The history of final vowels in english

The History of Final Vowels in English


Topics in English Linguistics
4
Editors

Jan Svartvik
Herman Wekker

Mouton de Gruyter
Berlin
· New York


The History of Final Vowels
in English
The Sound of Muting
Donka Minkova

Mouton de Gruyter
Berlin
· New York

1991


M o u t o n de Gruyter (formerly M o u t o n , The Hague)
is a Division of Walter de Gruyter & Co., Berlin.

© Printed on acid-free paper which falls within the guidelines of the
A N S I to ensure permanence and durability.

Library of Congress Cataloging in Publication

Data

Minkova, D o n k a , 1944 —
The history of final vowels in English : the sound of
muting / D o n k a Minkova.
p.
cm. — (Topics in English linguistics ; 4)
Includes bibliographical references and index.
ISBN 0-89925-784-4 (alk. paper)
1. English language —Vowels. I. Title. II. Series.
PE1157.M5 1991
42Γ.5 —dc20 ·
91-28082
CIP

Die Deutsche Bibliothek

— Cataloging in Publication

Data

Minkova, Donka:
The history of final vowels in English : the sound of muting /
D o n k a Minkova. — Berlin ; New York : M o u t o n de Gruyter,
1991

(Topics in English linguistics ; 4)
ISBN 3-11-012763-6
NE: G T

© Copyright 1991 by Walter de Gruyter & Co., D-1000 Berlin 30
All rights reserved, including those of translation into foreign languages. N o part of this
book may be reproduced or transmitted in any form or by any means, electronic or
mechanical, including photocopy, recording, or any information storage and retrieval
system, without permission in writing from the publisher.
Printing: Ratzlow-Druck, Berlin.
Binding: Lüderitz & Bauer, Berlin.
Printed in Germany.


Contents
Preface

ix

Chapter 1
Schwa in the history of English

1

1.1.
1.2.
1.3.
1.4.

Problems involved in the present study

Forces at play in schwa loss
Remarks on method
Analytical procedures

Chapter 2
More background: Accounts of schwa loss
2.1. Choice of sources
2.2. Lorenz Morsbach (1896): Mittelenglische Grammatik
2.3. Karl Luick (1921-1940): Historische Grammatik der englischen Sprache
2.4. Richard Jordan: (1934/1968) Handbuch der mittelenglischen
Grammatik
2.5. Wilhelm Horn/Martin Lehnert (1954): Laut und Leben. Englische Lautgeschichte der neueren Zeit (1400 — 1950)
2.6. Karl Brunner (1962): Die englische Sprache, ihre geschichtliche Entwicklung. Vols. I, II
2.7. Joseph Wright and E. M. Wright (1923): An Elementary
Middle English Grammar
2.8. Fernand Mossé (1949): Manuel de l'anglais du moyen âge
2.9. Jacek Fisiak (1970): A Short Grammar of Middle English
2.10. Peter Erdmann (1972): Tiefenphonologische
Lautgeschichte
der englischen Vokale
2.11. Summary
Chapter 3
Textual evidence
3.1. Principles and methods specific to the study of schwa loss
in Middle English
3.2. Types of evidence
3.3. Graphically marked morpho-syntactic loss of -e

1
2

3
10

15
15
16
20
23
24
24
25
26
27
27
28

35
35
37
45


vi

Contents

3.4.
3.5.
3.6.
3.7.

3.8.

Unetymological final -e
Elision in hiatus
Rhyme evidence
Scribal errors, editorial corrections
Graphically unmarked evidence

Chapter 4
Phonological aspects of schwa loss
4.1.
4.2.
4.3.
4.4.
4.5.
4.6.

The phonemic status of schwa in Middle English
The Middle phonemic inventory
Distribution
Phonetic parameters. Distinctive features
Phonological correlates of schwa loss
Some consequences

Chapter 5
Morphological aspects of schwa loss
5.1.
5.2.
5.3.
5.4.

5.5.
5.6.

The morpho-syntactic functions of schwa
The morphologization of schwa loss
Classification and ranking of factors for schwa loss
Syntactic correlates
Extralinguistic factors
Hierarchy of factors within individual word classes

Chapter 6
Early schwa deletion as a prosodie phenomenon
6.1.
6.2.
6.3.
6.4.

Early schwa loss data
The shared properties of early schwa loss
The rhythmic phrasing of early schwa deletion
The Early Schwa Deletion Rule

55
62
69
71
75

87
87

91
104
106
109
114

125
125
133
134
140
146
149

155
155
158
159
162

Chapter 7
Schwa preservation in Late Middle English as a prosodie phenomenon

171

7.1. Final -e in weak adjectival inflexions
7.2. The weak adjectives in Chaucer
7.3. The weak adjectives in non-Chaucerian Late Middle English

171

172
173


Contents

7.4.
7.5.
7.6.
7.7.
7.8.

Accounts of adjectival schwa preservation
The rhythmic nature of final -e
The metrical structure of adjectival phrases
The proposal reviewed: objections
Conclusion

vii

175
177
178
186
187

Bibliography

193

Index

217



Preface
Schwa has always been the poor relation of the stressed vowels in English
historical phonology. The academic searchlights have been directed on
the developments of the vowels in stressed position. Most studies in the
field do make some reference to the changes of unstressed vowels, yet the
supposedly predictable genesis and unidirectional fate of the ubiquitous
schwa-type vowels have attracted very little attention. Seen in a wider
context, however, the reduction of unstressed vowels to schwa and its
subsequent loss in final position has had a more p r o f o u n d effect on the
phonology, prosody, and grammar of the language than any individual
qualitative or quantitative change of the stressed vowels. This book is an
attempt to bring together traditional philological records and knowledge
of the ways and means of schwa loss and reexamine them in relation to
the entire language system.
The first three chapters of the book deal with the data base: how we
know that unstressed vowels in final position were reduced and lost, what
methods and resources can be used in this context, how we classify and
interpret the various types of textual evidence. Separating out poetic from
non-poetic schwa loss, graphic f r o m inferred evidence, loss of nonmorphemic, root-final schwa f r o m loss of inflexional schwa, proves useful
in establishing causes and consequences of the process on all levels.
Chapter 4 addresses rarely asked questions a b o u t the synchronic status
of schwa in Middle English. An examination of the evidence for phonemic
identification of the unstressed vowel in final position suggests that in
terms of traditional phonemic analysis schwa has always been part of the

phonemic inventory of English. In Early Middle English there was a
match between the underlying, or lexical, representation of schwa, and
its surface realization. Instances of early deletion are analysed as postlexical rules operating first within a narrow set of environments, and later
across the board in word final position. By the end of the Middle English
period, after c. 1400, schwa is disallowed as an underlying segment in
absolutely final position except in a peripheral subset of markedly foreign
borrowings. Optional insertion rules operate in very specific lexical items
and in some clearly defined prosodie contexts.
Chapter 5 covers the morphological correlates of the change. In Middle
English as many as eighteen different grammatical functions can be
associated with the morpheme represented by <-e>
in writing. This
suggests that lack of morphological distinctiveness is a m a j o r factor in


χ

Preface

schwa loss in those word classes, such as the nouns and the verbs, in
which schwa figures prominently in the inflexional paradigm. It is argued
that the familiar generalized statements concerning the morphological
attrition of schwa should be refined substantially by referring not to the
lexicon as a whole, but to separate word classes. The causes and rate of
loss in each class turn out to be different. The chapter offers a detailed
hierarchization of individual factors and their potential importance in
schwa loss for each word class.
Chapters 6 and 7 introduce arguments from metrical phonology into
the account of schwa loss. Schwa loss affected the language continuously
for at least three centuries; it is impossible to get a clear and systematic

notion of what items were subject to deletion at the peak of the change,
around 1250 — 1350. However, if we look at the very early instances of
deletion, as well as at the instances of late preservation of schwa, a clearer
pattern emerges. Early schwa deletion can be defined in terms of a specific
metrical foot structure — a definition which extends to later cases of
deletion and eventually embraces all occurrences of word final loss. At
the other end of the chronological span, Late Middle English, we find
schwa surviving consistently only in one isolated syntactic context, the
weak declension of etymologically monosyllabic adjectives. The motivation for the preservation is neither functional nor segmental; it can be
read off the metrical structure of the constructions involved and can be
seen as an instance of the principle of eurhythmy. Both chapters address
schwa loss from a new perspective — the perspective of prosodie organization of language. Both chapters demonstrate that the operations of schwa
loss should be constrained within two domains: the domain of the foot
and the domain of the clitic group.
The customary task of acknowledgements and thanks is not easy in
the case of a study which draws on the achievements of traditional
philological scholarship and on modern linguistic theory. Indebtedness is
not quantifiable, and one paragraph will inevitably fall short of expressing
my gratitude to all who have helped bring the project to this stage. Work
on this book was done in three universities: Edinburgh, Sofia, and UCLA.
The mixture of background can be my only excuse for the blend of
philology and at least two kinds of linguistics I have used in my analysis.
I acknowledge gratefully the generous assistance of the U C L A Senate
Research Committee since 1985 for the final research and writing. Two
great philologists and teachers, Professors Marco Minkoff and Maria
Rancova, encouraged and supported me in my early steps in the profession. To their memory this book is gratefully dedicated. Among the many


Preface

xi

colleagues to whom I feel indebted I want to mention specially Roger Lass,
Heinz Giegerich, John Anderson, Edmund Gussmann, Bruce Hayes; they
have shown me how exciting and complicated historical phonology can
be, though no study could do justice to their collective wisdom, nor
should any of them be implicated in my blunders. My closest professional
model and associate Robert Stockwell has tried in vain to teach me how
to keep effusiveness and verbosity in check; to his salutary influence the
readers owe the elimination of one third of the original bulk of the book.
My sons Dimo and Lubo, the culprits solely and jointly responsible for
all infelicities in the book, have shared me with schwa far too long —
their only reward must be that no other teenager in the world would
know what this weird word means. Finally, for years I have described
my task to friends and colleagues as "everything you always wanted to
know about schwa but were afraid to ask". What started out as a blind
date with the topic developed into a relationship which has enriched me;
I hope I have enriched it. More questions remain than have been asked
or answered; this is only a beginning.



Chapter 1
Schwa in the history of English

1.1. Problems involved in the present study
The relative significance of various factors, phonological, morphological, syntactic, sociolinguistic, and stylistic in relation to the causes and consequences of the
loss of final schwa in English has remained an unresolved and controversial issue
in the philological literature of the last 150 years. This study is an attempt to
clarify some recalcitrant aspects of the change and offer a hypothesis defining

schwa loss in terms of the more general linguistic properties of English, diachronically and synchronically.
Provisionally, before going into the phonetic and terminological details in
Chapter Four, I will use the term schwa to mean, vaguely, the unstressed, reduced,
mid, central, murmured, indeterminate, colorless vowel, the usual sound of the
first and last vowel of America, for which I use the IPA mid central vowel symbol [s]. Briefly, word-final schwa in English is generally assumed to have originated from neutralization of the oppositions between a number of distinctive and
stable sounds occurring prior to and during the time span between the ninth and
the eleventh centuries. Its subsequent history, i.e. between the eleventh and the
fourteenth centuries, is characterized by non-distinctiveness, instability, and progressive disappearance; by the second half of the fifteenth century the letter <-e>
was no longer pronounced, except in a handful of recognizably foreign words.
The spelling evidence for Old English (OE) suggests that the distance between
the vowels /e/, /a/, /u/, and /o/ in unstressed syllables was fairly well preserved
throughout the Classical period (excluding the /a/ and /o/ neutralization, and possibly other short vowel neutralizations before nasals). 1 After 950, in Late Old
English (LOE) and in Early Middle English (EME) this difference was no longer
maintained; the grapheme <-e> began to be used in all post-tonic syllables, no
matter whether they belonged to basically monomorphem«: words, e.g. OE guma
'man', ME gume; OE wundor 'wonder', ME wunder, OE lufu 'love', ME luve; OE
stede 'place', ME stede; or whether they were syllables functioning as inflexional morphemes, e.g. mac-ode 'made', ME makede, OE waer-on 'were', ME
weren; OE hwal-as 'whales', ME whales; OE dag-um 'to the days', ME dagen;
aris-αδ 'they rise', ME ariseth, OE scip-a 'of ships' ME shipe etc.^ The high vowels in some derivational affixes, especially the /i/, appear to have resisted the
weakening; thus /i/ shows no evidence of reduction and typically survives in


2 Schwa in the history of English
suffixes such as -ίξ, -ic, -isc, -ing; while -ung alternates with -ong. Morphemic
identification is probably responsible for the consistent preservation of the
unstressed vowel in words ending in -o(c)k: parrok, hassok, buttok, bullok. ^
There is no written documentation of reduction in a limited set of lexical items,
which behave indiosyncratically most likely for reasons of association with a
specific register (ME bishop, abbot) - no morphological justification can be
found for the apparent persistence of the original vowel in their final syllables.

1.2. Forces at play in schwa loss
Standard textbook descriptions of final vowel reduction in Late Old English and
Early Middle English attribute the process, manifested by the adoption of <-e>
spelling in post-tonic syllables to the following factors:
(1) Phonetic changes in progress: phonetic reduction and its corollary - the
neutralization of phonemic distinctions - resulting from the
concentration of stress on the root syllable, leading to a series of
important grammatical changes, and above all to a radical decrease in the
number and loss of inflexions.
(2) Morphological changes in progress; neutralization of the original differences
between the inflexional morphemes of nouns, verbs, etc., renders these
morphemes functionally deficient, and promotes the appearance of new
means of expressing grammatical relations. This leads to redundancy and
eventual loss of the final unstressed vowel in major class words.
(3) Phonetic and morphological changes in progress simultaneously; in this view
phonology and morphology are in a counterfeeding relationship: the
absence of stress prominence renders inflectional morphemes nondistinctive phonologically, and therefore redundant functionally; loss
due to the absence of function in some instances extends to cases where
schwa is not a grammatical unit, contributing to the phonological spread
of the change. Some variant of this compromise is the most frequently
encountered description of the loss. I believe this to be the only correct
account, yet it needs additional defense, expansion, and refinement.
(4) In addition to the system-internal factors in (1) - (3), the developments of the
post-tonic syllables in Middle English have been causally linked to the


Forces at play in schwa loss 3
extralinguistic factor of the existence of a large bilingual community in
areas of the North and North East Midlands after the tenth century.

This study surveys the evidence for the changes in post-tonic syllables in Middle
English. In the light of this evidence I examine the plausibility of the views
summarized in (1) - (4). There is little discussion in the philological literature of
the phonological properties and behavior of /a/ in Middle English, its phonemic
status, or its place and function within the prosodie system of the language. The
last two chapters will address the issue of schwa loss in terms of the prosodie
patterns and preferences of Early and Late Middle English.

1.3. Remarks on method
Part of my task is purely descriptive: I will review and gather in one place textual
information scattered or unavailable in the literature. This prompts the necessity
of constructing a data-specific taxonomy. Once this has been accomplished, and an
acceptable classification has emerged, I will reconsider the results in the light of
independently formulated traditional and more recent theories and assumptions.
Combining the philological with the linguistic approach holds the promise of
new insights into the history of this important and underresearched issue.
Providing explanations of language change has been the concern of most major
schools, branches, currents in both philology and linguistics. I state briefly my
indebtedness to the various hypotheses which may have direct bearing on the
history of schwa. The background assumptions and the descriptive and analytical
techniques I have used are self-selected and derived from the following
contributions to the historical study of language:

1.3.1.

Neogrammarian

The notions of sound law and analogical change are crucial for the Neogrammarian
model. In that framework sound change is characterized by a regularity which
elevates it to the status of a "law". Phonological change is also understood to

proceed mechanically, affecting all items exhibiting a particular phonetic
configuration; it is presumed to be gradual and impervious to direct observation.
The famous Neogrammarian identification of sound laws with universal
natural laws has been questioned and rejected on the basis of the now obvious fact
that the regularity of phonological change is never absolute. Post Neogrammarian


4 Schwa in the history of English
studies have revealed intricate ties between sound change and morphology, syntax,
and word formation - this now axiomatic contingency compromised the notion of
"mechanical" sound change. The gradualness hypothesis, although it overlaps
with our ideas of change through lexical diffusion, does not coincide with it, and
leads to unacceptable notions of pervasiveness, and has been challenged on this and
other scores. The same is true of the idea of "unobservability" of sound change,
which was rendered obsolete by the excellent sociolinguistic studies of the last
couple of decades.
Schwa loss is clearly a phonetic process. From that point of view the principle
of regularity supports the assumptions concerning the predictable relationship
between lack of stress as the conditioning factor, and weakening and deletion of
the final vowel. Under a magnifying glass, however, the detail of textual evidence
reveals an uneven rate of disappearance of -e in Middle English, which suggests a
morphological interference with the process of loss. Even this initial basic
knowledge would preclude any attempts to formulate the change in terms of a
Neogrammarian sound law.
Analogy is another tenet of the Neogrammarian school. It is a principle which
applies to the patterning of higher-than-the-phoneme units, i.e. it is most
frequently appealed to in morphology, syntax, word formation. Analogy has the
effect of counteracting phonological change. The extent to which this principle
can be called upon to explain the causes and mechanisms of language change
presents a complex and controversial problem and there is a vast body of

literature on analogy. This model of accounting for language change has been used
by scholars since the sixteenth century. The treatment of analogy in the
Neogrammarian school allows it to be a strategy which "tends to become a
terminological receptacle, devoid of explanatory power, a catchall for
irregularities in the operations of 'regular' sound laws" (King 1969: 127), and
therefore in more recent diachronic studies it has been overshadowed by newer
proposals and only mentioned in conjunction with such proposals as a peripheral
factor. Many instances of paradigmatic leveling unrelated to analogy have been
quoted in support of the idea that this explanatory model should either be
abandoned or resorted to only in very special cases (King 1969: 129-30). The
related idea of proportionality, however, has had many proponents. Studies by
Kuryfowicz (1947), Manczak (1980), Anttila (1977), etc. have attempted to
formulate and explicate the "laws" and tendencies of analogy and have succeeded
in eliminating some of the insidious vagueness discrediting this model.
Analogy enters into the discussion of schwa deletion for two reasons. First,
schwa loss is a thoroughgoing morphological change, and morphology is one of
the language domains where analogy is agreed to be most powerful and most


Remarks on method 5
effective. Second, traditionally, sound change and analogy are seen as standing in a
restorative relationship, i.e. "grammatical structures rendered opaque, or under
the threat of being rendered opaque, by sound change are likely to be 'repaired' by
analogical change" (Bynon 1977: 44). Yet, analogy is not necessarily only an
antidote used to reintroduce order into a paradigm into which sound change has
inducted lack of uniformity. At least partially, after the levelling of inflexions,
the Middle English picture is one of sound change and analogy both working in
the same direction - a situation unaccounted for by the standard Neogrammarian
views of the interplay between sound change and analogy. ^

1.3.2. Functionalism
Even if one accepts analogy as a valid and interesting model of language change, its
explanatory power is confined to the spread of already existing forms; it
obviously feeds on already existing forms and cannot create new forms of its own
accord. Functionalist accounts, on the other hand, have been concerned with the
actual origin of innovative linguistic forms. Scholars in this tradition, adopting
the idea of language as a more or less rigid hierarchy of mutually dependent
constituents, have contributed to the elucidation and understanding of the
relations obtaining among the various linguistic units within the language
system. The limitations of this approach are by now widely recognized (cf. Lass
1980a), yet it has been fruitful in stimulating further research and in prompting
alternative hypotheses. In the history of schwa in English a certain degree of
"functionality" has to be incorporated into the account of its phonemic
development - useful and insightful notions such as that of phonemic contrasts
and neutralization are functionalist, and they are central to the explication of the
synchronic relations of -e in Middle English.
Diachronic functionalism, the recourse to explanations of language change by
invoking the metaphor of survival of the fittest, of linguistic units as healthy, or
unhealthy, as superior or inferior in performing their duty, is more suspect.
Language change does not proceed irrespective of, or in spite of, the needs of
communication. The familiar adage "communication alone shapes language"
(Martinet 1960: 191) is an alternative way of saying that language is first and
foremost a social phenomenon. Considered in the narrow context of schwa loss,
though, extreme functionalism would imply that if at any point in its history -e
had been perceived as absolutely necessary, if its omission would have impaired
the communicative value of the utterance, its ultimate loss without prophylaxis
or further therapy would have been inconceivable. ^


6 Schwa in the history of English

Another ramification of the functionalist model concerns the reliance on a
statistically computed functional load of a unit, which is then assumed to be in
some way explanatory in terms of the unit's proneness to various diachronic
processes. In spite of the initial attractiveness of the idea of an empirically based
explanatory strategy, its application encounters too many serious problems for it
to be adopted without further comments or reservations.
Functional load, ultimately a teleological concept, is measured by Prague
School phonologists by the number and frequency of the minimal oppositions into
which a given phoneme can participate, a procedure supposed to predict the
diachronic stability or instability of a phoneme. This type of diagnostic runs into
difficulties both with reference to the actual mechanics of computing functional
load, and with reference to the interpretation of the results within a more general
picture of the history of a language. How does one select the correct amount of
linguistic material providing the necessary and sufficient context for the
evaluation of the functional load of one phoneme? Does the idea apply to morphology and syntax as well as to phonology? Is it phonemes, allophones, or
features that one measures the functional load of? These and other questions have
prompted extensive criticism of the whole notion, cf. King (1967), Bynon (1977:
87-89), Lass (1980a: Chapter Three). I limit my "functionalist" discussion of -e
to an evaluation of its phonemic and morphological status within the entire
system of the language, and I agree from the start with the cautionary note that
"it may be altogether inappropriate to look for a motivation of phonological
change which is based on the structural status of phonological units" (Bynon
1977: 88).
Another ramification of the structuralist model is the view that "the
phonological evolution of a language may be seen as a continuous effort to
maintain a state of balance between inertia on the one hand, and communicative
needs on the other" (Bynon 1977: 89). Economy is a general linguistic
"collective-psychological" factor which has been described as "basic in the
analysis of the mechanism of language change" (Polivanov 1931: 54). It is
common sense to assume that, especially in more relaxed and colloquial registers,

speakers can take shortcuts, and their language would be subject to variability and
change attributable to "inertia" or "human laziness". These notions can all too
easily be confused with the now totally rejected idea of diachronic developments
as instances of "deterioration". In any case, evoking "economy", in the sense of
"shortcuts", and not as "deterioration", as anything more than a marginal
phenomenon in historical linguistics is an extreme position which may be
dangerously loaded in the context of a change which amounts to deletion, as is the


Remarks on method 7
case with schwa loss. The pitfalls of oversimplification are all too obvious; I
will try and direct my analysis away from this type of "economy".

1.3.3. Rule-governed change
Within the model of generative phonology linguistic change affects the set of
rules according to which the grammar of a language is organized; therefore change
is described in terms of loss or addition of rules, rule reordering or rule
simplification, cf. King (1969), Kiparsky (1968, 1970). Originally in this
framework the traditional processes of sound change and analogy were stated as
rules affecting the speaker's competence.^ A huge body of studies in synchronic
and historical linguistics have shown that both with reference to phonology and
with reference to syntax surface, or performance phenomena must be given
equally serious consideration. The development of the theory along these lines has
resulted in what has been described as a rapprochement between the generativist
and the traditional philological views, cf. the discussion in Bynon (1977: 145).
Accumulation of diachronic data in phonology has confirmed the assumption
that "the motivation of phonological changes depends rather on performancerelated targets ...than on purely formal considerations" (Bynon 1977: 138). The
psychological reality and the analytical usefulness of the phoneme, initially
regarded as uninteresting by the early generative phonologists, have regained
legitimacy as minimal contrastive underlying units, albeit in the guise of

"underlying segments". The surface attestations of underlying forms can be
derived with the help of rules which operate on those forms. Stipulation of the
conditions for the operation of the rules translates any difference between a
postulated underlying form and its realization into a predictable, rule-governed
change. Stability and pervasiveness of a particular type of innovative surface
realization may suggest that there has been a reanalysis of the underlying
linguistic "system". Little has been done in studies of schwa loss from this point
of view; as a very preliminary outline of the problem one can say that during
Middle English schwa loss was initially a phonologically and morphologically
motivated rule. The spread of the rule caused reanalysis of the underlying
structure of the language. By the end of the period a schwa whose sounding is
guaranteed by e.g. the metrical pattern of a poetic piece, must be accounted for in
terms of insertion by rule, rather than as a deletion.


8 Schwa in the history of English
1.3.4. Socially motivated change
Language is a social phenomenon; it cannot be understood or described adequately,
let alone "explained", without reference to the social context in which it
functions. Sociolinguistic studies have contributed substantial insights into our
comprehension of the mechanisms of language change, and of sound change in
particular. Martinet's extreme position (1960: 188-189) that "only internal
causality concerns the linguist" has long been shown to be too limited; the
complete autonomy of the so called "internal" and "external" factors has also
been rejected, see Vachek (1962), Avrorin (1975: 25-28),** Berezin and Golovin
(1979: 189), Bynon (1977: Chapter Five).
The extent to which the sociolinguistic parameters of a dead language can be
established depends on the availability of texts representing various stylistic
registers. Some marginal information can be gleaned from knowledge of the
regional origin of a particular text, as well as from familiarity with the personal

history of the author. Such information is not readily accessible for Middle
English; in the absence of first hand comments and descriptions of change in
progress by contemporaries, we have to be satisfied with very broad and possibly
unilluminating observations on the "social" character of schwa loss. One would
assume, for instance, that in its incipient stage schwa loss could have a
phenomenon correlating with speech unaffected by familiarity with the spelling
system, "uncontaminated" by education. Conversely, since the written form of
the language never abandoned the symbol for schwa, one might speculate about the
possibility of a period of time when pronouncing final schwas could have been a
mark of a literate, highly conservative variety of English. One line of inquiry
which can prompt interesting observations about the rate of the change in relation
to the sociolinguistic background would be taking into account the register of the
textual source, i.e. one would expect alliterative poetry to be more conservative
than accentual-syllabic poetry, narrative type texts to be more advanced than nonnarrative texts. However, the balance and interplay between "internal" and
"external" forces in language change in the case of schwa loss can be sought more
profitably in the introduction of another parameter - outside linguistic influence.

1.3.5. Contact-induced, change
Contact between languages, in the broadest sense of the word contact, is a noncontroversial source of language change. In a strict dichotomy of "external" vs.
"internal" factors for language change language contact would fall under the


Remarks on method 9
former group.9 Evidence that linguistic developments can be generated by the
contact between languages is most readily available in the sphere of vocabulary,
but there is also abundant literature on the manner in which continuous exposure
to and simultaneous use of two languages can influence a speaker's phonology,
morphology, syntax. In the conditions of prolonged bilingualism on the British
Isles, both Anglo-Scandinavian, and Anglo-Norman, one can expect some evidence
of the varying rates of schwa loss depending on the linguistic situation of the area

from which a given text originates. A direct influence from either Scandinavian or
Norman French on the loss of -e in phonological terms cannot be postulated, since
there is no parallel development in Scandinavian, and as for the French -e muet, it
began to disappear from the end of words under certain conditions much later than
it did in English. The morphological history of {-e} is universally assumed to have
been influenced, at least to some extent, by the contact between the different
languages spoken on the British Isles. More specifically, there is a long tradition
of associating the disintegration of the inflexional system of English has to
Anglo-Scandinavian bilingualism (Jespersen 1938 [1962]: §79; Ilyish 1973: §275
etc.). According to this widely accepted view absolute accuracy of the
grammatical form is not essential in the process of communication because of the
genetic relationship between the language of the settlers and the language used by
the native population. The phonetic identity or similarity of the roots of words is
taken as a catalyst, and in some studies, as a sufficient condition for basic
understanding to take place between the speakers of English and the Scandinavian
settlers. More recently, detailed research of creolization processes has provided a
new framework within which the contact between Scandinavian and English can
be described. Middle English has been analyzed in terms of a system which cannot
be derived directly from its Anglo-Saxon predecessor (Fisiak 1977: 249). This
new system, the Middle English interlanguage or creole (the two terms are used
synonymously here), emerged under the decisive influence of Scandinavian. Poussa
(1982: 83) summarizes the first stages in the development of (Chancery) English
as follows:
866
918
1016
1016
1066
1066
14th c.

Old Mercian. Old Danish
East Midland Creole
Hybridization with Late West Saxon and Mercian
Stabilization as spoken koiné in London and the Midlands
Further Danish influence, hybridization with London dialect(s)
French influence (mainly written, apart from lexis)
English patriotic revival. Immigration to London from the
Midlands.


10 Schwa in the history of English
The earlier and familiar Jespersenian view, and the more recent, creolization
account of the way in which contact with other languages affected the history of
English are fully compatible. They both cover the attested dialectal distribution
of morphological changes in Middle English; the creolization hypothesis,
however, emphasizes processes which take place in the spoken form (Poussa 1982:
70), undoubtedly an important and neglected area of historical analysis. Stylistic
considerations are also taken into account in the study of creóles. Creoles are
generally characterized by loss of grammatical gender and simplification and loss
of the grammatical inflexions; assuming the existence of a Middle English creole
can help us sharpen our understanding of the very different behavior of -e in
English and in the other Germanic languages given identical phonological
conditions of reduced stress.

1.4. Analytical procedures
The brief introductory notes in 2.1 on the major ways in which the origins and
nature of language change have been approached in the literature, as well as their
applicability to the Middle English material must be supplemented by an outline
of concrete procedures of analysis. Once again my point of departure is fairly

familiar. Important methodological principles for the study of language change
were laid down in the classical study by Weinreich, Labov and Herzog (1968: 183188). I reproduce their Empirical Principles for the Theory of Language Change
below:
(1) Determine the set of possible changes and possible conditions for change (The
Constraints Problem ).
(2) Trace the ways in which the changing feature is embedded in the linguistics
structure (The Embedding Problem [a]).^
(3) Consider all factors, "stimuli and constraints both from society and from the
structure of language" which influence change in relation to one another
(The Actuation Problem). Heterogeneity underlies change. An overall
view of the process of change is best achieved through a multiplicity of
approaches. (186).^
This cursory survey of some basic background assumptions suggests that one
cannot expect a simple and unequivocal answer to the question of why a certain
change has occurred in language. The only purely internal factor about whith we


Analytical procedures 11
can be positive with respect to schwa is analogy; analogy, as noted, however,
provides a satisfactory explanation only of the spread of already existing new
forms, i.e. the source of the proportional model for the analogy should be
available before it can start affecting more and more forms in the l a n g u a g e .
External factors, sociolinguistic, or contact-induced processes, must be brought
into the picture; they can provide significant additional information in the account
of schwa loss, especially in reference to the surface morphological composition of
previously inflected grammatical classes: nouns, adjectives, and verbs.
The discussion of theories of the origin and mechanisms of language change
presented here has not been concerned with the "actuation" of change, in terms of
transmission patterns as described by scholars studying language acquisition.
Change can be attributed to the way children and adults acquire and modify

language. This particular aspect of the "transition" of change is unrecoverable in
the written records; in the absence of any relevant historical information
attempts at making reasonable use of the notions of "actuation" and
"transmission" in an acquisitionist sense is futile.
Emphasis in this study will be laid on "change", which brings to the fore the
dynamic aspect of language. "Change" is a concept which comprises two separate
and discrete stages in the development of any diachronic process: the stage of
innovations, and the stage of actual change, the latter involving "the
generalization, or codification of innovations" (Andersen 1974: 22).
At the same time an analysis which aims to uncover the network of functional
relations within the limits of one historical period must necessarily assume the
existence of an idealized, relatively homogeneous, synchronic bulk of language.
Each Middle English text examined will be taken as representing one such
synchronic state. Synchronic states can then be compared to each other to reveal
the dynamism, instability, variability, potential or actualized "change" in a
diachronic perspective. This theoretical position, ideally achieving a perfect
compromise between the synchronic and diachronic approaches, has gained
universal acceptance and requires no special justification.

Notes
Chapter 1. Schwa in the history of English

1. Angled brackets < > are used throughout the text for graphemes; squares [ ]
enclose allophones; slashes / / enclose phonemes. Italicized -e is used as a


12 Schwa in the history of English
cover notation for any final -e, prior to defining it as a grapheme, a phoneme,
an allomorph, or a morpheme. Curly brackets { } are either morphemic or
enclose members of a set.

2. The length of the stressed vowel in the forms quoted in the text has been left
unmarked if it is irrelevant to the discussion.
3. See Jordan (1974: § 133-135), also Luick (1921-1940: § 440).
4. This type of mutual enhancement of sound change and analogical change is
discussed in Zirmunski (1954: 38 ff.): "It is not always [the case] that
phonetic changes disorganize the grammatical system...." Here Zirmunski
refers specifically to the regular phonetic reduction of final unstressed
syllables as a very significant factor in the development of the inflection in
the Germanic languages. In this case the phonetic process eliminates the
discrepancies within the old grammatical system and thus interacts with
morphological analogy to create a new, more unified, and consistent system.
Zirmunski believes also that analogical tendencies ... "contribute to the
improvement and perfection of the grammatical rules of a given language in
accordance with special internal laws of its development" (1954: 28). I take
"improvement" and "perfection" to imply that a morphologically uniform
paradigm is somehow superior to a formally heterogeneous one, an
interpretation which renders this statement unacceptable. Of some interest is
another point which Zirmunski makes in this connection,however, namely
that internal, i.e. inflectional analogy restores the unity within the paradigm
of a word, separating more rigorously the root morpheme as the bearer of the
objective (material) meaning of the word (1954: 31). This is an important
typological observation about word structure in the Germanic languages and
in English in particular; morphological analogy related to the weakening to
and loss of {-e} results in realignment and concentration of the semantic and
grammatical information on the root.
5. The controversy over the legitimacy of functionalist arguments and
explanations continues (see Lass 1987a, 1987b; Samuels 1987).
6. The type of "economy" discussed here should be differentiated from the
unfortunately synonymous, but highly desirable economy of linguistic
description within the generative framework. The reference here is to

functionalist economy, an idea which most linguists have dismissed as
untenable. Yet as recently as 1981 a whole book (Krustev 1981) was devoted
to the absurd task of "explaining" a series of changes, phonological,
morphological, syntactic, derivational and even graphic, from the principle of
economy. This extreme and oversimplified view cannot be found in the works