.
Not all past tense and participle forms have the <ed> spelling. The group
burn, dwell, learn, smell, spill, spoil have two alternative forms (burned
/d/s<cd>, burnt /t/=<t> etc.). Some irregular verbs have two such forms but
with a different vowel in the III form as well (dream, dreamed and dreamt
/dremt/, lean, leaned and leant /lent/), or just one form with a different
20 A survey of English spelling

vowel and /t/=<t> (deal, dealt; creep, crept; buy, bought). The group bend,
build, lend, rend, send, spend, have a single form with /t/=<t> replacing the
!d/=<d> of the base form. The group sell, tell, say, hear have a single
/d/=<d> with a stem vowel change. These do not present a problem in
spelling since the <t> or <d> spellings only occur where the stem vowel
changes or where the /t/ suffix irregularly follows a vowel or voiced conso-
nant or both. There should be no temptation to write *<creped> (for crept)
or *<bened> for bent.
The system also maintains a constant morphemic spelling in instances
where an underlying phoneme does not surface and the letter in question
has a zero phonetic counterpart - cf. sign with /0/=<g> and signature with
/g/=<g>. The <g> in sign is an example of what is here called an 'inert
letter' (see §2.6.5 p. 41).
In arguments about the English writing system, a good head of steam can
easily be generated simply by regarding morphemic spellings as failed
phonemic spellings. For instance: 'it is often impossible to know for certain
which phonic rules apply. The rule that specifies how to pronounce ph in
telephone falls down in the face of haphazard, or shepherd, or cuphook'
(Smith 1971:168). These alarming lapses can easily be cured by recognizing
a morpheme boundary. Smith considers that there are three important
aspects of spoken language that 'our writing system does not even pretend
to represent'. They are, in his view, intonation, syntax (e.g. whether permit
is a noun or a verb), and 'finally, and perhaps surprisingly, spelling does not
really attempt to represent sounds at all, but rather phonemes'. Reluctantly
accepting that the English writing system does seem to represent
phonemes, he goes on to illustrate its failure to do so consistently:

Sometimes a single letter that we may think has only one sound is
pronounced as three different phonemes [sic]; the s at the end of a
plural noun or present tense verb may be pronounced /s/ as in 'walks',
Izl as in 'runs', or /ez/ [sic] as in 'judges'. Similarly the past tense ed may
be pronounced It! as in 'walked', Id/ as in 'charged', or /ed/ [sic] as in
'landed'. None of these distinctions is indicated by the spelling, and they
are generally disregarded in 'phonic rules '.
(Smith 1971: 168)
Here again, one can only note that regularity is in the eye of the beholder.
Smith assumes that the only function of spelling is to represent phonemes.
Then he criticizes the writing system when it chooses to represent
morphemes instead of phonemes.

Processes and underlying forms

Keeping the spelling of a morpheme constant entails some graphic units

having several different phonetic counterparts. Usually this variation is
 Literacy and English spelling 21

catered for by regular correspondences between letters and phonemes

already provided in the system. So the /s/ and Ikl values of <c> in electrical,
electricity are not only used to give morphemes such as electric a constant
spelling, they are in use throughout the system (cat, city). However, deriva-
tion processes do produce correspondences only found in allomorphic
variation and not elsewhere. Sound changes, such as palatalization, which
operate in word-formation, highlight the difficulty of working with surface
phonemes in setting up simple correspondences. Palatalization, in English
word-formation, is where an alveolar consonant has become palato-
alveolar by merger with a following HI or 1)1 or with a palato-alveolar
consonant (///, izl, lp, /dj/) and where special correspondences have to be set
up to capture the result. The word picture was probably pronounced as
/piktju:r/ in Early Modern English; now it is pronounced /pikfa(r)/. The
spelling reflects the earlier pronunciation, which would give the straight-
forward correspondences /t/s<t>, /ju:/=<u>, /r/s<r>. But how are we to
cope with the present-day Ipl Setting up /f/=<t> is inevitably going to add
to the complexity of the description.
We could of course avoid the problem and say that the spelling system
does not pretend to represent the present-day pronunciation: that as a
system it represents more or less the pronunciation of Early Modern
English. The modern pronunciation can be derived from the earlier 'under-
lying' form by a number of fairly simple rules. This type of description
requires phoneme-encoding (if and when phoneme-encoding takes place in
any stage or type of reading activity) to be a two-stage process. W e would
interpret <picture> as //piktju:r// and then apply some rough-and-ready
sound changes by which //ill+fljll became lp and the Ixl before the boundary
was lost. A t first sight this looks like an absurdly complex way of going
about it. However, the sound changes involved in this process are not
really long-dead and fossilized, they are still an active and normal part of
the speaker's control of allegro forms. In quick speech, palatalization occurs
across word-boundaries to produce might you as /maitfo/ from /mait/+/ju:/
and speakers of SBS and similar accents are aware of a word-final latent llxll
which will surface only before a vowel in a close-knit phrase (picture it).
We can note in passing Bernard Shaw's comic suggestion that fish might
equally well be spelt *<ghoti> (by analogy with rough, women, and nation).
It is not just that these correspondences never occur in the phonetic
contexts in which he has put them (Haas 1970: 57, Stubbs 1980: 51). The
correspondence /Jfe<ti> only occurs as palatalization in §Latinate words
and fish is clearly §Basic.

An optimal orthography? The views of Venezky and Chomsky and Halle

It is possible, then, to describe the English writing system as a system of

greater inherent simplicity by going beyond the surface correspondences in
Random documents with unrelated
content Scribd suggests to you:
 Chapter Eighteen.
By Day and by Night.

She put up both her small white hands as though to stay

the torrent of passionate words which I poured forth; but I
grasped her wrists and held her to me until I had told her
all the longings of my soul.

What she had said had caused me a stab of unutterable

pain, for my conscience was pricked by the knowledge that
I had for a brief moment forsaken her in favour of Yolande.
But she could not know the real truth. It was only by her
woman’s natural intuition that she held me in suspicion,
believing that by my neglect to write I had proved myself
attracted by some member of that crowd of feminine
butterflies who flit through the embassies, showing their
bright colours and dazzling effects.

At last she lifted her face, and in a low, faltering voice said:

“I do not wish that we should part, Gerald. I have no one

but you.”

“And God knows—God knows, darling, I have no one but

you!” I cried brokenly; and as I uttered these words she
cast her arms about my neck, clinging to me, sobbing, with
her face lying close against my breast.

“My darling—my own darling!” was all I could murmur as I

kissed away the tears that rained down her checks. I could
say nothing more definite than that.

“You will not be false, will you?” she implored at last. “You
will not break your promise, will you?”
“I will never do that, dearest,” I assured her. “I love you,
upon my word of honour as a man. I have loved you ever
since that day when we first met at the house-party up in
Scotland—the night of my arrival when you sat opposite me
at dinner. Do you remember?”

“Yes,” she answered, smiling, “I remember. My love for you,

Gerald, has never wavered for one single instant.”

“Then why should you be unhappy?” I asked.

“I really cannot tell,” she answered. She turned her face,

and I saw that there was a shadow across it, as though the
sunshine of her life had gone behind a bank of cloud. “All I
can compare this strange foreboding to is the shadow of an
unknown danger which seems of late to have arisen, and to
stand in a wall of impenetrable blackness between us.”

“No, no!” I hastened to urge, “the sweet idyll of our

blameless love must be preserved. That fancy of yours is
only a vague, unfounded one.”

She shook her head dubiously.

“It is always with me. During my long, solitary wanderings

here I think of you, and then it arises to overshadow me
and crush out all my happiness,” she said in a tone of
sorrow.

“Your life is dismal and lonely here,” I said. “You’ve become

nervous and melancholy. Why not have a change? Persuade
your aunt to bring you to Paris, or, if not, to some place
near, where we may meet often.”

“No,” she replied in a harsh tone. “My presence in Paris is

not wanted. You are better without me. You must leave
England again to-morrow—and you must forget.”
“Forget!” I gasped. “Why?”

“It is best to do so,” she faltered with emotion. “I am

unfitted to become your wife.”

“But you shall—you must!” I cried. “You have already given

me your promise. You will not desert me now!”

She made no response. I pressed her again for an answer,

but she maintained silence. Her attitude was one of firm
resolve, and gave me the distinct impression that she had
gained some knowledge of the reason of our brief
estrangement.

“Tell me the reason of your sudden disbelief in my

declarations,” I urged, looking earnestly into her eyes.
“Surely I have given you no cause to regard our love as a
mere irresponsible flirtation?”

“I have no reason to disbelieve you, Gerald,” she answered

seriously; “yet I recognise the impossibility of our
marriage.”

“Why is it impossible? We are both controllers of our own

actions. You will not remain here with your aunt all your
days?”

“We may marry, but we should not be happy, I feel certain.”

“Why?”

“Because if I were your wife I could not bear to think you

were out each night dancing attendance upon a crowd of
foreign women at the various functions which you are in
duty bound to attend.”
I smiled at her argument. Ignorant of the world and its
ways, and knowing nothing of Society beyond that gossiping
little circle of tea-drinkers and tennis-players which had its
centre in the town of Fakenham, and had as leader the
portly wife of the estimable incumbent, she saw herself
neglected among the brilliant crowd in Paris as described by
the so-called “Society” papers.

I hastened to reassure her, and as we strolled on through

the wood and, following the meandering of the river,
emerged upon the broad grass-lands before Sennowe Hall, I
used every argument of which I was capable in order to
dispel her absurd apprehensions. My protestations of love I
repeated a hundred times, striving to impress upon her that
I was actually in earnest; but she repelled me always, until
of a sudden I halted beneath the willows, and, placing my
arm around her slim waist, narrowly girdled by its crimson
ribbon, I drew her again to me, saying:

“Tell me, Edith, plainly, whether or no you love me. These

cold words of yours have struck me to the heart, and I feel
somehow that in my absence you have found some other
man who has your gratitude, your respect, and your love.”

She raised her hand, as though to stay the flow of my

words.

“No, no!” I went on passionately. “You must hear me, for

you seem to be gradually slipping away from me. You must
hear me! Cast away this cold sweetness that is enough to
madden any man. Give me a right to your love; give me a
right to it! You cannot be indifferent to such a love as mine
unless you love someone else.”

“Stop!” she cried, moved by a sudden generous impulse. “I

love no one else but you.”
“And you admit that you still love me? You will be the same
to me as before?” I cried eagerly.

“If you will swear that there is no thought of another

woman in your heart,” she answered seriously.

A pang of conscience smote me; but inwardly I reassured

myself that all the fascination of Yolande had been dispelled
and that my love was free.

“I swear,” I said; then slowly I bent until our lips touched.

Hers met mine in a fierce, passionate caress, and by that I

knew our compact was sealed.

“I admit,” she said, “that my instinct, if it were instinct, was

wrong. You have, after all, proved yourself loyal to me.”

“And I shall remain so, darling,” I assured her, kissing her

again upon the brow. “For the present you must be content
to remain with your aunt; but nevertheless, try to persuade
her to come to Paris. Then we can spend many happy days
together.”

“She hates the Continent and foreigners,” answered my love

with a brightening smile. “I fear I can never persuade her to
move from here. She went to Switzerland twenty years ago,
and has never ceased condemning foreign travel.”

“If she will not come, then why not engage a chaperon? You
surely know some pleasant woman who would be pleased to
have a holiday jaunt.”

“Well,” she answered dubiously, “I’ll try, but I fear Aunt

Hetty will never hear of it.”
“The life in the profound stillness of that house and the rigid
seclusion from all worldly enjoyment are producing an ill
effect upon your health, darling,” I said presently. “You must
have a change. It is imperative.”

But she only sighed, smiled rather sadly, and answered in a

low voice:

“The quietness of life here is nothing to me, as long as I am

confident that your love for me is just the same as it was
when you first told me the secret of your heart.”

“It is,” I assured her—“it is, darling. I love you—and you

alone.”

There was an instant’s hesitation, and then her arms stole

gently to my neck, and her lips were pressed to the cheek I
bent to them, but only for a second; then my lips were
upon hers, clinging to them softly, passionately; and in
those moments of ecstasy I drew my soul’s life from that
sweet mouth.

Heedless of time, we stood there in each other’s embrace,

repeating our vows of love and devotion, until the sun went
down behind the low hills beyond Raynham, and the broad
pastures were flooded by the purple glow of the dying day.
Happy and content in each other’s affection, we were
careless of the past, and recked not of the future. Edith
loved me, and I wished for naught else in all the world.

Now as I sit committing this strange story of my life—this

confidential chapter in the modern history of Europe—to
paper, I recall every detail of those hours we spent down by
the riverside, and contrast it with the curious events which
followed—events which were so strange as to be
inexplicable until the ghastly truth became revealed. But I
loved, and my affection was reciprocated. That surely was
sufficient, for I knew that I had gained the purest, most
beautiful, and sweetest woman I had ever met.

At last the fading sunlight impressed upon us the fact that

the dinner-hour was approaching; and, knowing Miss
Foskett’s punctuality at meals, we were compelled to strike
along the footpath over Dunham Hill, and take the shortest
cut across the fields through the little hamlet of Gateley,
and thence by a grass-grown by-road back to Great
Ryburgh, where we arrived just as the gong sounded.

When we re-entered the dining-room, Aunt Hetty glanced at

us keenly, as though she wished to make some sarcastic
comment upon our long absence; but our pleasant
demeanour apparently silenced her, and she contented
herself by taking her seat at table and inquiring of me if I
had had a pleasant walk, and whether I found the country
agreeable after the dusty boulevards of Paris.

“Of course,” I answered, “I always find England charming,

and I’m very frequently homesick, living as I do among
foreigners always. But why don’t you come abroad for a
month or so, and bring Edith?”

“Abroad!” screamed the old lady, holding up her hands.

“Never! I went to Lucerne once, and found it horrible.”

“But that was some years ago, was it not? If you went now,
you would find that travelling has greatly improved, with a
through sleeping-car from Calais to Basle; hotels excellent,
and food quite as good as you can obtain in England. During
the past few years hotel-keepers on the Continent have
awakened to the fact that if they wish to be prosperous they
must cater for English visitors.”

“Oh, do let us go abroad, aunt!” urged Edith. “I should so

much enjoy it!”
“Paris in summer is worse than London, I’ve heard, my
dear,” answered Miss Foskett, in her high-pitched tone.

“But there are many pretty places within easy reach of the
capital,” I remarked. “Edith speaks French; therefore you
need have no hesitation on that score.”

“No,” said the old lady decisively, “we shall not move from
Ryburgh this summer, but perhaps next winter—”

“Ah!” cried Edith joyously. “Yes, capital! Let us go abroad

next winter, to the Riviera, or somewhere where it’s warm.
It would be delightful to escape all the rain and cold, and
eat one’s Christmas dinner in the sunshine. You know the
South, Gerald? What place do you recommend?”

“Well,” I said, “any place along the Riviera except, perhaps,

Monte Carlo.”

“Monte Carlo!” echoed Aunt Hetty. “That wicked place! I

hope I shall never see it. Mr Harbur told us in his sermon
the other Sunday about the frightful gambling there, and
how people hanged themselves on the trees in the garden.
Please don’t talk of such places, Edith.”

“But, aunt, there are many beautiful resorts in the

neighbourhood,” her niece protested. “All along the coast
there are towns where the English go to avoid the winter,
such as Cannes, Nice, Mentone, and San Remo.”

“Well,” responded Miss Foskett with some asperity, “we

need not discuss in August what we shall do in December.
Ryburgh is quite pleasant enough for me. When I was your
age I employed my time with embroidery and wool-work,
and never troubled my head about foreign travel. But
nowadays,” she added with a sigh, “I really don’t know what
young people are coming to.”
“We’ve advanced with the times, and they’ve emancipated
women in England,” responded Edith mischievously,
glancing merrily across at me.

Miss Foskett drew herself up primly, and declared that she

hoped her niece would never become one of “those dreadful
creatures who ape the manners of men;” to which my love
replied that liberty of action was the source of all happiness.

Fearing that this beginning might end in a heated

argument, I managed to turn the conversation into a
different channel.

“If all we read in the newspapers is true, it would seem,”

observed Aunt Hetty presently, “that you diplomatists have
a most difficult task in Paris.”

“All is not true,” I laughed. “Much of what you read exists

only in the minds of those imaginative gentlemen called
Paris correspondents.”

“I suppose,” remarked Edith, smiling, “that it is impossible

for either a diplomatist or a journalist to tell the truth
always.”

“Truth, no doubt, is all very well in its place, and now and
then in diplomacy, but only a sparing use should be made of
it as a rule,” I answered. “But there should be no waste.
Only those should be allowed to handle it who can use it
with discretion, and who will ladle it out with caution.”

“Mr Ingram, I am surprised!” interrupted Miss Foskett,

scandalised.

“It is our creed,” I went on, “that truth should be always

spoken in a dead or foreign language, no home-truths being
for a moment tolerated. Now think what a happy land this
England of ours would be if only we were not so wedded to
the bare, cold truth! Suppose for its own good purposes our
Government has thought right to make a hasty dash for the
back seats in the international scrimmage, and to adhere to
them with all the tenacity of a limpet, why, for all that,
should the Opposition journals blurt out the fact for our
humiliation, when by a few deft scratches of the pen the
leader-writer might easily make us believe that no back
seat had ever in any circumstances been occupied by
Britain, and that the nose of the lion had never been pulled
out of any hole into which it had once been inserted? The
itch for truth is, judged from a diplomatists point of view,
responsible for the ruin of our policy towards our enemies.”

“Shocking, Mr Ingram! I’m surprised to find that you hold

such views,” said Miss Foskett in a soured tone; while Edith
laughed merrily, declaring that she fully agreed with my
argument, much to her aunt’s discomfiture.

The old lady loved the harsh truth as propounded by the

precisionist.

And so the dinner proceeded, each of us vying with the

other to dispel Aunt Hetty’s deep-seated prejudices and
narrow-minded views of the world and its ways.

Coffee was served in the drawing-room, where Edith went

to the piano and sang in her sweet contralto several of my
favourite songs, after which, at an early hour, as was usual
with the household at Ryburgh, we all retired.

To sleep so quickly after dinner was to me impossible;

therefore, on gaining my room, I lit a cigar, and, taking a
novel from my bag, sat reading. The book proved
interesting; and time had passed unnoticed, until of a
sudden my attention was attracted by the sound of low
voices. I listened, glancing at the clock, and noticed that it
was nearly two in the morning.

A suspicion of burglars at once flashed across my mind. I

blew out my candles, so as not to attract attention,
noiselessly opened the wooden shutters before my window,
and cautiously gazed out. The lawn, garden, and wide
sweep of country beyond lay bathed in the bright
moonlight, and at first I distinguished no one. Peering
down, however, until I could see the path running in the
shadow just below my window, I distinguished two figures
with hands clasped, as though in parting. I looked again,
scarce believing my own eyes. But I was not mistaken. One
figure was that of a woman, her dark cloak open at the
throat, revealing her white dress beneath; while the other
was the tall dark figure of a man in a long black overcoat,
the collar of which was turned up as though to conceal his
features. Even though they stood together in the dark
shadow, the astounding truth was plain to me. The woman
who had kept that midnight assignation was Edith Austin,
my well-beloved.

My heart stood still.

 Chapter Nineteen.
Whispered Words.

The revelation held me rigid. I stood there, peering down,

watching their movements, and straining my ears to catch
the whispered words. As I feared to open the window lest
the noise should attract them, I could do no more than
remain a spectator of Edith’s perfidy. To me it seemed as
though she had been walking with him, and he had
accompanied her back to the house. As he held her hand,
he was bending, whispering some earnest words into her
ear. She did not attempt to withdraw; indeed, it was
apparent that she was not unwilling. The conclusion to be
made was that they were lovers.

Reader, can you imagine my feelings at this astounding

discovery? Only six hours before we had stood beside the
river, and she had vowed that for no man save myself had
she any place in her heart; yet with my own eyes I was
watching her while she believed me sleeping in calm
ignorance of her movements. That she had been walking
with him was apparent, because of the shawl she wore
wrapped about her head; while the fact that the stranger
carried a stout stick showed that he had walked, or was
about to walk, a considerable distance. Because his hat was
drawn over his brow and his coat collar turned up, I could
not see his features. To me, as he stood there, he appeared
to be slightly round-shouldered, but, nevertheless, a
strongly built fellow, seemingly rather above the average
height.

How long she had been absent from the house I could not
tell. Her light step across the lawn had not attracted my
attention. Only his low, gruff voice on their return had
caused me to listen. There was a French window near where
they were standing, and it was evident that by means of
this she had secretly left the house.

Across the moon there drifted a strip of fleecy cloud, hiding

the lawn and garden for a few moments; then suddenly all
became brilliant again, and, looking down, I saw that she
had moved, and was unconsciously in the full white light. I
caught sight of her countenance, so that her identity
became undeniable. He was urging her to speak, but she
remained silent. Again and again he whispered into her ear,
but she shook her head. At last she spoke. I heard what she
said, for I had contrived to raise the sash an inch or two.

“Very well, I promise,” she said. “He leaves to-morrow.”

“And you will not fail?” asked the gruff voice of her
clandestine companion.

“No. Adieu!”

And as I watched I saw his dark figure striding away in the

full moonlight across the lawn. He did not glance back, but
went straight over to the belt of elms on the left, and a few
minutes later was lost to view, while the woman I loved had
apparently re-entered the house by the dining-room
window, and was creeping silently to her room.

The one thought that gripped my heart and froze my senses

was that Edith was false to me. She had a lover whom she
met at dead of night and with whom she had a perfect
understanding. She had made him a promise, the fulfilment
of which was to take place when I had left. Had such things
been told to me I would not have believed them, but I had
seen with my own eyes and heard with my own ears. The
truth was too terribly plain. Edith, the one woman in the
world whom I had believed to be pure, honest, and upright,
was false to me. I saw it all as I reclosed the shutters, relit
the candles in their old silver sconces, and paced that
ancient bed-chamber. The reason of her attempt to evade
me and to withdraw her promise of marriage was only too
apparent. She, the woman whom I loved and in whom I had
put all my faith, had a lover.

As I reflected upon our conversation of that afternoon I saw

in her uneasiness and her responses a self-condemnation.
She dreaded lest I should discover the secret within her
breast—the secret that, after all, she did not love me. The
dark silhouette of that man standing forth in the brilliant
light of the moon was photographed indelibly upon my
memory. His outline struck me as that of a man of shabby
attire, and I felt certain the hat drawn down upon his face
was battered and worn. Indeed, I had a distinct conviction
that he was some low-born lout from the neighbourhood—a
conviction aroused, I think, by her announcement that I
was to leave on the morrow. She would have freedom of
action then, I reflected bitterly. And her promise? What, I
wondered, had she promised? The fellow had evidently
been persuading her until she had at last given him her
pledge. His gait was that of a man who knew the place well,
the swinging step of one used to walking easily in rough
places. His stick, too, was a rough ash, such as a town-bred
lover would never carry, while his voice had, I felt certain,
just a tinge of the Norfolk accent in it. That they should
meet at dead of night in that clandestine manner was surely
sufficiently suspicious, but those words I had overheard
sounded ever in my ears as I paced from end to end of that
old room with its sombre, almost funereal, hangings.

A great bitterness fell upon my heart. The woman whom I

really loved had played me false; and yet, when I reflected,
I could not help admitting that perhaps, after all, I deserved
this punishment. I had wavered from her and gone back to
my old love, it was true. But I loved Edith well and truly,
whatever might have been the fascination of the smart,
gesticulating, foreign beauty. She was mine in heart and
mine alone.

All my belief in woman’s affection or devotion had, in that

instant, been dispelled. The truth had fallen upon me as a
crushing blow, which staggered me, wrecking all my hopes
and plans for the future.

I tossed my things heedlessly into my bag, in readiness for

early departure in the morning. I had been a fool, I knew. I
was ever a fool where women were concerned. In the old
days in Brussels my affection for Yolande had been strong
and impetuous, burning with all the ardour of a first love;
yet the awakening had come, and I had tardily discovered
that she had played me false. And in Edith’s case, although
I entertained towards her such a real and deep affection as
a man only extends to a woman once in his lifetime,
unfaithfulness had once again been my reward.

I flung away my cigar. My agony of heart was too acute to

be accurately described in words. You, my reader, who may
have experienced the sudden breaking of your most
cherished idol, can only rightly understand the chagrin, the
intense bitterness, the spiritual desolation of that night
watch.

My candles were as nearly as possible burnt out. At length I

took my hat, and, creeping noiselessly downstairs, passed
through the dining-room, and let myself out by the window
which Edith had entered. The first grey of dawn was
spreading, and a sudden desire for fresh air had seized me.
I felt stifled in that old room with its gloomy furniture and
hangings. With the cool wind of early morning fanning my
heated temples, I struck straight across the lawn in the
direction taken by the mysterious lover. For some distance I
traversed the boundary of the grounds, until I discovered a
break in the oak fence, and, passing through it, found
myself out upon the broad, undulating meadows which
stretched away to the Beacon Hill and the tiny hamlet of
Toftrees, noted for its ancient hall and quaint church
steeple. Heedless of where my footsteps led me, I went
straight on, my mind full of the discovery I had made, my
heart overflowing. Away to my left, from behind the low
dark hills, the sky became flushed with the crimson light of
dawn; but all was still save the distant crowing of a cock
and the howl of a dog in the far distance. Behind me the
bell of Ryburgh church solemnly chimed the hour, followed
by other bells at greater distances. Then all was quiet again
save for the soft rustling of the trees. The morning air was
delicious, with a sweet fragrance everywhere.

Suddenly, leaping a fence, I found myself upon the old

coach-road that ran over the hills to Lynn, and continued
along it without thought of distance or destination. I passed
a carter with his team, and he wished me good-morning.
His words aroused me, and I saw that I was nearing an
unfamiliar village.

“What place is this?” I inquired.

“It’s Harpley, sir.”

I thanked him and went my way. I had never heard of the

place before; but as I entered it the first rays of sunlight
shot across the hills, and it certainly looked picturesque and
typically English in the light of the dawn. I must have
walked fully eight miles, and, being tired and thirsty, I
noticed at the entrance of the village a small inn, upon
which was the sign The Houghton Arms. The door stood
open, and a burly man, evidently the landlord, was busy
chopping wood in an outhouse at the side.

“Nice mornin’, sir,” he observed, looking up at me, probably

astonished to see anyone who was not a labourer astir at
such an early hour.

I returned his greeting, and inquired whether it was too

early for a cup of tea and a rest.

“Not at all, sir,” he answered, laying down his axe and

conducting me within.

The place, in common with all village hostelries, smelt

strongly of the combined fumes of shag and stale beer.
Village innkeepers have a habit of polishing their well-
seasoned furniture with sour beer; hence the odour, which,
to the patrons of such places, seems appetising. The
perfume is to them as the hors d’oeuvre.

The man, having shown me into a little parlour behind the

tap-room, called loudly to “Jenny,” who turned out to be his
wife. After this I had not long to wait before a pot of tea
and a couple of poached eggs were at my disposal.

They were a homely pair, these two, full of local chatter.

Harpley, the man informed me, was nine and a half miles
from Great Ryburgh, and I saw by his manner that he was
much exercised in his mind to know whence I had come and
the reason for my being about at such an hour. The rural
busybody was extremely inquisitive, but I did not permit his
bucolic diplomacy to triumph. While I drank the tea and ate
the eggs the landlord stood leaning against the door-lintel
with his arms folded, garrulously displaying his Norfolk
brogue. He evidently regarded me as one of those summer
visitors from London who stay at the farmhouses, where
hypocrisy terms them “paying guests,” and I allowed him to
adhere to his opinion. I learned from him that at six o’clock
there was a train from Massingham station, half a mile
away, which would convey me direct to Fakenham. This I
resolved to take, for I could then return to Miss Foskett’s by
a quarter to seven. A map of the county was hanging on the
wall, and I had risen to look at the spot to which the
landlord was pointing, when a footstep sounded in the
narrow passage, and, turning, I caught sight of the dark
figure of a man making his way out. The hat, the black
overcoat, the figure, all were familiar. His head was turned
away from me, so that I could not see his features, but in
an instant I recognised him.

He was Edith’s mysterious lover!

 Chapter Twenty.
From Downing Street to Paris.

I sprang quickly to the door, and looked down the passage

out into the village street; but he had already made his exit.
By the time I had reached the porch of the inn he was
already striding quickly along the dusty highway. He turned
to glance back, and I perceived that he was thin-faced, with
high cheekbones and a small black beard. He was carrying
his thick stick jauntily, and walking smartly, with an easy
gait which at that moment struck me as being distinctly
military.

“Who is that man?” I inquired eagerly of the landlord, who

stood beside me, evidently surprised at my sudden rush
towards the door.

“A stranger, sir. I don’t know who he is.”

“When did he arrive?”

“He came by the last train to Massingham last night, sir, and
had a bed here. My missis, however, didn’t like the looks of
’im.”

“Why?”

“Well, I don’t exactly know. There was something about him

a bit peculiar. Besides, he went out before one o’clock, and
didn’t return till an hour ago. Then he went up, washed, had
a cup o’ tea in his room, paid, and now he’s gone.”

“Rather peculiar behaviour, isn’t it?” I suggested, hoping to

find some clue to his identity from what this man might tell
me. “Did he have no luggage?”

“None. He seemed a bit down on his luck. His clothes were

very shabby, and he evidently hadn’t had a clean collar for a
week.”

Then the opinion I had formed of him—namely, that he was

shabby genteel—was correct.

“You’re certain you’ve never seen him before?”

“Quite certain,” he replied.

At that moment his wife entered, and, addressing her, he

said:

“We’re talking of that stranger who’s just gone, missus. His

movements were a bit suspicious, weren’t they?”

“Yes. Why he should want to go out half the night

wandering about the neighbourhood I can’t make out,
unless he were a burglar or something o’ that sort,” the
woman answered, adding: “I shouldn’t be at all surprised to
hear that one of the houses about here has been broken
into. Anyhow, we’d know him again among a thousand.”

“What kind of man was he?”

“Tall and dark, with a beard, and a pair of eyes that seemed
to look you through. He spoke all right, but I’ve my doubts
as to whether he wasn’t a foreigner.”

“A foreigner!” I echoed quickly, interested. “What made you

suspect that?”

“I really can’t tell. I had a suspicion of it the first moment I

saw him. He pronounced his ‘r’s’ rather curiously. His
clothes seemed to be of foreign cut, and his boots, although
worn out, were unusually long and narrow. I brushed ’em
this morning, and saw on the tabs a foreign name. I think it
was ‘Firenze,’ or something like that.”

I reflected for an instant. The word “Firenze” was Italian for

Florence, the town where the boots had evidently been
made. Therefore the mysterious stranger might be Italian.

“You didn’t actually detect anything foreign in his style of

speaking?”

“He didn’t speak much. He seemed very glum and

thoughtful. I sent him up some toast with his tea, but he
hasn’t touched it.”

“He didn’t say where he was going?”

“Not a word. When he arrived he only explained that he had

come by the last train from Lynn, and that he wanted a bed
—that’s all. I should think by the look of him that he’s gone
on tramp.”

My first impulse was to follow him; but on reflection I saw

that by doing so I should in all probability lose my train, and
to dog the fellow’s footsteps would, after all, be of no
benefit now that I knew the truth of Edith’s perfidy. So I
stood there chatting, discussing the stranger, and
wondering who he could be.

“He’s up to no good, that I feel certain,” declared the

landlord’s wife. “There’s something about him that aroused
my suspicion at once last night. I can’t, however, explain
what it was. But a man don’t prowl about all night to admire
the moon.”
And thus I waited until it was time to catch the train; then,
wishing the innkeeper and his wife good-morning, left them
and strolled in the morning sunlight to the station, arriving
at Fakenham shortly before seven. I took the short cut
through Starmoor Wood to Ryburgh, and, finding Miss
Foskett’s maid polishing the door-handle, entered and went
upstairs.

Upon the toilet-table was a telegram, which the maid said

had just arrived, and on opening it I found a message from
the Foreign Office, which had been forwarded from the
Club, asking me to call at the earliest possible moment, and
to be prepared to return to my post by the afternoon
service from Charing Cross. I knew what that implied. The
Marquess desired me to bear a secret despatch to my Chief.

I washed, tidied myself after my dusty walk, strapped my

bag, and with a feeling of regret that I was compelled to
meet my false love again face to face before departure, I
descended the stairs.

She was awaiting me, looking cool and fresh in her white
gown, with a bunch of fresh roses she had plucked from the
garden in her breast. She smiled gladly, and stretched forth
her hand as though I were all the world to her. What
admirable actresses some women are! Her affected
sweetness that yesterday had so charmed me now sickened
me. The scales had fallen from my eyes, and I was angry
with myself that I had ever allowed myself to lose control of
my feelings and love her. She was false—false! That one
thought alone ran in my mind as she laughed merrily.

“Why, Gerald, wherever have you been? A telegram came

for you by special messenger from Fakenham at half-past
six, and when Ann knocked at your door she found you
were out. And you went out by the dining-room window,
too.”

“Yes,” I said, not without a touch of sarcasm, “I felt that I

wanted fresh air, so I went for a stroll.”

“You are an early bird,” she answered. “Did you go far?”

“No, not very far. Only down the Lynn road a little way.”

“I always thought that you people in Paris never got up till

your déjeuner at eleven?”

“I’m an exception,” I said shortly. “I prefer the morning air

in the country to lying in bed.”

“And the telegram? Is it anything particular?”

“Yes,” I answered. “I must leave at once. I am summoned

to Downing Street, and must leave London this afternoon.”

“What! return to Paris at once?”

“Yes,” I replied. “It is an order from the Chief. There’s a

train to London at 9:50, I think. I must not fail to catch
that.”

I had not kissed her, and I saw that she was somewhat
puzzled by my coolness. Did the fact that I had let myself
out by the dining-room window give her any clue to the
reason why I had chosen that mode of egress?

“I thought you would remain here with us at least to-day,”

she pouted. “That’s the worst of diplomacy. You never seem
to know what you may do next.”

We were standing alone in the dining-room, where

breakfast was already laid and the copper kettle was hissing
above the spirit-lamp. As Aunt Hetty had not entered, it was
upon the tip of my tongue to charge Edith with that
clandestine meeting; yet if I did so, I reflected, a scene
would certainly be created. Aunt Hetty would first be
scandalised and afterwards wax indignant, while my
departure would doubtless be fraught with considerable
unpleasantness. Therefore I resolved to keep my anger
within my heart, and on my return to Paris to write a letter
of explanation to this smiling, bright-faced woman who had
thus played me false.

“You cannot tell how wretched I am at the thought of your

departure, Gerald,” she said, her dark eyes suddenly grave
and serious. “Each time we part I always fear that we shall
not meet again.”

I smiled, rather bitterly, I think, and uttered some weak

platitude without appearing to be much interested. Then
with a quick movement she took my hand, but next instant
was compelled to drop it, for Miss Foskett entered suddenly,
and, after an explanation of my unexpected call by
telegram, we seated ourselves and breakfasted.

As the woman I had so dearly loved sat opposite me I saw

that she was strangely nervous and agitated, and that she
was eager to question me; but with feigned indifference I
chatted and laughed with the punctilious old spinster until
the boy brought round the pony-trap and it was time for me
to depart for Fakenham, where I could join the express for
London.

Edith drove me to the station, but, the boy being with us,
she could say nothing confidential until we were walking
together upon the platform. Then, looking at me in strange
eagerness, she suddenly asked:
Welcome to our website – the perfect destination for book lovers and
knowledge seekers. We believe that every book holds a new world,
offering opportunities for learning, discovery, and personal growth.
That’s why we are dedicated to bringing you a diverse collection of
books, ranging from classic literature and specialized publications to
self-development guides and children's books.

More than just a book-buying platform, we strive to be a bridge

connecting you with timeless cultural and intellectual values. With an
elegant, user-friendly interface and a smart search system, you can
quickly find the books that best suit your interests. Additionally,
our special promotions and home delivery services help you save time
and fully enjoy the joy of reading.

Join us on a journey of knowledge exploration, passion nurturing, and

personal growth every day!

ebookbell.com