Instant Ebook Access, One Click Away – Begin at ebookgate.

com

The Sound Shape of Language 3rd ed., with a new

preface Edition Linda R. Waugh

https://ebookgate.com/product/the-sound-shape-of-
language-3rd-ed-with-a-new-preface-edition-linda-r-waugh/

OR CLICK BUTTON

DOWLOAD EBOOK

Get Instant Ebook Downloads – Browse at https://ebookgate.com

Click here to visit ebookgate.com and download ebook now
Instant digital products (PDF, ePub, MOBI) available
Download now and explore formats that suit you...

The Sound Shape of Language 3rd ed., with a new preface

Edition Roman Jakobson

https://ebookgate.com/product/the-sound-shape-of-language-3rd-ed-with-
a-new-preface-edition-roman-jakobson/

ebookgate.com

Prosodic Phonology With a New Foreword 2nd ed. [with a new

preface] Edition Marina Nespor

https://ebookgate.com/product/prosodic-phonology-with-a-new-
foreword-2nd-ed-with-a-new-preface-edition-marina-nespor/

ebookgate.com

The Power of Identity 2nd Edition, with a New Preface

Edition Manuel Castells

https://ebookgate.com/product/the-power-of-identity-2nd-edition-with-
a-new-preface-edition-manuel-castells/

ebookgate.com

Mabel McKay Weaving the Dream First Edition, With A New

Preface Edition Greg Sarris

https://ebookgate.com/product/mabel-mckay-weaving-the-dream-first-
edition-with-a-new-preface-edition-greg-sarris/

ebookgate.com
The Shape of the Liturgy New Edition Dix

https://ebookgate.com/product/the-shape-of-the-liturgy-new-edition-
dix/

ebookgate.com

Politics Culture and Class in the French Revolution

Twentieth Anniversary Edition With a New Preface Lynn Hunt

https://ebookgate.com/product/politics-culture-and-class-in-the-
french-revolution-twentieth-anniversary-edition-with-a-new-preface-
lynn-hunt/
ebookgate.com

Khubilai Khan His Life and Times 20th Anniversary Edition

With a New Preface Morris Rossabi

https://ebookgate.com/product/khubilai-khan-his-life-and-times-20th-
anniversary-edition-with-a-new-preface-morris-rossabi/

ebookgate.com

Sound and Sentiment Birds Weeping Poetics and Song in

Kaluli Expression 3rd edition with a new introduction by
the author Steven Feld
https://ebookgate.com/product/sound-and-sentiment-birds-weeping-
poetics-and-song-in-kaluli-expression-3rd-edition-with-a-new-
introduction-by-the-author-steven-feld/
ebookgate.com

Power and Charity A Chinese Merchant Elite in Colonial

Hong Kong Second Edition, With A New Preface Edition
Elizabeth Sinn
https://ebookgate.com/product/power-and-charity-a-chinese-merchant-
elite-in-colonial-hong-kong-second-edition-with-a-new-preface-edition-
elizabeth-sinn/
ebookgate.com
Roman Jakobson
Linda R. Waugh

The Sound Shape

of Language
Assisted by Martha Taylor

Third edition, with a new preface by Linda R. Waugh

Mouton de Gruyter
Berlin · New York 2002
Mouton de Gruyter (formerly Mouton, The Hague)
is a Division of Walter de Gruyter GmbH & Co. KG, Berlin.
Second Edition: 1987.
This work was originally published in 1979
by Indiana University Press, Bloomington and London.

® Printed on acid-free paper which falls within the guidelines

of the ANSI to ensure permanence and durability.

Die Deutsche Bibliothek - CIP-Einheilsaufnähme

Jakobson, Roman:
The sound shape of language / Roman Jakobson ; Linda R.Waugh.
Assisted by Martha Taylor. - Berlin ; New York : Mouton de Gruy-
ter, 2002
ISBN 3-11-017285-2

© Copyright 2002 by Walter de Gruyter GmbH & Co. KG, 10785 Berlin.
All rights reserved, including those of translation into foreign languages. No part of this
book may be reproduced 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.
Typesetting: Georg Appl, Wemding.
Printing and binding: Werner Hildebrand, Berlin.
Cover design: Sigurd Wendland, Berlin.
Printed in Germany.
What fetters the mind and benumbs the spirit
is ever the dogged acceptance of absolutes.
Edward Sapir, 1924
The Grammarian and His
Language
 CONTENTS

ACKNOWLEDGMENTS XI
NOTE ON TRANSCRIPTION XII
PREFACE TO THE THIRD EDITION (Linda R. Waugh) XIII
PREFACE TO THE SECOND EDITION (Linda R. Waugh) 1
Chapter One
Speech Sounds and Their Tasks 7
I. Spoonerisms 7
II. Sense Discrimination 8
III. Homonymy 9
IV. Doublets 12
V. Early Search 13
VI. Invariance and Relativity 17
VII. Quest for Oppositions 22
VIII. Features and Phonemes 29
IX. Speech Sounds and the Brain 32
X. Redundancy 39
XI. Configurative Features 41
XII. Stylistic Variations 43
XIII. Physiognomic Indices 45
XIV. The Distinctive Features in Relation to the Other
Components of the Speech Sound 46
XV. The Identification of Distinctive Features 53
XVI. Sense Discrimination and Sense Determination 57
XVII. Autonomy and Integration 59
XVIII. Universals 60
XIX. Speech Perception 64
XX. Life and Language 67
XXI. Role of Learning 73
XXII. Speech and Visualized Language 74
XXIII. Multiformity and Conformism 77
XXIV. Inner Speech 81
VIII

Chapter Two
Quest for the Ultimate Constituents 83
I. To the Memory of Pierre Delattre 83
II. Vowel ~ Consonant 87
III. Syllabicity 89
IV. Markedness 92
V. Grave ~ Acute 95
VI. Production and Decoding 98
VII. Compact-Diffuse 101
VIII. Sharpness and Flatness 113
IX. Interrelation of Tonality Features 119
X. And What Now? 123

Chapter Three
The Network of Distinctive Features 125
I. Significance of the Distinctive Features 125
II. The Two Axes 128
III. Nasality 134
IV. Voiced ~ Voiceless and Tense ~ Lax 138
V. Strident- Mellow 142
VI. Consonantal Correspondences to the Prosodic Features . 145
VII. Vowel Harmony 149
VIII. Glides 153
IX. The Nascent Sound Shape 156
X. Dynamic Synchrony 168
XI. Vistas 176

Chapter Four
The Spell of Speech Sounds 181
I. Sound Symbolism 181
II. Synesthesia 191
III. Word Affinities 198
IV. Sound-Symbolic Ablaut 203
V. Speech Sounds in Mythopoeic Usage 208
VI. Verbal Taboo 211
VII. Glossolalia 214
VIII. Sound as the Basis of Verse 218
 IX

IX. Children's Verbal Art 220

X. Samsure'spoetiquephonisante Seen from Today 224
XL Inferences from a Cummings Poem 225
XII. Language and Poetry 233

AFTERWORD 235

APPENDIX ONE
The Role of Phonic Elements in Speech Perception 241

APPENDIX TWO
On the Sound Shape of Language: Mediacy and Immediacy
by Linda R. Waugh 255

REFERENCES 273

INDEX OF NAMES 317

INDEX OF LANGUAGES 325

INDEX OF TOPICS DISCUSSED 328

ILLUSTRATIONS
1. Magic runes from Bryggen 16
2. Neuroanatomical schema for the auditory asymmetries . . . . 35
3. Delattre's spectrographic pattern of French consonants . . . . 97
4. Spectrograms of American English ba,da, and ga 103
5. X-ray photographs of Czech vowels and consonants 106
6. Title page of C. F. Hellwag's dissertation of 1781 129
7. The vowel triangle in C.F. Hell wag's dissertation 130
8. A cubic graph of the Turkish vowel system 151
 ACKNOWLEDGMENTS

For support of our work on this book, we thank the Ford Foundation
and its director, McGeorge Bundy.
For gracious assistance in the preparation of our study, thanks are
due to the Massachusetts Institute of Technology; Cornell University;
Ossabaw Island Project and its president, Eleanor Torrey West; Per-
ception Technology Corporation and its director, Hussein Yilmaz; the
Bryggen Museum in Bergen, its director, curator Asbjörn Herteig, and
Aslak Liestol, curator of the Universitetets Oldsaksamling in Oslo; the
Department of Indo-Pacific Languages at the University of Hawaii;
Stanford University Phonology Archives; Harvard University Interli-
brary Loan; and the National Library of Medicine, Bethesda, Mary-
land.
For valuable advice and suggestions we are gratefully indebted to
our friends and colleagues: Milada Blekastad, Sheila Blumstein, Ro-
bert Blust, Dwight Bolinger, the late Jacob Bronowski, Noam Chom-
sky, George N. Clements, Helge Dyvik, Samuel Elbert, Donna Erick-
son, Rachel Erlich, Sigurd Fasting, Charles Ferguson, Eli Fi-
scher-Jorgensen, Ivan Fonagy, Joseph Greenberg, Morris Halle, Einar
Haugen, Charles Hockett, Marcia Howden, Andrew Kerek, Michael
Krauss, Peter Ladefoged, Alvin Libermann, Björn Lindblom, Andre
Malecot, Robert H. Maurer, Sven Öhman, Colin Painter, Donald Pre-
ziosi, Ronald Scollon, Michael Silverstein, Edward Stankiewicz, Ken-
neth Stevens, Michael Studdert-Kennedy, William S.-Y. Wang, Calvert
Watkins, David Waugh, Roger Wescott, Dean S. Worth, and in partic-
ular Gunnar Fant and the late Pierre Delattre. Permission to reproduce
illustrations and texts was kindly granted by Doreen Kimura and Cor-
tex, by S. Blumstein and K. Stevens, by Harcourt Brace Jovanovich, by
Granada Publishing Limited, and by Chappell Music Company.
 NOTE ON TRANSCRIPTION

For the vowels we use the official IPA system, except that «(and not y)
designates the high front rounded vowel corresponding to /'. Long
vowels are noted by a macron (e.g., ä), nasal vowels by a tilde (ä).
For the consonants we use the symbols t, d for dental or dentialveo-
lar stops where there is no further distinction. If there is a difference, t
and d are usable for the dental stop, and t° and d° for the alveolar. For
the palatal consonants - stops, fricatives, nasals - we use the principle
used by Slavistic literature and going back to the spelling system estab-
lished by Jan Hus for Czech: a v is placed over or at the top right of the
corresponding dental: , d~ s, , n, c, and j". The last two are palatal af-
fricates corresponding to the dental affricates c and z. (In some sys-
tems of transcription quoted here s and z are rendered as J and 3.) f is a
strident r (as in Czech), and ö are the tense (voiceless) and lax
(voiced) interdental nonstrident continuants. Secondary modifications
of consonants are noted in the following manner: palatalization by '
(e.g., t'); rounding, velarization, or pharyngealization by ° (e.g., t°); ret-
roflexion by. (e.g., t); aspiration by h (e.g., th); glottalization by ? (e.g.,
t?); devoicing by 0 (e.g., d); syllabicity by , (e.g., r).
 PREFACE TO THE THIRD EDITION

This third edition is dedicated

to the memory of Roman Jakobson and to the
furtherance of his intellectual legacy

Roman Jakobson and I began our work on what was to become

The Sound Shape of Language in January 1977. This third edition,
published as part of the jubilee celebration of the merger of Mouton
and de Gruyter, appears, therefore, on the 25th anniversary of our
writing the first paragraphs about spoonerisms, sense discrimina-
tion, homonymy, doublets, and so on. The book, as it emerged from
our collaboration, changed from a technical monograph (much like
Jakobson's earlier Preliminaries to Speech Analysis [Jakobson, Fant,
Halle 1952] and Fundamentals of Language [Jakobson, Halle 1956])
on the acoustic definitions of "the ultimate constituents of lan-
guage" (our first, provisional title for the work), to a shorter work
on the distinctive features and their structural interrelations, to the
present book with its insistence on both the structural and the func-
tional aspects of sound and on the propensity of distinctive features
to signal meaning directly (as discussed, in particular, in Chapter
Four). Ultimately, our aim was to address the vast scope of issues
relevant to the understanding of these smallest units of sound and
their combinations (phonemes), while taking account of new, rele-
vant research of the 1960's and 1970's on acoustic phonetics, speech
perception, language and the brain, neurolinguistics, the genetic
code, language change, language variation, language universals, the
sound systems of little known languages, child language acquisition,
sound symbolism, synesthesia, anagrams, children's play with lan-
guage, mythopoeic usage and other uses of language where sound
seems to cast a "spell" on its users. At the same time we paid ho-
mage to our predecessors, from antiquity to the first half of the 20th
century, whose insights into the sound pattern of language we hailed
as precursors to the present work.
Jakobson often claimed that it took twenty-five years for many
of his most original ideas to be heard and understood. Recent inter-
est in functionalist, emergentist, corpus, laboratory-phonological,
XIV PREFACE TO THE THIRD EDITION

discourse-pragmatic, interactional, anthropological, sociological

and socio-cognitive approaches to language is, I hope, a sign that
this third edition of The Sound Shape of Language will find new
sympathetic readers.

Linda R. Waugh
Tucson, Arizona
 PREFACE TO THE SECOND EDITION

Since the first edition of The Sound Shape of Language^ is out of

print and the demand for copies continues, this second, augmented
edition is being published. Minor, mostly typographical changes were
made in the text itself; and an index of names has been provided. I
have added a second appendix, my "On the Sound Shape of Lan-
guage: Mediacy and Immediacy". This article, which was written in
1979, after this book went to press, reviews one cluster of its basic
findings.
Roman Jakobson died on July 18, 1982. This second edition is de-
dicated to his memory.
In general, it could be said that this book represents the fourth
(and final) stage in Jakobson's quest to uncover the function and
structure of sound in language. In the first stage (1920's and 1930's),
he collaborated with the linguist Nikolai S. Trubetzkoy. Together,
they developed the concept of the phoneme as a functional element
and adumbrated various structural patterns which underlie phono-
logical systems. Their fundamental concepts were based on invar-
iance and contextual variation. Thus, phonemes are defined by those
constant characteristics which relate them to other phonemes of the
same system; and these common denominators are related to the var-
ious alterations which a phoneme manifests as it is used in different
contexts. Moreover, Jakobson and Trubetzkoy showed that there are
different types of phonological systems: e.g., triangular ( a ) vs.
quadrangular (a ae) vocalic systems. ,Λι
u i
Structural relations were also discerned among phonemes; the most
important of these were binary opposition (markedness relations) and

1
Published in 1979 by Indiana University Press and prepared with the assis-
tance of Martha Taylor.
2 PREFACE TO THE SECOND EDITION

implicational laws, especially asymmetric implication (X—>Y but not

vice versa). These new concepts were applied to language synchrony,
language typology (types of systems), and language diachrony
(change). Jakobson was also able to show that these structural pat-
terns and relations determine the order of acquisition of phonemes in
children: in general, simpler structures tend to be learned before
complex ones (unmarked before marked, triangular before quadran-
gular, Υ before X [if X—>Y]).
In the second stage (late 1930's and 1940's), Jakobson developed
the notion of binary distinctive features as the minimal building
blocks of language. He argued that phonemes are complex since they
are combinations of distinctive features, and that it is these features
which underlie the different types of phonological systems. More-
over, he came to the general conclusion that distinctive features are
signs whose signatum is "mere otherness" or pure differentiation;
thus language is a completely semiotic system, from the largest com-
ponents (texts) to the smallest ones (distinctive features).
In the third stage (1950's and 1960's), Jakobson collaborated with
the acoustician C.Gunnar Fant and his student, the linguist Morris
Halle, in developing acoustic definitions of the distinctive features
which function to differentiate words perceptually. They also promul-
gated the theoretical concept that there is a small set of features
which underlies the phonological systems of all the languages of the
world. They found that they could unify in one acoustic feature a
variety of articulatory properties (e. g., the feature of flatness encom-
passes labialization, rounding, velarization, and pharyngealization).
And they developed a more relational account of invariance: for ex-
ample, the /a/ in triangular systems is structurally, and thus linguisti-
cally, different from the /a/ in quadrangular systems.
The fourth stage is marked in particular by the present book (writ-
ten in 1977-78, but in preparation for a few years before that). In it
we attempted to give as broad a coverage as possible to the continuity
of Jakobsonian thought. But there are new themes as well - or new
perspectives on old ones. In general, this work underscores the im-
portance of functional concerns in the analysis of sound. Thus, there
is a recognition that everything in the speech sound plays some lin-
guistic role and that, therefore, there are redundant, configurative, ex-
pressive and physiognomic features in addition to the distinctive
ones. In fact, the sound shape as a whole is a linguistic creation, and
oppositions such as (phon)etic vs. (phon)emic are outmoded. All of
 PREFACE TO THE SECOND EDITION 3

this is shown to be corroborated by modern research on the hemis-

pheric specialization of the brain. Each of the distinctive features is
discussed in turn and redefined as rigorously as possible. In addition,
we further develop properties of feature systems: for example, the na-
ture and interconnection of the two basic axes, compact ·%. diffuse and
grave-^ acute; the interrelation of the tonality features (grave -N. acute,
flat·>-plain, sharp ~plain); consonantal correspondences to the pro-
sodic (vocalic) features; glides as prime examples of zero phonemes.
Many of these themes are developed, at least in part, as a reaction
to the work of other phoneticians and phonologists of our time, many
of whom were responding in turn to Jakobson's earlier work. This is
particularly true of generative phonology and of its leading practitio-
ners, Noam Chomsky and Morris Halle, and of their book, The
Sound Pattern of English (1968). In certain ways, the present book
provides a (usually implicit) corrective commentary to the generative
trend. Examples abound on almost every page. The most salient are
the following. We object to the abandonment of the notion of pho-
neme: even though the distinctive features are primordial, the pho-
neme has its place in language structure. Various attempts to replace
the original set of distinctive features proposed by Jakobson are eval-
uated and rejected. It is affirmed that invariance is crucial to any
analysis of sound and is to be applied rigorously at the level of the
feature. Features are defined in acoustic terms: articulatory means
are to be seen only in the light of their ends, namely their use to dis-
tinguish perceptually words which are different in meaning. The con-
cept of markedness is given a relational definition such that either
pole of an opposition (+ or —) is marked, depending on the context
in which it is found. This is correlated with the order of acquisition in
children, language change, and language typology and universals, es-
pecially implicational laws. And sound systems are defined as dy-
namic, heterogeneous, multiform structures in which time (older and
newer forms) and space (social and geographical variants) have a
semiotic value. Moreover, the interlacing of learning and innate struc-
tures in language acquisition is affirmed, with emphasis on the for-
mer; as is the centrality of dialogue for learning and usage, even for
our own inner speech (thinking).
Perhaps the most important tendency within current linguistics
which we sought to address is the widespread disregard for the func-
tional, pragmatic, social, and communicative basis of sound (and of
language in general). This was particularly evident in the abandon-
4 PREFACE TO THE SECOND EDITION

ment by generative phonology of the fundamental division between

two different functions of the distinctive features. The first is their use
to keep apart words which differ in meaning. This is distinctiveness
proper, which serves as the basis for the set of distinctive features giv-
en here. It is primary in all languages of the world and exhibits regu-
lar patterns across languages. The second function, which is built on
the first, includes the study that Jakobson and Trubetzkoy originally
called morphonology. But actually it encompasses questions not only
of the alternations of a word (or a morpheme), but also the informa-
tion that features supply about derivational and inflectional structure
and lexical and grammatical meaning. This is called "sense-determi-
nation": sound is necessarily linked with, and thus informs about,
meaning. The ways in which this is accomplished exhibit great diver-
sity across languages.
This distinction is crucial to an understanding of the diverse ways
in which sound functions in language. However, in much generative
phonology, these two functions have been collapsed into one, called
"distinctive"; and their study has been called "phonology". Because
of this, we once again attempt to define "distinctive" through the syn-
onymous expression "sense-discriminative": the capability of signal-
ing, with a probability near to 1.0, the semantic likeness or nonlike-
ness of two meaningful verbal units. And we avoid the words
"phonology" and "phonological" - terms which Jakobson himself
had helped to launch decades earlier with Trubetzkoy - because we
felt that their meaning had become too blurred.
The important, but widely neglected, theme of the relation be-
tween sound and given units of meaning, which in our original plan
occupied only a few lines in Chapter One, eventually became the top-
ic of Chapter Four. Our concern was to relativize the popular notion
of arbitrariness. This led to the definition of a new antinomy, namely,
mediacy vs. immediacy: the tendency for sound to have an indirect,
or mediated, relation to meaning (sense-discrimination) versus its
tendency to have direct, and thus immediate, signification. In the for-
mer, the distinctive features connect sound and meaning solely by vir-
tue of contiguity; but in the latter, the sound symbolism inherent in
these features sustains a similarity relation between sound and mean-
ing. As evidence of the "spell" which the sound shape may cast over
the language user, we explore domains as widespread as synesthesia,
word affinity relations, taboo substitutions, and glossolalia. But no-
where is the direct interplay of sound and meaning more salient than
 PREFACE TO THE SECOND EDITION 5

in poetry, which is based on similarity (and equivalence) at all levels.

In fact, poetry is the fullest accomplishment of the synthesis between
contiguity and similarity and the most important locus of linguistic
creativity. Thus we endeavor to affirm that any science of language
must find a proper place for the mutual implication of the two inse-
parable universals: Language and Poetry.

Linda R. Waugh
 CHAPTER ONE

SPEECH SOUNDS AND THEIR TASKS

A thing without oppositions

ipso facto does not exist.
Charles Sanders Peirce
Collected Papers, 1.457

I. SPOONERISMS

"May I sew you to another sheet?"

This is a joking substitute reportedly originated by the Reverend
William Spooner (1844-1930) for the familiar question: May I show
you to another seat? The mere exchange of two initial sibilants, one
hushing and the other hissing, or in other words, a metathesis of the
two opposite terms of the feature compact ~ diffuse, changes the
meaning of two words (show to sew and seat to sheet) and creates a
comic effect. In like fashion, the interchange of the two liquids r and /
in blade of grass produces the unusual braid of glass. Algernon Charles
Swinburne wrote in the 1890s to his cousin Mary Gordon Leith, the
dark lady of his ballads, in a childish code: "By Merest Dozen" instead
of My Dearest Cousin, using a playful metathesis of the initial conso-
nants: M.D.K. >B.M.D. Such reversals, labeled 'Spoonerisms', fre-
quently occur as simple slips of the tongue (cf. Mac Kay 1970a), but are
also widely used as intentional, "laboriously fabricated" humorous
constructions, customary in English (see Robbins 1966) and even more
so in French, where this device is known under the name contrepetene
(see, e.g., Etienne 1957). Spoonerisms transposing separate features,
such as the metathesis of compactness and diffuseness (velarity and
labiality) in Fromkin's (1966) example, "plear glue sky" in place of
clear blue sky, are one of the clear-cut testimonies for the free play of
'distinctive features'. The punlike metathesis of distinctive features
may serve to weld together words etymologically unrelated but close in
their sound and meaning, as in the chapter title of a scholarly book,
"Identification of Roles and Rules", which playfully confronts the
sound and meaning of the last two nouns. Wordplays recorded from
unwritten languages often cling to the same principle.
8 THE SOUND SHAPE OF LANGUAGE

II. SENSE DISCRIMINATION

Indisputably, the grammatical pattern of the sentence, the verbal

context of the words at issue, and the situation which surrounds the
given utterance prompt the hearer's apprehension of the actual sense
of the words so that he doesn't need to pick up all the constituents of
the sound sequence. Still, the degrees of expectancy are variable and
the addressee may suddenly be faced with a message requiring him to
pay full attention to the various distinctions possible:
It shows the strange zeal of the mad sailor with neither mobility
" showed " " deal" " bad tailor " " nobility
nor fashion.
" passion.
Such attention to single features is needed in order both to grasp the
vocabulary (words and lexical morphemes) and to assign to the gram-
matical morphemes such as affixes their meaning. It is true that in the
latter case listeners may find a substantial cue in the syntactic arrange-
ment of the sentence, but the choice between, for instance, the present
tense suffix [z] and the preterit suffix [d] in the utterance quoted still is
not helped by the syntactic context, and thus cannot be derived from
other representations of the sentence, notwithstanding the hearer's
knowledge of the grammatical rules. Given the two basic operations -
selection and combination - one may, when dealing with a construc-
tion such as children showed, conclude from the preceding word chil-
dren that showed and not shows is in question, but the subject it would
not constrain the selection. In this way, the 'paradigmatic' (selectional)
axis keeps its relevance independently of the 'syntagmatic' (combina-
tional) axis.
Such components as those which in the sentence discussed enable
the perceiverto distinguish zeal from deal, shows from showed, and sail-
or from tailor, or mad from bad, mobility from nobility, and fashion
from passion are termed 'distinctive features'. But the attribute 'distinc-
tive', apparently introduced by the astute English phonetician Henry
Sweet, is unfortunately sometimes confused with 'distinct'. "Distinc-
tive", that is "discriminative", "having the ability to distinguish", has
often been misinterpreted as "discriminable", "distinguishable", "ca-
pable of being easily perceived". It is true that "discriminative" com-
ponents of speech are readily perceivable, but not all particulars "dis-
tinguishable" in the speech sounds by the language-user may effect the
distinction of meanings.
 SPEECH SOUNDS AND THEIR TASKS 9

III. HOMONYMY

The primary task of distinctive features is the discrimination of ver-

bal meanings. It is, so to speak, their raison d'etre in the makeup of lan-
guage. Sometimes the (sense-)discriminative role is questioned by the-
oreticians whose chief argument against this function is provided by
the existence of homonyms. It is a priori dear, however, that a language
devoid of homonyms is conceivable, whereas a purely homonymic
language is a reductio ad absurdum. By differing in sound shape and
meaning, the surrounding words as a rule serve for the disambiguation
of homonyms and thus for their correct semantic interpretation by the
listener. When a language, such as Chinese, radically reduces the num-
ber of distinctive features and of their simultaneous and sequential
combinations and confines words to a uniform, monosyllabic model,
the number of homonyms among simple (noncompound) words in-
creases excessively (e.g., an ί in the 'fourth tone' offers no less than
thirty-eight words of totally noncognate meanings). But the question
"How is it in practice possible to use a language having so many hom-
ophones?" finds an exhaustive answer in Bernhard Karlgren's classic
study (1962). The language frames such simple words into different
types of "elucidative compounds" by adding to a given monosyllable a
synonymous or generic apposition (cf. the Chinese pidgin English
look-see), by accompanying the monosyllabic verb with its tautological
object (cf. / eat food), or finally by prefixing the noun with a semantic
'classifier' (such as "mouth" as a classifier for objects with a round
opening or "branch" for long objects). (See also Klima 1975:262 ff.) In
all such cases the basic goal of these devices is the buildup of a wider
range for a possible application of the sense-discriminative features, in
order to overcome any annoying ambiguity.
When no help is given by the verbal context (cf. Appendix I), it is
often the nonverbalized situation which provides the clue. Homonyms
within a spoken sentence can have a double sense for the listener in
such sentences as The children had a \p£r\. Three words may remain
ambiguous: my children, or some children in general? ate, or merely
possessed! pair, or pearl The reason for the ambiguity lies in the elliptic
character of this example, translatable into such explicit sentences as
My children (or the children I know) each ate (or possessed) a pear (or a
pair of shoes or pants or socks, and so on). Most of the questions of ver-
bal ambiguity cited by linguists belong to various degrees of elliptic
speech. The phenomenon of ellipsis remains largely minimized or dis-
10 THE SOUND SHAPE OF LANGUAGE

regarded in the science of language, despite the insistent and repeated

suggestions of the sixteenth-century Spanish linguist Franciscus Sanc-
tius Brocensis, who clearly viewed explicitness and ellipticity as two
extreme aspects of linguistic operations. They may be seen as two
poles of the verbal code, because the scale of transitions from the ex-
plicit makeup of language to the elliptic offers a set of ordered rules.
Rules of gradual omission have a specific order in each language, just
as they differ in such customary patterns as folkdress. In contradistinc-
tion to some Western habits, Russian rural tradition allowed people to
go outdoors barefoot; however, a married woman was required to
have her head covered and a man to wear a belt when outside. This ex-
plains such pejorative Russian verbs as raspojasat'sja ('to stay ungir-
dled, throw aside all restraint') and oprostovolosit'sja ('to make a fool of
oneself, literally referring to an improperly bared woman's head).
The convenience of the omission of linguistic elements which are
redundant for the comprehension of the message is by no means the
sole impetus for ellipsis. Socially or individually tabooed words and
expressions are currently censored and excluded from speech, or are at
least camouflaged, for instance by being distorted in their sound shape
(cf. below, pp. 211 f.), and it is worthy of note that words tabooed and
omitted by a patient may be deemed highly significant when recovered
by psychoanalysis (cf. Lacan 1966).
We are constantly faced with elliptic structures, and thorough or-
ders of rules underlie the different degrees of ellipticity on each level
of language, from discourse through syntax and morphology and fi-
nally to reductions in sound sequences and concurrences of distinctive
features. The elliptic abbreviation of the English conjunction and first
entails the reduction of the initial vowel, then the abolition of the final
d, next the loss of the vowel (with concomitant syllabification of ri) af-
ter a terminal consonant of the preceding word, and finally, in a post-
vocalic position, the possible loss of n with the assignment of nasality
to the final vowel of the preceding word.
Some linguistic ambiguities cited by scholars are actually confined
to the elliptic variety of oral language and may find support in the writ-
ten form, which does not take into account intonational and pausal
distinctions characteristic of explicit oral speech but omissible in ellip-
tic style. Hence, in the instructive example discussed by Noam Chom-
sky (1972:104f., 162f.) - "John is certain that Bill will leave", and "John
is certain to leave"- a pause may occur after "is certain" in the first sen-
tence but before the same words in the second; thus in their explicit oral
 SPEECH SOUNDS AND THEIR TASKS 11

shape the so-called "superficial similarity" of these two sentences dis-

appears.
Any elliptic message is translatable by the speaker into a more ex-
plicit statement and vice versa. Thus, with respect to its two varieties,
the elliptic and the explicit subcodes, the overall code may be charac-
terized as 'convertible'. Fra^ois Dell (1973b), who pays close atten-
tion to "styles of diction", namely 'careful style' ("diction soignee") on
the one hand and 'casual speech' ("conversation familiere") on the oth-
er, cites the curious example of two French phrases a votre tour and a
votre retour; in elliptic style they become avottour and avotretour re-
spectively, so that the elliptic variant of the phrase a votre retour coin-
cides with the explicit form of the other phrase. The allegro variety of
American English is full of such elliptic phonations as the frequently
quoted ten min sem 'ten minutes to seven', or jijcet 'did you eat yet?'
and produces numerous homonyms, such as gone and going, put him
and put them, allude and elude, affect and effect, homonyms normally
disambiguated in lento speech.
One is apt to designate the explicit variety of language as the opti-
mal facet of the code. It is in fact optimal with respect to the plenitude
of informational distinctions; however, on the other hand, explicitness
risks carrying too much redundant information in the flow of speech
when the topic of discourse is familiar to the interlocutors. In such sit-
uations the economy of the tersely elliptic exchange of messages is pre-
ferred by both addressers and addressees.
Besides codified, conventionalized elliptic omissions, there are oc-
casional, individual inadvertences in both the speaker's and the listen-
er's performance. One could label the latter 'elliptic perception', which
causes a loss of features, words, and syntactic units in the reception of
the message; some of the signals omitted by the speaker as well as by
the hearer are restored and correctly identified by the latter thanks to
the context which he succeeds in picking up, while others remain un-
construed or even misunderstood. Familiar words or even larger
wholes are easily recognized in spite of emissive or perceptual ellipsis
and hence enable the recipient of the message to apprehend immedi-
ately and directly many higher units without the need for prior atten-
tion to their components.
12 THE SOUND SHAPE OF LANGUAGE

IV. DOUBLETS

Besides homonyms, another peculiarity of language has sometimes

been drawn in as evidence against the generality of the sense-discrim-
inative function of the distinctive features: namely, in single raris-
sime cases, which Ferdinand de Saussure (1857-1913) in his Geneva
lectures used to call 'linguistic dust' ("poussiere linguistique"), as his
student Karcevskij recalled, there occur among the tens or even hun-
dreds of thousands of vocables of a given language a handful of spo-
radic doublets with no semantic difference. Yet the occurrence of lexi-
cal doublets such as the two noticeable examples discussed by Sebas-
tian K. Shaumyan (1965: 24 f.) are quite exceptional. Particularly rare
(probability near zero) is the use of the two doublets by one and the
same member of the speech community without any semantic or at
least stylistic differentiation. Of the instances cited by Shaumyan, the
pair skap/skaf^cupboard, wardrobe', etc.) goes back to two different
borrowings from German; with respect to the latter form, the /f/ is still
felt to be a foreign phoneme at the end of Russian nouns (cf. RJ I:
728 ff.), and individual use tends to differentiate both forms. The other
pair, okolotok/okolodok ('ward'), both with the same cluster [tk] in the
genitive, etc., seems, whatever its origin, to be based on two different
folk etymologies, the first alluding to a fence (cf. kol 'stake, picket',
okolotit' 'to fence') and the second to a heavy barrier (cf. koloda 'log,
block' and the association between okolodocnyj 'police officer' and kol-
odki, -dok'shackles, fetters, stocks'; cf. also zakolodit"to shackle').
Usually such coexisting doublets belong to different dialectal varie-
ties, as is emphasized in the jocular song by Ira and George Gershwin:
You say eether [I] and I say eyether [ay],
You say neether [I] and I say nyther [ay];
Eether [I], eyether [ay], neether [I], nyther [ay],
Let's call the whole thing off!
You like potato [e] and I like potahto [a],
You like tomato [e] and I like tomahto [a];
Potato [e], potahto [a], tomato [6], tomahto [a]!
Let's call the whole thing off!
* * *
So, if you like pajamas [ae] and I like pajahmas [a],
I'll wear pajamas [ae] and give up pajahmas [a].
For we know we need each other,
So we better call the calling off off.
Let's call the whole thing off!
 SPEECH SOUNDS AND THEIR TASKS 13

Such doublets as either [ay]/[I] are felt by speakers as proceeding

from different dialects, or else, when used by the same speaker, take on
differing emotional tinges. For instance, some users of both varieties
feel that either with [ay] is the more prestigious variant and is thus en-
dowed with a more elevated significance. Within the same dialogue by
the same speaker one would rarely find the use of doublets different in
their sound shape but with no functional difference in their applica-
tion. The minimal differentiation of the two similar forms is their dissim-
ilation within a narrow context. When such an interdialectal doublet
is enrooted in the use of one speech community, the meaning-giving
capacity of the distinctive features endeavors to assign two different
lexical meanings to the members of the doublet: e.g., Polish dziewka,
dziwka, the former, with an [ε], meaning 'girl' or 'servant girl', the latter,
with an [i], taking on the specialized designation of 'whore'. Curiously
enough, Czech uses the two corresponding variants of the same word
in the opposite semantic distribution: divka ([I]) 'girl' - devka ([ε])
'whore'. Cf. baba 'grandmother, grandma', and baba Old crone, old
woman, old hen'. As in most of the essential questions of linguistic
analysis, we must when dealing with the semantic uses of distinctive
features operate with usability, the capacity to be utilized, as a concept
which implies certainty (probability 1.0), whereas the actual use of any
features for semantic discrimination can admit apparent exceptions
and thus displays a probability less than but very near to 1.0.

V. EARLY SEARCH

The international quest for the ultimate constituents of language

able to serve to discriminate meaning has endured through millennia.
Its first steps are manifested in the history of writing systems, with their
gradual dissociation of the speech sequence into words, syllables, and
finally the shortest successive segments. This last step was first attained
over three thousand years ago in the Aramaic and subsequently in the
Greek alphabetic script. "It has been rightfully stated that in the oldest
Greek inscriptions there occur but a few abbreviations as compared
for example with Latin, and that early Greek writing shows no disposi-
tion to ligatures; each letter remains separate and retains its atomic
character, its own place, its individual validity" (Patocka 1964:48). The
consistent segmentation of the Greek alphabet into single sound units
14 THE SOUND SHAPE OF LANGUAGE

was a powerful inspiration for the theory of ultimate, discrete verbal

constituents.
In the Greek philosophical literature indivisible sound units capa-
ble of forming meaningful strings were termed STOICHEIA, 'the under-
lying primes of sounds and letters'. The sound shape of language and
correspondingly its alphabet were viewed as a joint coherent system
with a limited number of discrete and interconnected formal units.
This concept proved to be so persuasive that Democritus (fragment
A6; cf. Diels and Wilpert) and his adherent Lucretius, in searching for
an analogy which might confirm their theory of the atomic structure of
the physical universe, cited STOICHEIA as the minimal components of
speech. Thus Lucretius, in his philosophical poem De rerum natura
(Liber II, verses 694 ff.) teaches:
Sic aliis in rebus item communia multa,
Multarum rerum cum sint Primordia longe
Dissimili tamen inter se consistere summa
Possunt: ut merito ex aliis constare ferantur
Humanum genus, ac Fruges, Arbustaque laeta.
("So likewise in other matters, many elements,
as they are the primary principles of many things,
may yet exist in dissimilar combinations among
themselves; so that the human race and the fruits
of the earth and the rich groves may justly be
considered to consist each of distinct original-
particles.")

Just as in Antiquity the term STOICHEION, used originally for linguis-

tic elemental units, was extended to the physical world, in a similar
way, but reversely, linguistic theory of the last hundred years in its
quest for the ulimate constituents has appealed in turn to the model of
atomic physics.
In Plato's Dialogues and Aristotle's writings some focal concepts
and problems of modern discussions on the theory of language (cf.
Koller 1955,1959; Burkert 1959; Moravcsik 1960a; Gallop 1963; De-
mos 1964; and Patooka 1964) are anticipated. The verbal STOICHEIA are
"distinguishables not detachables; abstractables not extractables" (see
Ryle 1960:436). To the cardinal tasks pointed out by Plato with regard
to the STOICHEIA belong their identification, recognition, classification
within the system, and an exact rendering of their compatibility and in-
terwovenness (namely the hierarchy of the rules for their combinability
into meaningful sequences). In his dialogue Philebus Plato states that
 SPEECH SOUNDS AND THEIR TASKS 15

none of us could learn any STOICHEION by itself without learning all of

the members of the given SYSTEMA and their mutual bonds. Both the
account of the distribution of elements in the sequence (SYMPLOKE)
and, first and foremost, their underlying classification on the basis of
their distinctive properties (DYNAMEIS) are foreseen in Plato's attempt
"to discriminate the DYNAMEIS of the STOICHEIA" (Kratylos, D 424).
The STOICHEION as such is for the philosopher a relational unit denot-
ing otherness and, to cite Julius Moravcsik's comment on the dialogue
Sophist (1960a: 48 f.), "whatever is Other', it is always other than some-
thing else." According to Plato, it is the art of grammar that enables us
to recognize the elements, to separate them from each other, and to dis-
cern their appropriateness for combinations.
In the theory of language richly and widely developed for many
centuries by the Sanskrit grammarians, the concept of sphofa, inter-
preted as the relation between form and meaning in language on its
various levels, occupied a pivotal position (cf. Brough; Ruegg; and Iyer)
and gave rise to vivid and broad discussions producing such major
achievements as the treatises of Patanjali in the second century B.C.
and Bhartrhari in the fifth century A.D. - that fixed attention to empiri-
cal data which was proper to the Indie linguistic tradition finds a pat-
ent expression in Patanjali's 'minimal pairs test', according to the cur-
rent American terminology - yüpa/cüpa/süpa (see Allen 1953: 81) -
and on the other hand this scholarly tradition daringly touched upon
the fundamental puzzles of linguistic theory such as the perennial an-
tinomies between permanence and mutability or between real percepts
of the language user and heuristic constructs imposed by grammarians
upon the verbal material, and it was Indie theory that brought to dis-
cussion the paradoxical status of units which by themselves are devoid
of meaning but at the same time are endowed with an indispensable
significative value.
The problems of the relationship between a verbal sound shape and
its meaning, far from being abandoned, gave rise to inquisitive theoret-
ical approaches in the philosophy of the Middle Ages (cf. RJ 1975:
292 f.) and to new, striking examples of concrete experimentation with
script. Thus an ingenious insight into the discrete ultimate segments
and their autonomous sense-discriminative value is characteristic of
the Old Norse tradition. Besides the entangled sound/meaning prob-
lems hidden in skaldic poetry and poetic treatises, one may quote a
few simple examples. One of the terse medieval runic texts recently un-
earthed in Bergen (see Liestol 1964:18 f.) constitutes a set of seven sev-
16 THE SOUND SHAPE OF LANGUAGE

en-letter words differing only in their initial consonant - mistill, tistill,

pistill, kistill, ristill, gistill, and bistill - written in such a way as to point
out the difference of the initials and the identity of each further seg-
ment, i. e., a 'commutation test' with a magical purpose (see Figure 1):

_L

Figure 1.
Magic runes from Bryggen (Bergen, Norway),
mtpkrgb iiiiiiissssssstttttttiiiiiiilllllll

An anonymous Icelandic grammatical treatise of the twelfth centu-

ry arguing for a revised adaptation of the Latin alphabet for the local
language resorts, as is shown in the excellent edition by Einar Haugen
(19722), to the device of 'minimal pairs' in order to establish the reper-
tory of sense-discriminative vowels, each of which, because of the
function it performs, requires a separate graphic designation. Thus, to
split off vocalic "distinctions [grein] each of which changes the mean-
ings", the author places each of the numerous vowels he finds in his
language "between the same two consonants" in order to show which
of these vowels "placed in the same position make[s] a different sense".
The next step of the treatise is to distinguish those vowels which are
"spoken in the nose" by placing a dot above them, because this vocalic
distinction appears to him to be systematic and to show "that it too can
change the meaning". Exemplifying sentences are used to justify this
attempt: "har ['hair'] grows on living creatures, but har ['shark'] is a
fish", etc. Another "distinction which changes the meaning" is the di-
vision of vowels into long and short ones; the treatise proposes to
 SPEECH SOUNDS AND THEIR TASKS 17

"mark the long ones with a stroke to distinguish them from the short
ones", and again minimal pairs and illustrative sentences follow: "far
['vessel'] is a kind of ship, but far ['danger'] is a kind of distress". In this
way, as the author states, all thirty-six distinctive vowels of the Icelan-
dic language are taken care of in a new "alphabet for us Icelanders".
On the other hand, the treatise makes use of one and the same letter for
cognate sounds without distinctive capacity, e.g. for the voiced and
voiceless interdentale.
Many remarkable theoretical and empirical assertions of the past,
after suffering long-term disregard and oblivion, reappear, often with
no reference to the original model, and turn into new, effective propo-
sitions. Such was, for instance, the historical destiny of the 2000-year-
old Stoic thesis which treated the sign, SEMEION, as an entity constitut-
ed by the two correlatives: the SEMAINON ('the signifier') and the sE-
MAINOMENON ('the signified'). In the last semester of his series of
courses in linguistic theory, Saussure took over and emphatically rec-
ommended this formula - "le signifiant et le signifie sont les deux ele-
ments composant le signe" (1916: 152; cf. Golebiewski) - and it en-
tered into his posthumous Cours de linguistique generate, compiled by
Saussure's disciples Charles Bally (1865-1947) and Albert Sechehaye
(1870-1946) and published in 1916. This thesis, often mistakenly
viewed as an invention of the Genevan, is unsurpassed because of its
clear ascertainment of the two semiotic constituents, one (SEMAINON,
'signans', signifiant) directly given and the other (SEMAINOMENON,
'signatum', signifie) prompted by the first. Both abstract and concrete
questions of the relationship between signans and signatum in the
realm of signs (signa) and especially in the various aspects of language
belong to the continuously increasing penetrations into the cultural life
of humanity which engender ever-new solutions and ever-new puzzles.

VI, I N V A R I A N C E AND RELATIVITY

Consistent efforts toward the delineation and identification of

those further-indivisible constituents of language which serve to differ-
entiate fundamental linguistic units endowed with their own meaning
such as words and their meaningful components ('morphemes') was
begun in the 1870s by a few young pioneers scattered throughout dif-
ferent countries. Three of these far-sighted scholars, all three born in
the mid-1840s - the Englishman Henry Sweet (1845-1912), the Pole
18 THE SOUND SHAPE OF LANGUAGE

Jan Baudouin de Courtenay (1845-1929), and the Swiss Jost Winteler

(1846-1929) - resolutely advanced the basic question of the relation-
ship between sound and meaning. From 1877 on, Sweet outspokenly
separated the "independently significant" sounds which may corre-
spond to differences of meaning from all other "differences which are
not significant and cannot alter the meaning" (cf. RJ II: 456 ff.). As ear-
ly as 1869 Baudouin broached the subject of the differences in sounds
which are "used to differentiate meanings" (see 1974: 258), and the
elaboration of "connections between sounds and meaning" became
the main concern of his research and teaching through the next de-
cades (see 1963 I; cf. RJ II: 394ff.; Stankiewicz 1972). In his thesis of
1876, Winteler unswervingly distinguished variations labeled "accid-
ental features" from "essential properties", sense-discriminative in-
variants on the sound level of language. In order to extract and identify
these invariants he made deliberate use of minimal pairs (cf. RJ 1972)
and was followed in this technique by Sweet.
No matter whether there was a convergence in thought or whether
linguists received a new impetus from mathematics, the seminal idea
of invariance developed into the dominant principle for both fields of
knowledge, especially in the 1870s. In either case, according to Felix
Klein's (1849-1925) Erlanger Programm of 1871-72, "the 'given' is a
multiplicity and within it a transformational group; the patterns to
which this multiplicity is related have to be investigated with respect to
those properties which remain unaffected by the transformations of
the group" (1921: 463).
This view found a series of close correspondences in the broad aims
and purposes outlined in Baudouin de Courtenay's introductory lec-
ture, published at the same time as the Erlanger text (see 1963 1:47 ff.),
and in the comprehensive programs of Baudouin's linguistic courses,
delivered chiefly during the next ten years and printed in the same Bul-
letins (see 1963 I: 78ff.) of Kazan' University in which half a century
earlier Nikolaj Lobacevskij had issued his epochal sketch of a non-
Euclidean geometry. Baudouin stated that "we have to put aside the
divergent, accidental properties of individual sounds and substitute a
general expression for the mobile sounds - an expression that is, so to
speak, the common denominator of these variables" (1963 1:120). The
linguist's aim was to discover the relational invariants in the flux of
speech, with its countless contextual and optional sound variations.
He predicted also that the acoustic correlates of these linguistic invari-
ants would be more precisely determined when speech analysis could
 SPEECH SOUNDS AND THEIR TASKS 19

utilize such technical achievements as the telephone, microphone, and

electricity and could obtain a visual portrayal of sound waves (Baudou-
in 1881-82: 4 f., 61 f., 65). This prediction, which up to our time is still
objected to by regressive dogmatists, found a convinced advocate in
Leonard Bloomfield (1887-1949), who in 1934 foresaw that "the phys-
ical (acoustic) definition of each phoneme of any given dialect can be
expected to come from the laboratory within the next decades" (see
Twaddelll935:23).
Both to Baudouin and to the prematurely deceased Mikolaj Kru-
szewski (1851-1887), Baudouin's omniscient disciple and uncompro-
mising collaborator (cf. RJ II: 428 ff.), it was clear that every linguistic
unit occurs in diverse modifying environments and that all its occur-
rences are equivalent to each other: i.e., in the terms of 'Group The-
ory', they are merely different expressions of one and the same linguis-
tic essence. Baudouin approached this extracted, purely relational unit
as an indivisible phonetic particle (or 'phoneme' in his later terminolo-
gy), comparable with the atom as the unit of matter and with 1.0 as the
unit of mathematics.
Since both penetrating Polish explorers, Baudouin and Kruszewski,
were aware of invariances in relationship, the connection between the
initial aspirated // of English and the final unaspirated /?, parallel to an
equivalent connection between the other initial and final stops within
that language, quite naturally authorized the assignment of both vari-
ants to the same phoneme. The identical relation between the aspirat-
ed initial stops and the unaspirated final stops cannot serve in English
to discriminate meanings of two words because the two sounds do not
occur in the same position (in contradistinction to those languages,
such as the Swiss-German dialect described by Winteler (1876), in
which [th] and [t] appear in the same position and carry a sense-discrim-
inative capacity). On the other hand, the feature distinguishing [ph]
from [th] or [p] from [t], all other things being equal, is capable of signal-
ing such lexical pairs as pot and tot, top and tot, pill and ////, pick and
tick, lip and lit in English, and thus [p] and [t] function as two different
phonemes.
Both Baudouin and Winteler, when dealing with sound patterns of
diverse languages, invoked verbatim the principle of "relativity [otnosi-
tel'nost'} of sound categories" (Baudouin, see 1963 1: 80), since relativi-
ty and invariance are necessarily two complementary concepts, or as
physicists say, "the reverse side of invariance {...) is called relativity"
(Morgenau 1961: 82). Gunnar Pant's insistent reminder (1973: 163) to
20 THE SOUND SHAPE OF LANGUAGE

speech sound analysts of different training - "the invariance is gener-

ally relative rather than absolute" - becomes more and more oppor-
tune. From the early 1870s Baudouin detected and pointed out "sets of
parallel sound oppositions \protivopoloznosti]" and emphasized "the
intimate connection of such oppositions with the meaning of words
and of their constituents" (see 1963 I: 80). Both Baudouin and Sweet
developed the same approach in similar ways not only in intralingual
but also in interlingual analysis. Thus, as stated in the program of Bau-
douin's university course for 1876-77, "within different languages
physiologically identical sounds may possess different values in con-
formity with the whole sound system, i. e. in correspondence with their
relations to the other sounds. Their scales are unlike" (1963 I: 90).
Sweet taught in 1876 not only that each language utilizes only a few
sound distinctions for differences of meaning, but moreover that there
are universal restrictions. Thus, if two vowels are "formed in a totally
different way" but are "never employed together in the same language
to distinguish the meanings of words {...} they may be considered as
variations of the same vowel". Herewith the extraction of invariants
from intralingual variations is supplemented by a search for interlin-
gual, universal variations and for corresponding invariants.
Saussure, obviously inspired by Kruszewski's and Baudouin's ex-
ample, concentrated on problems of the relation between sound and
meaning during the final, Geneva period of his academic activities. In
his comprehensive sketch for a phonetic treatise, drafted in the 1890s
and now part of the Harvard manuscript collection (see RJ 1: 743 ff.),
Saussure developed the view of speech production as a programmed,
intentional, anticipatory activity with an auditory, perceptual effect.
Correspondingly, he discussed the different aspects of phonemes and
advanced the thesis that "Phonemes = Valeur semiologique", or, in
other words, that the leading role is played by the "rapport entre le son
et 1'idee". Briefly, phonemes were treated here as simple signs which
endow "acoustic oppositions" ("oppositions acoustiques") with a
"sense-discriminative value" ("une valeur pour I'idee"). Therefore, such
solely positional differences as the divergence between the syllabic
and nonsyllabic / in Indo-European proves to be "devoid of any semi-
otic value" ("sans valeur semiologique") and both "semiotically equiv-
alent" specimens are to be assigned to one and the same phoneme and
may be transcribed by the same letter. In another note of this same
time, Saussure added that in the forms srutos, sreumen, sreuo, "the
phoneme 'u' appears in two acoustic forms" and reveals a "unity in di-
 SPEECH SOUNDS AND THEIR TASKS 21

versity". He emphasized that this identification should not be viewed

as a mere scientific construct. Herewith Saussure attempted to inaugu-
rate a new discipline, "semiologic phonetics", which later was tenta-
tively renamed "phonology" by his follower Albert Sechehaye in his
book of 1908, and this label, adopted in the early 1920s by the Prague
linguists, has entered into international use.
Thus, the astute warning of the young Baudouin de Courtenay
against "an unjustified and paralogical jump" from the treatment of
such semantic units as sentences, words, and their minimal grammati-
cal components ("morphemes" as he named them) to the gross sound
matter without any regard for its semiotic function has been corrobo-
rated by decades of linguistic research, and the question of the rela-
tionship between the two sides of verbal signs - the signans and the
signatum - has finally been extended to the sound units of language as
well. Saussure's view that any linguistic constituent is "une entite ä
deux faces" and that "any material unit only exists through the sense,
the function it is endowed with", has been echoed, discussed, and
worked through in the subsequent development of linguistics.
Habent suafata libelli: as one the landmarks in linguistic research
based on the conjunct concepts of relation and invariance, Saussure's
Cours de linguistique generate, in the version elaborated by his devotees
Bally and Sechehaye, appeared in 1916, the same year as the first edi-
tion of Einstein's General Theory of Relativity. It is noteworthy (see RJ
1972: 74) that in 1895, while Albert Einstein was a student of the can-
tonal school at Aarau near Zurich, he was a boarder in Jost Winteler's
household and was treated as a member of the family. Until the end of
his life Einstein paid high tribute to the memory of the "clairvoyant Pa-
pa Winteler" and when as an Aarau student he started those experi-
ments which gradually led him to the theory of relativity, he was aware
of the principle of "situational relativity" (Relativität der Verhältnisse]
enunciated in Winteler's historic dissertation of 1876.
"Sound Patterns in Language" (1925 [see 1949]) was Edward Sapir's
momentous contribution to the first issue of the first volume of the re-
view Language, published by the newborn Linguistic Society of Amer-
ica. This first American pathfinder (1884-1939) in the theoretical in-
sights into the sound shape of language said that "a speech sound is
not merely an articulation or an acoustic image, but material for sym-
bolic expression in an appropriate linguistic context"; and it was on
"the relational gaps between the sounds of a language" that Sapir put
the chief emphasis. Similarly, the topological idea that in any analysis
22 THE SOUND SHAPE OF LANGUAGE

of structure "it is not things that matter but the relations between
them", an idea which found a manifold expression in contemporane-
ous sciences and arts, was a main guide for the exponents of the
Prague Linguistic Circle, founded in 1926. They endeavored to derive
the characteristics of phonemes from the interrelations of these units
and in the "Project of Standardized Phonological Terminology" of
1930 they defined a 'phonological unit' as a term of an opposition. The
concept of Opposition' took on fundamental importance for the inqui-
ry into sound differences able to serve in a given language for the dif-
ferentiation of cognitive meanings. The question of the relationship
between the sense-discriminative units became the necessary require-
ment for any delineation of functional sound systems.

VII. QUEST FOR OPPOSITIONS

In the 1920s the analysis of sense-discriminative constituents of lan-

guage did not go beyond the successive segments of the sound se-
quence ; in other words, phonemes were viewed as those terms of op-
positions which cannot be dissociated into smaller oppositions (see for
instance RJ I: 8). This view was a corollary of the traditional, especial-
ly the enduring Saussurian, thesis ascribing a mere linearity to the
sounds of language which are measurable in one dimension only:
"c'est une ligne" (1916:157). This principle is based on a vicious circle:
sounds defined as segments of a temporal sequence are not supposed
to be perceived as simultaneous. The striking successivity of sound
units tempted one to disregard the simultaneity of their components,
even though hints about the idea of simultaneous features of sounds,
each of these elements carrying a sense-discriminative capacity, at
times must have attracted the attention of experts. Thus, Baudouin no-
ticed that "phonemes are not separate notes but chords composed of
severals elements" (1910) and later added in his Petersburg University
lectures that a semantic and morphological role is carried out not by
total and indivisible phonemes, but only by their fractional motor-au-
ditory "constituent elements", for which he coined a special Russian
label, kinakema(see 1963: 290). Similarly, Saussure's lectures in gener-
al linguistics stressed the need to resolve the phonemes into their dif-
ferential elements while taking into account the participation of nega-
tive factors as distinguished from positive ones, for instance the ab-
sence of nasal resonance in opposition to its presence (1916:110).
 SPEECH SOUNDS AND THEIR TASKS 23

The Preparatory Committee of the First International Congress of

Linguists, held in The Hague in 1928, raised the question of the meth-
ods appropriate to a comprehensive view of a given language; a pro-
position sent in answer from Prague outlined a particular class of signif-
icative differences, namely a phonological correlation "constituted by
a set of binary oppositions all of which are defined by a common crite-
rion conceivable apart from each couple of opposites": comparative
phonology was entrusted with the formulation of general laws which
underlie the interconnection of these correlations within any given
sound system. The division of the couples in question into the princip-
ium divisions and the common substratum which unifies the two mem-
bers of any opposition meant the first step in the componential analy-
sis of phonemes into their distinctive features. The first concrete appli-
cation of this prerequisite, the successful effort of Nikolaj S. Trubetz-
koy (1890-1938) to approach the vowel patterns of languages from
such a point of view, appeared in the first volume of the Travaux du
Cercle Linguistique de Prague (1929), which was dedicated to the First
International Congress of Slavists held that year in Prague.
After the intense discussions on the foundations of phonology at
the two aforementioned congresses, and then in a more comprehen-
sive way at the Prague International Phonological Conference of 1930,
the necessity for a consistent dissociation of a phoneme into its simul-
taneous components was made increasingly clear. At the beginning of
the 1930s, the Praguians designated the phoneme as a "set, bundle, to-
tality of those concurrent sound properties which are used in a given
language to distinguish words of unlike meaning". These properties
were tentatively labeled differential or distinctive qualities or attributes
and later were called 'distinctive features'. The English term emerged
in the American linguistic literature of 1933, when it was used by Ed-
ward Sapir in his encyclopedia article "Language" (see 1949: 25) and
by Leonard Bloomfield, who in his book Language oscillated on one
and the same page (1933: 79) between the earlier conception of the
phoneme as "a minimum unit of distinctive sound-feature" and its
newer, innovative definition as a "bundle of distinctive features". (On
contradictions in Bloomfield's use of the term 'distinctive feature' see
Twaddell 1935:19 ff.) With reference to the Praguian work on the dis-
sociation of phonemes into 'binary oppositions', Zellig Harris brought
to the attention of American linguists the fact that the focus of interest
had been shifted "toward discovering what are the differences among
the phonemes in terms of the relative speech-feature categories" (1951:
24 THE SOUND SHAPE OF LANGUAGE

146ff.), and the principle of basing such an analysis upon a relative ap-
proach met with his express approval.
Saussure's basic definition of differential units as "negative, relative
and oppositive" has been seminal. The idea of opposition as the pri-
mary logical operation universally arising in humans from the first
glimmerings of consciousness in infants and from youngsters' initial
steps in the buildup of language was viewed as the natural key to the
inquiry into verbal structure from its highest to its lowest levels. The in-
alienable property of opposition which separates it from all other, con-
tingent differences is, when we are dealing with one opposite, the obli-
gatory copresence of the other one in our minds, or in other words, the
impossibility of evoking long without a simultaneous, latent idea of
short, or expensive without cheap, 'voiced' without 'voiceless', and vice
versa, as was for the first time (1938,1939) brought to light and lucidly
demonstrated by the Dutch theoretician of language Hendrik Pos (cf.
below pp. 176 f.). When contending in the 1930s with the idea of oppo-
sition, linguists were unacquainted with Charles Sanders Peirce's
(1839-1914) writings, which offer an astute insight into the "particular
features of language" and into the concept of opposition: "The natural
classification takes place by dichotomies" (Peirce 1.437); "A dyad con-
sists of two subjects brought into oneness" (1.326); "Essential dyadic
relation: existence lies in opposition merely" (1.457). In Peirce's par-
lance "a dual relative term, such as 'lover', 'benefactor', 'servant' is a
common name signifying a pair of objects". Every "relative has also a
converse, produced by reversing the order of the members of the pair.
Thus the converse of 'lover' is 'loved' " (III.238 & 330).
Curiously enough, some American linguists insisted on the replace-
ment of the allegedly un-American term Opposition' by the less telling
and more ambiguous label 'contrast', despite the remarkable commen-
tary on opposition given by this most prominent American thinker. It
becomes a fundamental, operational concept not only for language (cf.
Ivanov 1974), but also for social structure in general (cf. Lorrain 1975;
Levi-Strauss 1958: 37 ff., 93 ff., 257 ff., 1963, and 1971: 240 ff., 498 ff.,
539f.; Parsons & Bales 1955; Blanche 1966; Fox 1974, 1975, 1977).
Franfois Lorrain opens his research on social systems by arguing for
"the privileged place of binary oppositions in the human mind" (p. 17).
In his dense essay on intellectual structures, R. Blanche (p.15) insists
on the importance of such oppositional structures: they should not be
underestimated, and in particular the organization of concepts by ad-
versative couples appears to be an original, permanent form of thought
 SPEECH SOUNDS AND THEIR TASKS 25

(cf. Ungeheuer). In Wallon's formulation, "le couple, en meme temps

qu'il oppose, unit" (1:117; cf. 75).
Saussure's and earlier Baudouin's (see above, p.20), recourse to the
idea of opposition was an efficacious event; however, this device was
not applicable to phonemes as wholes. The question "what is the op-
posite of the English [m]?" makes no sense. There is no unique oppo-
site. But the feature nasality finds its true and single opposite in the ab-
sence of nasality, as Saussure anticipated (see above, p.22): all other
things being equal, the nasality of [m] has its self-evident opposite in
the nonnasality of [b], or of [n] in [d], or of (French) [δ] in [o]. Step by
step it became clear that from the dichotomous analysis of 'correla-
tions' it was possible and indispensable to proceed to a similar, binary
dissociation of all phonemes into distinctive features. The direction of
inquiry was quite natural and had been foreseen in diverse patterns of
spelling, for instance in the diacritic signs used by the above-men-
tioned (see above, p. 15f.) Icelandic treatise (with strokes for the long
vowels and dots for the nasal ones), in the treatise on Czech spelling by
the many-sided reformer of the fifteenth century Jan Hus, who distin-
guished the palatal (compact acute) set of consonants by an inverted
circumflex in contradistinction to the corresponding diffuse conso-
nants (e. g., C, d", n, c, s, z), and especially in Enmun, the Korean vernac-
ular script of the same epoch with its radical division of letters into
particular strokes referring to the distinctive constituents of the native
phonemes.
A systematic search for what later, in the early 1950s, was metaphor-
ically described as the "elementary quanta of language" (cf. e.g. RJ II:
224) was devoted chiefly to the decomposition of consonants habitual-
ly aligned in the phonetic textbooks according to their place of articu-
lation from the bilabials to the uvulars and laryngeals. This linear ar-
rangement obstructed any questions of conceivable oppositions. It was
impossible to ask: "what is opposed to the dentality of [t] or to the bi-
labiality of [p]?" or "what is in dichotomous terms the relation between
labials and velars?" However, despite a striking distance on the string
of their point of articulation, labials and velars often show their close
connection in the history of phonetic changes from velars into labials
and vice versa.
Particularly familiar are mutual changes between the velar and la-
bial continuants - [x] and [fj and the corresponding voiced [γ] and [v] -
such as Jespersen's example of the English "jump" enough; or the
Celtic change of [ft] into [xt] (cf. Irish secht 'seven' with [pt] > [ft] > [xt]);
26 THE SOUND SHAPE OF LANGUAGE

and the Dutch shift to [xt] from [ft] with achter for after (Kaiser 1929:
119); or - for stops - the Celtic change of vesper into fescor; or the in-
terchange between [fj and [x] in various Slavic dialects, for instance in
Polish na ftorym, xtorek and Slovenian kozuch> kozuf, chruska>
fruska, krxka> krfka, plexko> plefko, and the reverse Czech substitu-
tion of the labial preposition by a velar one, [x] before voiceless obstru-
ents: ch Tumove, chpravit, ch Cechach and [y] before voiced ones: hbeh-
nou, h Vysokym, h Jablonci, or Russian dialectal forms with velars in-
stead of labials: dexka, verexka, xlex kuzox, x cerkox (v cerkov), krox,
oxtomik, and a corresponding change of [v] into [γ]: y lese, γ wy/ιί (v ug-
lu), ynuk, ydovec, y dom, as well as reverse changes of velars into la-
bials \fto, lefko, if,fodit',fajat'. The Russian dialectal changes of [g] or [γ]
into [v] may be exemplified by the masculine genitive ending of adjec-
tives and pronouns, -ovo, and also by such forms as povost from pogost
or mnovo from mnogo, as well as the exchange between [x] and [fj in
words of foreign origin, such as kufarka instead of kuxarka, xrancus in-
stead offrancus, and xrukt instead offrukt. In some Indonesian (name-
ly Eastern Toba-Batak) dialects, [p] in general changes into [k]: for in-
stance piso 'knife' > kiso (see Meillet & Cohen 1924: 418 f.). Also typi-
cal are the Czech dialectal variant kafez instead ofparezand such chil-
dren's oscillations between velars and labials as Czech telefon and vou-
sy changed into exon and xosi or in Peking Chinese xurj from fug (see
Ohnesorg 1959: 30,44).
It was especially in this and similar connections that the question of
a property joining labials with velars in opposition to the common fea-
ture of dentals and palatals, as well as the question of the feature com-
mon to labials and dentals in contradistinction to that of velars and
palatals, was brought into consideration, especially since each of these
paired groups displays characteristic interchanges in the history of
world languages. The preliminary solution to this complex of ques-
tions and to the problem of interconnections between the consonantal
and vocalic patterns was presented for discussion at the Third Interna-
tional Congress of Phonetic Sciences in Ghent in 1938 (See RJ I:
272 ff.), on the eve of the world events which put an end to the Prague
phonological deliberations and in general radically changed the to-
pography of international scientific activities.
Any notion of opposites is inseparable from the notion of opposi-
tion as such and neither of the two opposites can function in the neigh-
borhood of other concurrent or successive features if such a neighbor-
hood excludes the appearance of the other opposite. If instead of the
 SPEECH SOUNDS AND THEIR TASKS 27

explicit designation of an opposition, for example nasality ~ nonna-

sality, we use the abbreviation 'the nasality feature', ambiguities can-
not arise, because the feature is present only in those contexts where
both members of the opposition - nasal and nonnasal - are admissi-
ble. In Russian the opposition of voiced and voiceless obstruents func-
tions only before phonemes other than obstruents; therefore the voice-
lessness of the word-final consonant in [luk] loses its distinctiveness,
and the nominatives of the two nouns whose datives are [luku] 'to the
bow' and [lugu] 'to the meadow' become homonymous. The sound-
form [d'etks] corresponds to both d'etka 'kid' and d'edka 'grandpa';
and in the sentence - highly improbable precisely because of its fla-
grant homonymy - eto ne [d'etka} a [d'etka] prokazit 'it's not the young-
ster but the grandpa (or not the grandpa but the youngster) who plays
tricks', the question of the detected trickster remains ambiguous. Com-
pare cognate derivatives with an initial vowel of the second, antece-
dent suffix, such as [d'etuska] 'kiddy' and [d'eduska] 'grandpa'. It is
self-evident that to look for a distinctive feature in positions where no
distinction is possible is a gross contradiction in terms. If in certain
contexts only one of the opposites can appear, the feature loses its dis-
tinctiveness and becomes inactive and disabled. The opposition is
alive when both opposites are able to occur in the same context, given,
that is, the identity of concurrent and adjacent features.
The question of whether the dichotomous scale is inherent in the
structure of language or whether it is only a profitable principle for in-
quiry into that structure, a device imposed by the analyst upon the lin-
guistic material, has been repeatedly raised. The Russian linguist Vja-
ceslav V. Ivanov has devoted several penetrating studies to binary rela-
tions as an intrinsic qualitative property "permeating the entire system
of language and enabling one to describe it as a monolithic whole and
not an accumulation of scattered data" (1972). Far from being a mere
heuristic guide, the question of binary choices faces any interlocutor,
for example whenever it is necessary for the addresser to bring to the
knowledge of the addressee whether in the sentence cited near the be-
ginning of this chapter, deal or zeal, showed or shows, tailor or sailor,
etc., are meant. A choice between the two alternatives, stop or continu-
ant, through the recognition of the appropriate one, is of course de-
manded not only from the linguistic inquirer, but first and foremost
from the listener of the message, regardless of whether the identifica-
tional process is conscious or subliminal (cf. Muljacic 1977; Mel'cuk
1977: 292 fif.)·
28 THE SOUND SHAPE OF LANGUAGE

The difference between a single and a multiple distinction is pal-

pable for the speakers. Thus, there is a single binary distinction be-
tween seal and zeal; similarly, the distinction between seal and feel is
single but that between seal and veal is double, a simultaneous pair of
two binary distinctions; between zeal and dill there is a sequentially
double distinction - one in the initial consonant and the other in the
vowel. Finally, the discrimination of sill and bill is based on a triple
concurrent distinction. And again we must say that the difference be-
tween a single and a multiple distinction and between a simultaneous
and a sequential one is a palpable operation both for linguistic analysis
and for actual interlocutors. Lev Tolstoj acknowledged that he en-
dowed some characters of War and Peace with generally known aris-
tocratic Russian family names, which he changed only slightly, so that
they would sound familiar to his readership. In each case he played
with one single featural distinction: he made Bolkonskij from Volkons-
kij and Drubeckoj from Trubeckoj. These names sound quite natural,
whereas with a double distinction they would have become grotesque,
for instance *Dolkonskij or *Polkonskij or *Zrubeckoj. Children also
show an acute feeling for single distinctions. As Cukovskij (1966) re-
ported, a Russian child complained about the word doslyj ('shrewd'),
which he felt to be too close to doxlyj ('putrid'). Among various charac-
teristic examples, Willem Kaper (1959) quotes a four-year-old Dutch
boy who amused himself with the sound-meaning differences between
the rhyming words peertje 'little pear' and beertje 'little bear'. Another
Dutch boy, according to Kaper's report, announced with pleasure to
his mother that both Piet (proper name) and biet 'beet' really exist. A
similar pleasure is felt by children in the juxtaposition of an extant
word with an invented one distinguished by a single feature: pink is
bink, bink is pink.
Of course, any speech sound exists for its producer as well as for the
perceiver both as a whole and as a concurrence of those parts which
are imposed by the code of their language and imply single distinc-
tions. The parts are exemplified for the consonant [s] by its distinctions
from [z], [t], [f], and [s] in French and also from [Θ] in English and from
palatalized [s'] in Russian. In an analogous way, seal exists simultane-
ously for speakers of English both as a lexical whole and as a sequence
of its speech sounds, just as in turn a whole conventional sentence,
such as take it easy, necessarily coexists in the mind of speakers both as
a phraseological whole and as a set of single words. When discussing
the pattern of features (1973), the adept analyst of speech Gunnar Fant
 SPEECH SOUNDS AND THEIR TASKS 29

rightly viewed their pattern as "a matter of coding convenience only",

but the code and its conveniences are much more enrooted in the users
of language than in its explorers despite the fact that the users of the
verbal code may be unaccustomed to taking its conveniences into de-
liberate examination. For the sake of efficiency the perception of the
sense-discriminative cues naturally has recourse to the polar differen-
tiators facing the native decoder with a set of bare yes-or-no decisions
between any two members of binary oppositions. In this way, the need
for maximal simplicity, not only in the scientist's approach to the
sound pattern of language, but first and foremost in the daily strategy
of the language user, is fulfilled, especially since the number of opposi-
tions in any given language is prefabricated and strictly limited for the
apprehension of its speakers and perceivers.

VIII. FEATURES AND PHONEMES

It has been questioned whether linguistic operations with pho-

nemes would not be more advantageous than those dealing directly
with their ultimate constituents. The comparison of phonemes and dis-
tinctive features with respect to the most productive concept of opposi-
tion shows, as mentioned above (pp. 25 f.), that this concept, consis-
tently applicable to the distinctive features, cannot be extended to the
phoneme. A phoneme, being a bundle of distinctive features, proves to
be an ever important but derivative unit, a complex, simultaneous con-
struct of a set of elementary concurrent units. It can be compared in
this respect to the syllable, which is a constructive complex unit within
the verbal sequence. However, it is to be noted here that the sequence
implies concurrence, while the reverse is not always the case. For in-
stance, French vocabulary offers such homonymous words as /u/
'where' and /u/ Or'. And the Latin imperative /i/ 'go' may function as
an entire utterance.
The difficulty or even the impossibility of a consistent segmentation
of a sequence into phonemes has again and again been confirmed by
instrumental studies both on the motor and on the acoustic level. The
process of multifarious coarticulation was first thoroughly pursued by
Paul Menzerath (1883-1954) and was impressively demonstrated by
him with the help of an X-ray sound film at the Rome International
Congress of Linguists in 1933. Through manifold experiments and in-
strumental observations of the articulatory process of speech, he came
30 THE SOUND SHAPE OF LANGUAGE

to the conclusion that "a speech sound has no position; speech is a

continuous, uninterrupted movement, irregardless of whether the
sounds are vowels, diphthongs, or consonants, even plosives. All of
them are gliding sounds (...) A sound sequence in an articulatory sense
does not exist. The parts combined into a word prove to form not a
chain but an interweaving. An acoustically later sound may begin arti-
culatorally before the acoustically earlier sound" (Menzerath & Lacer-
da 1933).
These observations have recently been supplemented by the com-
prehensive acoustic, primarily spectrographic data of Haskins Labora-
tories. Yet, these impediments disappear as soon as we go from the lev-
el of phonemes to the segmentation of the speech sequence into the
succession of distinctive features. The diverse features can exhibit dif-
ferent limits of duration in the sequence, for they often capture a large
portion of the preceding phoneme or on the contrary begin in the mid-
dle of the phoneme to which they belong; they may spread over into
the next phoneme or cease in the middle of their own phoneme. The
relative order of these features, however, usually remains the same,
apart from rare and insignificant digressions. As a rule, the divergence
between the limits of the implementation of different features does not
cancel their concurrence in a certain at least minimal segment of the
string, so that localization of their co-occurrence is usually preserved,
except in the negligent, elliptic variant of speech. Thus, the segmenta-
tion of a sequence into successive features permits its further segmen-
tation into phonemes. The temporal order of phonemes remains an
undeniably significative factor - cf. Russian rvu Ί tear' vs. vm Ί lie', or
Czech vfe 'boils' vs. fve 'shouts' vs. rev 'roar'.
One cannot but agree that "we are a long way off from a really well-
developed model of the speech perception process" (cf. Pisoni 1975:
98), yet indeed direct recourse to "the basic primes" defended by Shei-
la Blumstein and William Cooper (1972: 208) is indispensable, for
"within any language system every phoneme is characterized by the
minimum number of features needed to distinguish it from all other
phonemes of that system", and such direct recourse ensures the most
exact application of those two vital principles of relativity and invar-
iance which underlie any present scientific task. The strategy of these
fundamental tasks will be touched upon below, but the direct ap-
proach to the listener's experience seems to be often a simpler pathway
than the search for the speaker's commands to the articulatory mus-
cles, especially since "we cannot at the present time observe those pro-
 SPEECH SOUNDS AND THEIR TASKS 31

cesses nor can we directly measure their output" (Liberman, Cooper, et

al. 1967:446 ff.); moreover, "when the temporally overlapping gestures
for successive phonemes involve more or less adjacent muscles that
control the same structures, it is of course more difficult to discover
whether there is invariance or not". In a recent survey volume edited
by Ljudmila Cistovic on theories of speech perception, the motor the-
ory is said to have been very popular some ten years ago; although this
theory gave hope for finding the key to speech perception by detecting
the "motor commands", the unjustified assumption of their standard
nature was superseded by proof of their high variability. It became ob-
vious that "in order to describe and model the process of governed
speech production, it is necessary to turn to such a notion as the aim of
the motor act, and the solely evident aim is the production of a definite
acoustic effect accessible and 'understandable' to the listener" (Cisto-
vic et al. 1976: 31 ff.).
At the beginning of our century, one of the most perspicacious
phoneticians of his time, Alexander Thomson (1860-1935), again and
again rejected the still ineradicable efforts to exaggerate the articulato-
ry criteria for the listener's speech interpretation: "It is not the move-
ments of the speech organs but the speech sounds themselves which
are primary in language. The sounds are a more uniform and more
constant element in language." Deprived of acoustic analysis, the
"physiology of any sounds whatever remains incomprehensible, since
the articulation is only a means to an end". Through the analysis of
consonants, Thomson brings to light the fact that "the dorsal or coro-
nal way of articulation, a difference which fallaciously is highly re-
garded, is by itself quite inessential for the sound and the choice is
chiefly determined by the height of the tone assigned to the noise"
(Thomson 1909 and 1934).
An additional reason for following the path of analysis from the
features to phonemes and not vice versa and for the advantages of a
featural transcription over a simple notation of mere unanalyzed pho-
nemes is the possibility of an unambiguous answer to the question of
the presence vs. absence of a given distinctive feature in its two terms,
whereas the makeup of a phoneme is different according to the con-
text. Thus for instance, in the Russian adverb zc/es"here', the initial
consonant lacks two distinctive features proper to the phonemes /z/,
/z'/, /s/, and /s'/ in prevocalic position, e.g. zad = [zat] 'back', zjat'
= [z'af] 'son-in-law', sad = [sat] 'garden', s/arf'[s'at'] 'sit down'. Hence
the initial sibilant appears here as an 'incomplete phoneme', because
32 THE SOUND SHAPE OF LANGUAGE

before obstruents the normally distinctive opposition of voicing and

voicelessness is canceled and likewise the opposition of palatalized
and nonpalatalized dentals is discarded before another dental. The
question of rendering such 'incomplete' phonemes in transcription
creates unavoidable complications and disagreements as long as spell-
ing is based not on features themselves but on entire phonemes. The
distinction between such tense and lax consonants as /t/ and /d/ or
/s/ and /z/ is lost in English after obstruents (cf. the final [d] and [z] in
ribbed and ribs and the final [t] and [s] in ripped and rips). In German
dialects where only [z] occurs initially and only [s] finally, the pho-
nemes /s/ and /z/ do not display the tense ~ lax distinction in other
than some internal positions (e. g., weisse and weise) and this distinc-
tion remains valid only as long as the antecedent diphthong [ay] in re-
lation to the following consonant does not carry its own significant dis-
tinction. In those positions where the tense ~ lax distinction is not op-
erative, the German hissing sibilant phoneme is quite incomplete: non-
distinctively lax in a prevocalic and nondistinctively tense in a postvo-
calic position. Cf. Saus [zaus] 'rush', gen. Sauses [zauzas].

IX. SPEECH SOUNDS AND THE BRAIN

The frequent French designation of distinctive features as traits per-

tinents easily leads to misconceptions; besides distinctive features the
sound shape of language contains a few other kinds of likewise perti-
nent, functional features. Only the degree and not the fact of their per-
tinence can be questioned. One may be an adamant 'featurist', but one
should not disregard the manifest copresence and functioning of
classes of features other than the distinctive ones. It should not be for-
gotten that speech sounds are tools of verbal communication and that
their entire makeup is an ensemble of diverse types of features, all of
which fulfill an interplay of tasks essential for communication. In the
process of communication, none of these features remains insignifi-
cant or immaterial. The obsolete but recurrent view of a phonetic de-
scription of articulatory and physico- and psycho-acoustic phenomena
as one which disregards their role in language and their communica-
tive significance not only inhibits inquiry, but above all arbitrarily sup-
presses the fundamental question of the manifold goals that these phe-
nomena pursue.
Such lack of respect for the multifold significance of sound shapes
 SPEECH SOUNDS AND THEIR TASKS 33

causes a dangerous truncation of the analyst's task and curbs rational

classification. Particularly vain is psychoacoustic, perceptual speech
analysis if done without regard for those diverse linguistic values of the
sensory stimuli which are picked up by the native perceiver, whose se-
lectional response to such stimuli depends precisely on their informa-
tional cues, most of them socially codified.
In light of the functional load of these sensory elements, which all
fulfill some semiotic duty, the view of phonetics as abstraction faite de
function now appears to us to be outdated, unrealistic, and emasculat-
ed. More and more we realize that speech sounds as a whole are an ar-
tifact built precisely for speech and are thus self-evidently goal-direct-
ed. The idea of "gross, raw" phonic matter, "amorphous substance", is
a fiction. Discrete articulated sounds did not exist before language,
and it is pointless and perverse to consider such "phonic stuff" with-
out reference to its linguistic utilization. The growth of language and
the development of the human supralaryngeal vocal apparatus are in-
terconnected innovations (cf. Lieberman 1975: 35); in particular the
hominid dental evolution turned the oral cavity into the best resonat-
ing chamber for linguistic use (Sheets 1977). And it is for their verbal
purpose that speech sounds were formed and submitted to a special
hierarchical organization.
The Old Indie theoreticians of language made a clear-cut distinc-
tion not only between varna sphota, the sense-discriminative constitu-
ent in speech sounds, and dhvani 'speech sound' in general, but also
between the latter and sabda 'nonspeech sound'. Any individual utter-
ing of a speech sound, which Bhartrhari named vaikrta-dhvani 'modi-
fied sound', presents differences of diction, but behind these fluctua-
tions inherent in given messages there is a codified, fixed sound-de-
sign, prakrta-dhvani 'primär}' sound', encompassing all those sound
properties which are normally produced by the speakers and per-
ceived by the listeners of a given speech community. Only some of
these integral properties, those connected with meaning, pertain to the
varna sphofa.
Thomas Aquinas opposed speech sounds, voces, to sounds emitted
naturaliter by animals. The former he characterized as voces signiß-
cantes, ex institutione humanae rationis et voluntatis. His definition of
speech sounds as significantia artificialiter and given ad significandum
appears to be the most valid one, especially now that a radical division
between speech and nonspeech sounds is becoming increasingly evi-
dent (cf. Manthey 1937). Sapir taught that "speech sounds exist merely
Discovering Diverse Content Through
Random Scribd Documents
appears to have acquired and used it[769]. It may be that this implied
the possession of the magic ring[770].
The circumstances of the find recall two With the story of
anecdotes in the Mirabiles the finding of the
[771]
Auscultationes , located the one in ring cp. stories of
finds in mining
Paionia the other in Pieria, in which, as in districts.
the Gyges story, we have the rains, the
chasms, the find of gold and the taking of it to the palace. In at least
one of the two cases the find is treasure trove, like the ring of Gyges.
But the point to be noticed is that both Paionia and Pieria are in the
famous mining district that played so great a part in the history of
Athens and Macedonia. Considering the fame of the mines of Lydia it
is natural to ask whether they did not play their part in the evolution
of the story of the ring. Stories of men buried in the mines of Lydia
did actually exist. One of them, told in the Mirabiles Auscultationes,
shows that in Lydia, as in the region of Mt Pangaion and the river
Strymon, chasms containing skeletons and gold were not unlikely to
be mine shafts, and the power secured by their discoverer to be
simply the power of suddenly and perhaps secretly acquired wealth.
As to the date at which Tmolus was first mined nothing is known
except that the workings were disused in the days of Strabo[772]. It
may well have been worked from the earliest days of Gyges if, as is
probable, he mined at Pergamus and further afield[773]. Lydians
mining in those regions would hardly be neglecting mines so near
their own capital, the more so as they would be directed to them by
following the golden stream of the Pactolus[774].
In the face of these facts it is tempting to go one step further than
Radet when he explains Gyges’ talisman as economic science. We are
reminded of Byron’s description of the most Rings as money.
potent of talismans:
Yes, ready money is Aladdin’s lamp[775].

Gyges, it is true, discovered a ring, not a lamp, but the particular

form of his talisman only adds point to the Byronic interpretation.
Till somewhere about the time of the story rings were probably ready
money in the literal sense of the phrase[776]. In many parts of the
world before the introduction of a regular stamped coinage trade had
been conducted largely by means of rings of specific weight[777].
Much of the evidence for this use takes us back well into the
second millennium. Some of the localities where it is known to have
prevailed earliest were connected with Lydia at this time. Egypt was
the ally of Gyges, the Hittites were the predecessors of the Lydians in
the land of Lydia. Troy probably formed part of Gyges’
dominions[778]. The book of Genesis specifically associates ring
money with caravans[779]. It is therefore probable that rings
circulated in Lydia itself until they were supplanted by the new
stamped coins, which may have owed their shape to its convenience
for stamping. In Argos, according to a With Gyges’ ring
tradition which is defended in the next cp. perhaps
chapter, the stamped coinage was preceded Pheidon’s spits,
the Attic obol
by a currency of metal spits (below, fig. 21): and drachma,
if the tradition is to be trusted, these spits and the Roman
remained permanently associated with the as.
name of the tyrant who displaced them by
coined money. Have we a parallel in the case of the Gyges story? Is
its hero the inventor of the new stamped coinage and has his name
become associated with the rings that he displaced? It is perfectly
possible that the new coins were at first regarded as so many rings’
worth of precious metal. The name νόμισμα implies that the stamped
coin had won general acceptance; it would of course be quite
appropriate to a currency in rings or kettles or cows; but there is no
evidence of its use previous to the introduction of a stamped metal
coinage, and it was probably the general acceptance of this latter that
gave rise to the word. To call the earliest Lydian coins rings
(δακτύλιοι) would be precisely like calling the earliest coins of
Athens spits (ὀβελοί) or bundles of spits (δραχμαί) or the earliest
coins of Rome bars (asses). If obol and drachma and as survived
while δακτύλιος did not, the fact is sufficiently accounted for by the
histories of Athens, Rome, and Lydia.
The explanation just offered of the ring of But if so, why is
Gyges omits one detail that is essential to the ring a seal
the story as told by Plato. Plato’s ring is a ring?
signet ring, and it is the seal (σφραγίς) that does the magic. Babelon
holds that the seal makers (δακτυλιογλύφοι) of the period were
probably also the coin strikers. But if we are to follow one of our own
leading numismatists, G. Macdonald, the connexion between seal
and coin was closer than this. Originally, according to him[780], coins
were simply pieces of sealed metal.
The minting of money in its most primitive form “Coins are pieces
was simply the placing of a seal on lumps of electrum of sealed metal.”
that had previously been weighed and adjusted to a
fixed standard: the excessive rarity of coins that have only a striated surface on the
obverse proves that this primitive stage was of short duration[781].
Greek seals were normally attached to rings[782]. The definition of a
coin as a piece of sealed metal was started by Burgon, who used it as
an argument for the religious character of coin types and described
the seal as “the impress of the symbol of the tutelar divinity of the
city[783].” Macdonald very rightly rejects the theory, so widely current
before the appearance of Ridgeway’s Metallic Currency, that all coin
types had a religious origin. But he himself proceeds to hang on to
the seal theory views of his own that are equally untenable, at least in
the sweeping form in which he states them. He assumes that the seal
must be always that of a state or king or magistrate, and that the
device is usually heraldic. The latter point does not concern us
here[784]: the former does, since it is incompatible with the views of
Babelon and Lenormant that the earliest coins were not state issues.
But the seal theory does not depend on the nature either of the seals
or of the sealers. There is no reason why we should not accept
Babelon and Lenormant on the private character of the earliest coins
simultaneously with Burgon and Macdonald on the character of the
first coin types as being simply seals. Indeed some of the evidence
collected by Macdonald positively suggests that we should. In the
fifth and fourth centuries B.C. at Athens public property was
stamped with the public seal, τῷ δημοσίῳ σημάντρῳ[785]. The same
practice was followed in the fifth century at Samos[786], and perhaps
at Syracuse[787]. This use of public seals to stamp public property
suggests corresponding private seals similarly used. At this point the
literary evidence fails us, but archaeological documents come to our
aid.
Metal was not the only material that the ancients stamped in this
way. Wine jars, bricks, and tiles were similarly stamped. Thousands
of these stamps have come down to us, and many of them bear city
symbols and a magistrate’s name. But they often bear also the name
of the maker, and the Danish scholar Nilsson[788], who has made a
thorough study of them, is inclined to regard them as essentially the
seals of private manufacturers. The magistrate’s name may merely
indicate the date: the state symbol may mean that the state was the
consignee, or that the manufacturer enjoyed some sort of state
protection, or was subjected to state taxation and control.
If then such was the history of the stamp as applied to bricks and
wine jars, we have a further reason for thinking that stamps as
applied to the precious metals may have originally been put on them
by private owners. May we not here have the true explanation of the
marvellous seal of Gyges’ ring? May not the owner of this ring have
been the first person to use his signet for stamping coins of metal,
and may not this fact be the origin of the stories about its marvellous
powers?
Whether this person was Gyges is another question. His claim to
the ring is not beyond dispute. We have seen already that it is
sometimes attributed to Midas. Even in the story as told by Plato the
MSS. make the hero “the ancestor of Gyges,” not Gyges himself, and
the fact that in other writers and elsewhere in the Republic[789] the
ring is called the ring of Gyges is not in itself decisive. Both the ring
and its magic powers may have passed from hand to hand. But if the
evidence and opinions that have been adduced in this chapter have
any value, our interpretation of the ring story does not depend on
establishing Gyges as the original owner.
Let us now revert for a moment to the illuminating theory of
Babelon that the striking of coins was at first a private undertaking of
merchants or miners or bankers, and that the final step in the
evolution of a gold and silver coinage was reached only when this
business of stamping and issuing the pieces was taken over by the
state. We are entirely without record of the actual transference that
the theory implies. Babelon assumes that the government created the
monopoly. This chapter suggests a modification of Babelon’s view in
one important point, namely that it was not the government that
made the monopoly, but the monopoly that made the government.
As in the case of Babelon’s own main thesis, we are forced to trust
largely to analogies. But the analogies are striking, and they strike in
two directions. We have first the commercial or financial antecedents
of the rulers of this period both in Lydia and in various Greek states:
and secondly we have the history of later Lydian rulers and aspirants
to the throne, notably Croesus and Sadyattes, Cyrus and Pactyes,
Xerxes and Pythes. The moral of their stories is that no ruler could
feel safe at Sardis until he had secured some sort of financial
supremacy. When we further consider that Gyges was famous for his
gold[790], and that his gold and his tyranny are spoken of by
Archilochus, probably in the same breath, then apart from
speculative interpretations of the story of the ring, there is at least a
clear possibility that the monopolist policy of Croesus and his
successors goes back at least to Gyges and perhaps even a generation
or so earlier, to some such ruler as Ardys or Spermos, and that it was
the monopoly in stamped pieces of electrum that brought the first
tyrant to the king’s palace and placed him on the throne[791].
 Chapter VI. Argos

τὰν ἀρετὰν καὶ τὰν σοφίαν νικῶντι χελῶναι.

Fig. 20. Early Aeginetan “tortoises.”

Our earliest account of the one tyrant of “Tyranny” of

Argos is found in Herodotus and runs as Pheidon the
follows: “and from the Peloponnesus came legitimate
sovereign “who
Leokedes the son of Pheidon the tyrant of created for the
the Argives, that Pheidon who created for Peloponnesians
the Peloponnesians their measures and their measures”
behaved quite the most outrageously of all (Herodotus),
the Greeks, who having removed the Eleian
directors of the games himself directed the games at Olympia[792].”
Pheidon belonged to the royal house of Temenos[793], and appears to
have succeeded to a hereditary throne in the ordinary way.
Nevertheless he is deliberately classed by Aristotle as a typical
tyrant[794].
 Some years ago I suggested that it was Pheidon’s “invention” of
measures rather than his outrageous behaviour or his warlike
achievements that caused him to be regarded as a different kind of
ruler from his forefathers—as a tyrant instead of a king[795].
Herodotus speaks of him simply as the and also,
man who made their measures for the according to
Peloponnesians. But Ephorus and later Ephorus, struck
in Aegina the
writers declare that Pheidon invented a
first silver coins.
system of weights as well as measures, and,
most important of all, that silver was first coined by him in
Aegina[796]. The reign of Pheidon probably covered the first third of
the seventh century. Thus in Greece Proper as in Asia Minor there is
evidence for ascribing the earliest coins to the earliest tyrant. If there
is any weight in the Argive evidence, then Pheidon the
the accounts mutually confirm one another, European
and it becomes distinctly improbable that counterpart of
Gyges.
the association of coinage and tyranny was
a mere accident.
Unfortunately the evidence as regards The evidence for
Pheidon is all very much disputed. The Pheidon is
greater part of this chapter will therefore be disputed. This
chapter is
devoted to examining its credibility and devoted to
endeavouring to show that the doubts that maintaining its
have been cast upon it are not well founded, credibility.
that on the most probable showing the reign
of Pheidon opened the epoch known as the age of the tyrants, that
Pheidon lived just about the time when the first coins were struck in
Aegina, and that the institution of the Aeginetan weight system was
the direct result of an Argive occupation of the island.
One or two points may be assumed to Date of earliest
start with as generally admitted. The Aeginetan coins.
Aeginetan “tortoises” (fig. 20) were the first coins struck in European
Greece[797], and they were first struck fairly early in the seventh
century[798]. The points that have been most disputed are the date of
Pheidon and his connexion with Aegina and the Aeginetan coinage.
It is the date that is the real centre of the Evidence for the
controversy, and it will be best to deal with date of Pheidon:
it first and to begin by briefly recalling the evidence and arguments.
 (i) The later genealogies[799] which make (1) from
Pheidon seventh from Temenos and genealogies:
eleventh from Heracles and thus put him early in the ninth century
have been shown by Busolt[800] to be due to fourth century
tamperings with the pedigree of the royal house of Macedon. In
ascribing the foundation of the Macedonian dynasty to a certain
Karanos, who according to Theopompus was a son of Pheidon,
according to Satyrus a son of Pheidon’s father[801], they were
influenced by the incurable Greek belief in symmetry which required
the Macedonian royal family to be as old as that of its great rivals the
Medes which latter, following Ktesias, they dated from 884 B.C.[802].
(ii) The genealogy that makes Pheidon tenth from Temenos and
thus puts him about the middle of the eighth century can be traced
back only to Ephorus[803]. In other words, as already pointed out by
Bury[804], its credibility depends in great measure on that of the
writer who is also our earliest authority for the statement that
Pheidon coined in Aegina[805].
(iii) Yet a third statement as to Pheidon’s family is that of
Herodotus[806]. According to Herodotus Leokedes son of Pheidon was
one of the suitors of Agariste at Sicyon early in the sixth century. The
statement occurs in a plainly romantic setting, and must not be
pressed too far. It may however be fairly claimed as an argument
against a date as early as 750. Even admitting the possibility that
παίς (son) in the singular may be loosely used for ἀπόγονος
(descendant)[807], yet it remains highly unlikely that Herodotus
would have mentioned Pheidon at all in connexion with Leokedes if
he regarded them as separated by over 150 years[808].
Lehmann-Haupt[809], who, in spite of all the difficulties just
summarized, dates Pheidon eighth Olympiad (748 B.C.)[810],
imagines an obscure Pheidon, father of Leokedes, and formulates as
a characteristic of Herodotus the practice of assigning the deeds of
famous men to obscure namesakes. Besides Pheidon he quotes only
Philokypros, tyrant of Soli and friend of Solon, whom he proceeds, in
direct contradiction of Herodotus[811], to differentiate from the father
of the Aristokypros who fell during the Ionic revolt. His view is
discredited by the one illustration that he quotes in its support.
Solon’s young friend need not have been born before 608 B.C., and
the son of a man born in that year might well be alive in 498 B.C.
Even if Herodotus is mistaken on this point, his mistake would only
illustrate the comparatively narrow margin of error to which his
anachronisms on matters of this kind are limited.
There is yet another group of statements bearing on Pheidon’s
pedigree. Pausanias says that the last king of Argos was Meltas son of
Lakedes[812]. The latter is equated by Beloch[813] with the Leokedes of
Herodotus[814]. Meltas is said by Pausanias to have been tenth in
descent from Medon, grandson of Temenos. Pheidon, as has been
noted already, is described by Strabo as tenth in descent from
Temenos himself. Thus Strabo, Pausanias, and Herodotus might be
taken as mutually confirming one another, if we accept Herodotus
literally, and make Pheidon father of Leokedes and consequently
grandfather of Meltas. As there was still a king of Argos in 484 B.C.
[815]
, and there is nothing to show that the office did not continue well
after that date, Beloch’s argument brings Pheidon well down into the
sixth century. But as it allows only twelve generations from the
Dorian invasion under Temenos to the indeterminate date after 480
B.C. when kingship was completely abolished, it only helps to
emphasize the fact that Argive royal pedigrees are not a safe guide
for determining Pheidon’s date[816].
The assertion of Pausanias[817], that (2) from his
Pheidon interfered with the eighth interference at
celebration of the Olympian games, is not to the Olympian
games.
be reconciled with a seventh century date.
But serious doubts have been thrown on Pausanias’ dating, which
may very possibly have been influenced by the Macedonian
genealogies, in which case it is no confirmation of the date arrived at
by reckoning Pheidon tenth from Temenos. The arguments for
emending eighth to twenty-eighth are weighty[818]. Pausanias’ exact
statement is that “at the eighth Olympian festival the Pisatans called
in Pheidon ... and celebrated the games along with Pheidon[819].” But
Strabo declares it to be “more probable (ἐγγυτέρω τῆς πίστεως) that
from the first Olympiad till the twenty-sixth the presidency both of
the temple and the games was held by the Eleians[820].” Julius
Africanus likewise knows nothing of any disturbance at the eighth
Olympiad, but records one at the twenty-eighth.
The difficulties in accepting the twenty-eighth Olympiad, as set
forth by Unger[821], are not very impressive. He argues that at the
twenty-eighth Olympiad the Eleians had arms[822], whereas Pheidon
made his attack “when the Eleians were without arms[823],” and that,
as the Pisatans celebrated the twenty-seventh Olympiad, Pheidon in
the twenty-eighth would have displaced not the Eleians, but the
Pisatans. But when Strabo says that the Eleians were without arms,
he or his source may mean that their Dymaean war left them
unequipped for home defence. Assume that this was so and that they
were preoccupied with their Dymaean war both in 672 B.C. and in
668 B.C., and the whole situation is explained easily. At the twenty-
seventh celebration the Pisatans unaided might secure the
presidency by a surprise attack. It would be at the next festival, when
the Eleians were forewarned, that the Pisatans would need Pheidon’s
help to displace them at Olympia. Weaker still is Unger’s argument
that a notice about Pheidon may have fallen out in Eusebius under
Olymp. VIII, as one has in the same chronicle about the emperor
Caligula.
We may indeed with Mahaffy[824] and Busolt[825] doubt whether
these early parts of the Olympian victor lists are contemporary
records. But it is easy to be unduly sceptical. Mahaffy, for instance, is
inclined to argue that the Olympian lists cannot have existed in the
fifth century because they are not then used for purposes of dating.
He assumes that Hippias who made his edition of the list in 370 B.C.
can have had little more evidence at his disposal than Pausanias, who
lived over five hundred years later. Plutarch[826], whom he quotes as
discrediting the list, merely expresses an opinion which is no more
final than that of Mahaffy himself. If our chronological data are
untrustworthy, we are thrown back on Pheidon’s achievements for
determining his date, a position long ago maintained by C.
Mueller[827]. Regarded as a fact of indeterminate date Pheidon’s
interference at Olympia is more likely to have been remembered if it
was not made so early as the close of the eighth century, when the
festival had not yet attained its subsequent reputation[828].
The evidence as to Pheidon’s date is Pheidon
therefore quite compatible with the probably
statement that makes him the first to strike contemporary
with the earliest
coins in Aegina. For if it be allowed that he
Aeginetan coins.
may have lived in the first half of the
seventh century, there would be nothing unique in his having his
mint away from his capital in an outlying but commercially
important part of his dominion. Ridgeway and Svoronos have
already compared the Romans, who struck their first coins in
Campania[829], and the Ptolemies who coined very largely in
Cyprus[830].
But had Pheidon anything to do with Did he strike
Aegina and its coinage? Against the them?
statements of Ephorus and later writers must be set the silence of
Herodotus, who makes no reference either to coins or to Aegina in
his account of Pheidon. This omission, combined with the diversity
of views as to Pheidon’s date, has led to a general distrust of
Ephorus, so much so that the majority of the most competent
authorities are either agnostics[831] or utter unbelievers[832].
Herodotus’ silence is not by itself a serious argument for rejecting
the additions of the later historians. It should be remembered that
his whole account of Pheidon in VI. 127 extends to barely four lines.
He can hardly be expected to state at all completely even the main
facts about so important a personality in so short a space. To assert,
as has been done recently[833], that Herodotus knew nothing about a
coinage issued by Pheidon is to beg the question.
To the fifth century Greek, for whom Herodotus wrote, the origin
of the system of weights on which the Aeginetan coins were struck
may have been of greater interest than the remoter question of an
invention which they doubtless all took for granted. Peloponnesian
weights and measures stood for the lack of standardization in all
matters metrical from which Herodotus and his hearers must have
suffered daily.
There is no need to pursue these criticisms in further detail. They
start with various assumptions as to Pheidon’s date. Some of them
would be found to be mutually destructive. Nearly all of them
overestimate the difficulties raised by the apparent lack of
confirmation of Ephorus in earlier writers. It is by no means certain
that his statements about Pheidon’s connexion with the Aeginetan
coinage must be a fanciful expansion of Herodotus VI. 127. There are
in fact two lines of evidence that point in quite the opposite direction.
One of them rests on notices about the Argive Heraeum
supplemented by recent finds on the site; the other is based on a new
interpretation of a passage in the fifth book of Herodotus. It will be
necessary to examine in some detail the evidence from these two
sources.
In the famous temple of Hera, the Argive A. Evidence from
Heraeum, that lies between Argos and the Argive
Mycenae, there was preserved a dedication Heraeum.
that was said to have been made by Pheidon in commemoration of
his coinage. The notice is preserved only in the mediaeval
Etymologicum Magnum. It runs: “Pheidon the Argive was the first
of all men to strike a coinage in Aegina, and on account of this
coinage he called in the spits (ὀβελίσκοι) and dedicated them to the
Hera of Argos.” There is nothing suspicious in this notice. The word
drachma means a handful, and according to Plutarch a drachma is a
handful of obols (spits or nails), which in early times were used as
money[834]. In modern times nails are said to have served as a coinage
in both Scotland and France[835]. Plenty of evidence is to be found in
antiquity for offerings of disused objects to the gods[836]. The ultimate
source for the statement of the Etymologicum Magnum may well be
the official guide at the Heraeum itself. Temple traditions are not
always above suspicion. All the same the indications of a Pheidon
tradition preserved in the Argive Heraeum are valuable, because they
show a source from which Ephorus might very well have
supplemented Herodotus far older and more valuable than his own
imagination.
 Fig. 21. Bundle
of spits found in
the Argive
Heraeum.

This however does not end the evidence of the Argive Heraeum.
Some thirty years ago the site was excavated by the American School
of Archaeology at Athens. Among the finds was a bundle of iron spits
or rods about four feet long (fig. 21) which Svoronos[837] has plausibly
associated with the ὀβελίσκοι of the Etymologicum Magnum.
The Americans ascribed the foundation of the Heraeum to the
Mycenaean period, so that the dedication of the spits could be put
anywhere in the three centuries that form the limit of controversy as
to Pheidon’s date. But more recently this dating has been shown to
be mistaken by the Germans who excavated Tiryns. Whole series of
miniature vessels which the Argive Heraeum excavators had
regarded as Mycenaean were shown by the Tiryns excavators to be
seventh century or later, and when one of them, Frickenhaus, visited
the Argive Heraeum, he found fragments of Geometric and proto-
Corinthian pottery in positions which proved the sherds to be older
than the temple foundations. From this fact he argues conclusively
that the abundant series of dedications at the Heraeum begin in the
seventh century. Pheidon’s ὀβελίσκοι cannot therefore go further
back than that[838]. The Mycenaean objects from the Argive Heraeum
site must all come from a small secular settlement that preceded the
temple. The latter becomes possibly contemporary with Pheidon
himself.
This is a fact of possible significance. It suggests that the Heraeum
may have been the religious centre of Pheidon’s imperial policy, a
sort of religious federal capital carefully placed away from the chief
cities of the Argolid much as the federal capital of Australia has been
placed away from the capitals of the various Australian states. It
looks indeed as though the analogy may have been closer still, and
that Pheidon was the builder of his federal capital. If so his date was
some time, probably early, in the seventh century.
This ends the evidence of the Heraeum B. Fresh
and brings me to the most important evidence from
section of my argument. We have just seen Herodotus.
how little need there is to mistrust Ephorus simply because he does
not exactly reproduce Herodotus. All the same the earlier writer is of
course by far the more reliable. The account of Pheidon’s coinage in
Aegina would gain enormously in credibility if any evidence for it
could be found in the pages of Herodotus. Modern writers without
exception have taken it for granted that no such evidence is to be
found. In this I believe them to have been mistaken. There is a
passage in the fifth book which, though it does not mention
Pheidon’s name, I believe to describe the conquest by him of Aegina
and the institution as a result of that conquest of the weight standard
on which the earliest Aeginetan coins were struck. If my explanation
has not been anticipated, there is no reason for surprise. The passage
contains references to pottery, ships, dress, and jewellery, and my
interpretation of it is based on archaeological evidence much of
which has only quite recently become available.
In the passage of the fifth book which we Herodotus, V. 82
[839]
are now to examine Herodotus is f. describes an
explaining the origin of the hatred that Argive
intervention in
existed between Athens and Aegina in 500 Aegina,
B.C. Aegina had once been subject to
Epidaurus[840]. Then the Aeginetans, having built triremes and made
themselves masters of the sea[841], revolted. Through their revolt they
got embroiled with the Athenians, who had at that time very close
relations with Epidaurus. At the suggestion of the Epidaurians, the
Athenians sailed against Aegina. The Aeginetans appealed to Argos,
and with the help of an Argive force that crossed undetected from
Epidaurus, utterly defeated the Athenians in a land battle on the
island. The various measures[842] taken in common by the Aeginetans
and the Argives immediately after the war suggest that Aegina, when
she had revolted from Epidaurus, became in some sort a confederate,
or possibly a subject, of Argos[843]. We may assume too that Argos
secured some sort of control over Epidaurus in the course of the war.
Otherwise it is inconceivable that an Argive force should have set out
from Epidaurus with the double purpose of aiding Epidaurus’
revolted subjects and attacking those very Athenians, whose
expedition against the island had been suggested by the Epidaurians
themselves[844]. The crushing defeat that the Athenians sustained
may have been due to the collapse of her Epidaurian allies.
One further point about Herodotus’ narrative should be noticed.
There is nothing in it to suggest that at the time when the Aeginetans
revolted from the Epidaurians, either of them was dependent on
Argos. The narrative points rather to a previous confederation or
dominion in which the chief cities were Epidaurus, Aegina, and not
Argos but Athens. Are there any indications as to when all this
occurred?
The reference is to a time considerably[845] previous to 500 B.C.
Macan thinks that the most probable date for the expedition to
Aegina is somewhere in the lifetime of Solon or Peisistratus[846]. He
points to various circumstances that generally
certainly might well have led to a collision ascribed to the
between Athens and Aegina during that first half of the
sixth century,
period[847]. All the same it is difficult to
accept a date within those limits. The Aeginetans are not likely[848] to
have been dependent on Epidaurus after it was conquered by
Periander, about 600 B.C.[849] If therefore the revolt from Epidaurus
and the Athenian invasion are incidents in one and the same
struggle, both must go into the seventh century. Macan prefers to
assume a long interval between these two events. But Herodotus
gives no hint of one. On the contrary, his narrative hangs excellently
together as a description of successive and correlated incidents in a
single struggle. Not only so, but even if the invasion be separated
from the revolt, it is difficult to believe that it occurred after 590. So
crushing a defeat for the Athenians, who themselves admitted that
only one of their number got back to Attica[850], could hardly have
taken place in the time of Solon or Peisistratus without being
associated with their names. After all, a fair amount is known about
sixth century Athens. There are no traces of any such overwhelming
disaster, or of the inevitable set back that would have followed it. The
relations of Athens to Argos during the period seem to have been
friendly rather than the reverse. Peisistratus had Argive mercenaries,
not to speak of an Argive wife[851]. Argive support of Peisistratus is of
course quite consistent with hostility to the government that
Peisistratus overthrew. It has indeed been suggested[852] that the
Aeginetan expedition took place while Peisistratus was in exile. But,
apart from the entire absence of evidence, and all the other
difficulties involved by a sixth century date, this suggestion means
that Peisistratus sought a bodyguard and wife in the most unpopular
quarter imaginable, hardly a probable proceeding on the part of a
ruler so tactful and popular as Peisistratus must have been.
A date late in the seventh century is rendered unlikely by what is
known of Procles of Epidaurus[853], the father-in-law of Periander,
who ruled Epidaurus during the last part of the seventh century[854],
apparently as a dependent of the Corinthian tyrant, by whom he was
eventually deposed. C. Mueller indeed[855] claims Aegina for Procles,
but only on the more than dubious evidence of a more than dubious
story of Plutarch’s, which tells how Procles once used an “Aeginetan
stranger” to get rid of the corpse of a man whom he had murdered
for his money[856].
On the whole the narrative seems to fall but more
best into the first half of the seventh probably to be
century. That is the time that best suits the dated early in the
seventh, as
naval situation during the war, and the shown by
effect that it is said to have had upon archaeological
costume, ornament, and pottery. As the evidence on
archaeological evidence for all these points allusions in the
is based largely on the evidence of the narrative to:
pottery, it will be best to take the pottery
first.
 In the temple of Damia and Auxesia on (i) pottery,
Aegina it became the practice (νόμος) after
the war “to introduce into the temple neither anything else Attic nor
pottery, but to drink there henceforth only out of native jars[857].”
Herodotus mentions this embargo on Attic pottery only as applied to
the one temple on Aegina[858]. But he states that it was observed by
Argives as well as by Aeginetans, which points to the possibility that
the practice prevailed in Argos as well as Aegina. Macan goes as far
as to suggest that it is an “understatement and pseudo-explanation of
a measure or custom for the protection of native ware from Attic
competition[859].” The other measures recorded in this connexion, the
changes in Attic dress and in Peloponnesian brooches, support
Macan’s suggestion. But in the matter of dating he follows earlier
writers who, using very inadequate material, came to conclusions
which can now be shown to be improbable. They date this embargo
in the middle of the sixth century. But in Aegina at any rate Attic
pottery continued to be imported throughout the second half of the
sixth century, while in Argos, where the evidence is less decisive and
abundant, there is no sign of a cessation of Attic imports about 550
B.C. On the other hand both in Argos and Aegina there does appear
to be an abrupt cessation of Attic imports early in the seventh
century. As, further, the general history of Greek pottery shows that
an Argive-Aeginetan embargo on Attic pottery would have had a
strong commercial motive early in the seventh century and none in
the middle of the sixth, there is a strong presumption that the date of
the embargo was not the middle of the sixth century but somewhere
about the beginning of the seventh. To examine the archaeological
evidence here in detail would take us too far from our main enquiry.
It will be found presented in full in an appendix[860].
The war was a great disaster for Athenian (ii) sea-power
naval power. Now the period of greatest and ships,
eclipse for Athens from this point of view was the seventh century.
Throughout it there is no indication whatever of naval activity at
Athens, except a possible war against Mitylene. Even that must be
put at the earliest close on the year 600 B.C., and is to be regarded as
announcing the beginning of the new epoch of activity in the sixth
century[861]; and against it must be set the failure in the struggle with
Megara for Salamis[862]. This had not been the naval position of
Athens earlier. During the dark ages she appears to have been a
considerable naval power. A tradition preserved by Plutarch makes
Athens succeed Crete in the command of the sea[863]: naval power is
implied in Theseus’ expedition to Crete; a poem of Bacchylides[864],
which is illustrated by a vase painting of Euphronios[865], tells how
Theseus went to the depths of the sea to fetch up the ring of Minos,
and the story has been brought by S. Reinach into connexion with
rings such as those of Polycrates and the doges of Venice, and
explained as symbolizing the winning by Theseus of the sea which
had been previously the bride of Minos[866].
The date of these events must not be pressed. The period of this
sea-power is plainly the dark age that followed the breaking up of the
Cretan and Mycenaean civilization. It is the period of the pottery
known as Geometric, and the Athenian Geometric, the Dipylon ware,
again and again shows pictures of ships. Thirty-nine examples are
quoted by Torr[867], enough, as pointed out by Helbig[868], to prove the
important rôle played by the Athenian navy in the life of Athens of
that age. The Dipylon ships, as remarked twenty years ago by
Helbig[869], show that already in the eighth century Athens was
preparing to found her power on her navy. It requires some such
catastrophic explanation as has just been offered to account for her
complete set back in the seventh.
One result in Athens of the reverse in (iii) dress.
Aegina, so Herodotus declares, was a
revolution in the dress of the Athenian women, who gave up the
Doric costume, which was made of wool and fastened with pins, and
adopted in its place the Ionic, which consisted of sewn garments
made of linen. The passage is a locus classicus among writers on
Greek dress, and it must be at once admitted that nearly all of them
accept a date late in the first half of the sixth century[870]. So late a
date seems to me to be untenable. It can be reconciled neither with
the statements of Thucydides on the subject of Athenian dress[871],
nor with the evidence of extant monuments[872]. The sumptuary laws
on women’s dress passed by Solon in 594 B.C.[873] were plainly
directed against the Ionian costume. They show that it must have
reached Athens by about 600 B.C. and offer no evidence that it had
not done so considerably earlier. Bury dates the introduction of
Ionian dress into Athens “c. 650 (?)[874].”
 Among the Aeginetans and Argives as a After the war the
result of their victory over Athens a change Argives and
was introduced in what Herodotus calls the Aeginetans make
their brooches
“measure” (μέτρον) of Aeginetan and Argive “half as big
brooches (περόναι). Herodotus states that again.”
this change affected both the dedications at
the temple of Damia and Auxesia[875], and also the general
manufacture and use. The way he tells the story explains why he goes
beyond the temple when speaking of the pins, but does not do so in
the case of the pottery. The exclusion of Attic pottery from the
Aeginetan temple, or rather the exclusive use for temple purposes of
local ware, was in Herodotus’ days a ritualistic survival. The large
brooches on the other hand had continued in general use. “Now the
women of Argos and Aegina even to my own days wore brooches of
increased size.” Very possibly Herodotus had himself noticed them.
It is the account of this change in the “measure” of the Aeginetan and
Argive brooches that confirms the connexion of Pheidon with the
origin of the Aeginetan coinage.
The new practice was in Herodotus’ own words: “to make the
brooches half as big again as the then established measure.” It is
probably significant that, both before and after the change, the
brooches have a standard “measure.” The tendency of articles of
jewellery in early periods to be of a fixed weight is a familiar one.
Numerous instances are quoted in Ridgeway’s Origin of Metallic
Currency[876]. Not only so, but these fixed weights are repeatedly
found corresponding with or anticipating the coin standards of the
places they belong to.
It may be objected that the word μέτρον does not mean weight.
This is so when it is contrasted with σταθμός[877]; but it appears to
have been used also in a more comprehensive sense[878]. The
Athenian μετρονόμοι[879] must have inspected weights as well as
measures. μέτρον is presumably applied to both, and to a fifth
century Greek there would be no question of its referring to anything
but weight when applied to jewellery[880].
The change introduced by the Argives and The Aeginetan
Aeginetans after driving the Athenians from drachma was
Aegina was to make the “measure” of their half as big again
as the Attic.
brooches half as big again as what it had
previously been. Now this is approximately the relationship in weight
of the earliest Aeginetan drachmae to the earliest drachmae struck
on the Euboeic standard. Later, in Herodotus’ own times, the relative
weights were four to three. But the earliest Aeginetan drachmae
weighed a little more than those of later issues[881]. On the other
hand, as stated by Percy Gardner[882] in discussing Solon’s
“augmentation” of the Athenian coins, the earliest Attic or rather
Euboeic drachma[883] weighed less than those of post-Solonian times.
The weight of the Aeginetan drachma as determined from the early
didrachms quoted above (p. 171, n. 6) is just over six grammes, as
compared with the 5·85 grammes of later issues, while that of the
earliest Attic Euboeic drachma as determined from the coins of p.
171, n. 8 is just over four grammes, as compared with the 4·26
grammes of later issues[884].
Thus the original Aeginetan drachma seems to have been just half
as heavy again as the earliest Attic[885]. This ratio is accepted by
Ridgeway[886], who regards it as invented to make ten silver pieces
worth one gold when gold was fifteen times as precious as silver,
while later, when silver rose to be worth 3/40 of its weight in gold,
the silver pieces were slightly diminished in weight, in order that ten
of them might still be the equivalent of one of gold[887].
Let us now return to the one passage of Summary of the
Herodotus, in which he refers by name to evidence of
the Argive tyrant. Herodotus.
In that passage he speaks of Pheidon as “the man who made their
measures for the Peloponnesians[888].” The force of the definite
article that precedes the Greek μέτρα has not always been sufficiently
stressed. More than one recent writer begins his discussion of the
passage by translating τὰ μέτρα “a system of measures.” The
subsequent argument has naturally suffered. τὰ μέτρα can be no
other measures than those associated with the Peloponnesus in
Herodotus’ own days, namely those of the famous Aeginetan
standard, employed in particular for the coinage of the island[889].
Other scholars have regarded the statement that Pheidon struck the
first coins in Aegina as merely an amplification by later writers of
these very words. They argue that “the measures” plainly meant the
Aeginetan standard, and so suggested the famous Aeginetan coinage.
This latter view assumes of course that the amplifications of Ephorus
are not to be found in Herodotus himself. But what are the facts? The
establishment of Aeginetan measures in the Peloponnesus are
alluded to by Herodotus not only in the passage about the Argive
tyrant in Book VI but also very possibly in the passage in Book V that
describes the early Argive expedition to Aegina. In this latter passage
the measures are said to have been the result of the expedition. Both
expedition and tyrant are probably to be dated early in the seventh
century. That is also the date to which numismatists generally assign
the first drachmae struck in Aegina, struck too on a standard that,
like that of our brooches, was half again as great as that previously in
use.
It is hard to avoid the inference that when the fourth century
writers say that Pheidon coined in Aegina, they are faithfully
reporting a genuine tradition.
It has indeed been maintained that the Sceptical views
whole Herodotean account of the early on these
relations of Argos, Aegina, and Athens is chapters of
Herodotus stated
unhistorical[890]. The arguments brought and answered.
forward to support this destructive view
are: (i) that the episode is timeless and its timelessness must be due
to its unhistorical character, (ii) that it must be unhistorical because
it cannot, as alleged by Herodotus, have been the cause of the war of
487 B.C., which must have been due to the natural rivalry of the two
neighbouring states. As regards the first of these two arguments, the
preceding pages have, it is hoped, shown that the episode is not
timeless: as regards the second, it is enough to point out that it
assumes that war cannot breed war, that no war can be due to two
causes, and that an incident cannot be historical if it is alleged as
leading to results that it cannot have in fact produced. The fact that
arguments such as these were accepted for publication in a periodical
of high repute less than a generation back shows how much the
whole world of scholarship was infected by the spirit of uncritical
scepticism that has left its mark in some quarters on that of the
present age.
Others again like Wilamowitz[891] regard the narrative of
Herodotus V. 82–88 as simply a reflexion backwards of the state of
affairs existing in 487 B.C.[892], when Athens attacked Aegina, and the
Aeginetans “called to their aid the same people as before, the
Argives[893].” They argue that (i) the story is our only evidence for
hatred between Athens and Aegina much before 506 B.C., (ii) the
Argive-Aeginetan brooches as compared with the broochless
Athenian costume[894], the embargo on Attic pottery at the Aeginetan
temple, and the posture of the kneeling statues (pleading before the
Athenian invaders) may all have been referred in Herodotus’ days to
the existing hatred and recent wars between Athens and Aegina, (iii)
Herodotus puts back the Athenian disaster into the timeless period
because the miracle and the change of costume required an early
date, and the story does not fit the war of 487 B.C., since the famous
Sophanes[895], who fought in it, lived till 464. Herodotus, they say,
gives no account of a disaster to the Athenian fleet in 487 because he
had used it up for this early reflexion.
Of these points (i) is answered by the whole of this chapter, (ii) and
(iii) fall with (i), besides which (ii) contains many improbabilities,
e.g. that the pottery in an Aeginetan temple should without historic
reason have suggested to any fifth century Greek an early war with
Athens, while (iii) assumes an Athenian disaster in 487 B.C., whereas
Thucydides declares that Athens was successful in that war[896].
There is nothing suspicious in the Aeginetans having twice in two
hundred years attained some sort of thalassocracy, and having on
both occasions come as a result into collision with Athens. It is
perfectly natural for the Aeginetans on a second occasion to appeal to
allies who had previously helped them so effectively and with such
profit to themselves. Macan[897] observes that the Herodotean
account of the feud between Athens and Aegina is remarkably
uninfluenced by contemporary politics and interests. He suggests[898]
dating the subjection of Aegina to Epidaurus to the reign of Pheidon,
and the revolt of the island from Epidaurus to the time of Pheidon’s
fall. But why in that case does the account speak of a revolt from
Epidaurus, if it was really a revolt from the famous Argive tyranny?
The whole narrative finds a more appropriate setting if regarded as
one chapter in the history of Pheidon himself.
Only, why in this case is the name of Why Pheidon is
Pheidon nowhere mentioned? It is one not mentioned in
thing to omit details in a biography four them.
lines in length. It is quite another to omit so prominent a name in a
narrative that runs to seven whole chapters. But the omission,
though at first sight surprising, is capable of explanation. The
Herodotean story appears to have been derived from the temple of
Damia and Auxesia[899]. It was told Herodotus not in connexion with
any royal monument, but to explain certain offerings of pottery and
jewellery that he saw in the temple. Not a single personal name
occurs in the whole narrative, and there is no particular reason why
there should. There may actually have been motives for not
introducing them. The account of the events given to Herodotus in
the Aeginetan temple of Damia and Auxesia would naturally not
emphasize the part played by the Argive tyrant. The Athenian
version, to which also Herodotus alludes, would have still better
reason for trying to forget the name of Pheidon. If my whole
interpretation of these events is not entirely wrong, Pheidon dealt
the Athenians what was probably the most crushing blow they had
ever received down to the days when Herodotus wrote his history.
The personal name may be omitted from the same motive that made
the Athenians speak of the Aeginetan drachma as the “fat” drachma,
which they are said to have done, “refusing to call it Aeginetan out of
hatred of the Aeginetans[900].” Sparta again had taken sides against
Pheidon at Olympia[901], and would have had no interest in
perpetuating the name of the man who had almost barred their way
to the hegemony of the Peloponnese.
Ephorus’ account of Pheidon’s conquests and inventions is derived
neither from Attic nor from Aeginetan sources. As seen already[902]
the source of his statement about Pheidon coining in Aegina was
most probably the Argive Heraeum. Herodotus claims to use Argive
sources, but for him the war is primarily a matter between the
Athenians and the Aeginetans, whose subsequent hatred of one
another it is intended to explain. Thus we appear to have three rival
or even hostile traditions confirming one another, so that the variety
of sources adds in a real way to the credibility of the resultant
narrative.
The notices about the coinage are not the Pheidon and
only evidence for associating Pheidon with Aegina, further
Aegina. According to Ephorus “he evidence from
Ephorus:
completely recovered the lot of Temenos, Pheidon
which had previously been split into several recovered the lot
parts[903].” Temenos appears in the of Temenos,
genealogies as great great grandson of which included
Heracles, and founder of the Dorian Aegina.
dynasty at Argos[904]. He and his sons and his son-in-law between
them are represented as securing the greater part of the North-east
Peloponnese. Aegina fell to his son-in-law Deiophontes, who went to
the island from Epidaurus[905].
The operations described in Herodotus V. 82–88, by which the
Argives crossed from Epidaurus and drove the Athenians out of
Aegina and put an end to the Epidaurians being tributary to
Athens[906], are almost beyond doubt to be identified with the
recovery by Argos of the portion of the lot of Temenos that had been
secured by Deiophontes.
It is true that this account of the recovery Traces of this
of the lot of Temenos is first certainly met recovery in other
with in Strabo, whose authority is only the passages of
Herodotus.
fourth century Ephorus. But there are hints
that Ephorus is here to be trusted. There is the evidence of
Herodotus that from an unspecified earlier date down to about 550
B.C. the Argives had possessed the whole east coast of the
Peloponnesus and “the island of Cythera and the rest of the
islands[907].” The most likely period for Argos to have acquired this
territory is the reign of Pheidon. Pheidon according to Strabo[908]
“had deprived the Spartans of the hegemony of the Peloponnese,”
and it is the Spartans who shortly before Croesus asked for their
help, had wrested from the Argives “Cythera and the rest of the
islands.” About 668 B.C., i.e. probably in Pheidon’s reign, the Argives
had beaten the Spartans in the battle of Hysiae, which decided the
possession of the strip of coast land south of the Argolid[909].
Aegina is not mentioned in these proceedings, but C. Mueller may
be right in including it among “the rest of the islands[910].” The Hysiae
campaign is roughly contemporary with the second Messenian war,
in which Argos took part against the Spartans[911], and of which
indeed it may have been an incident. Now in that war the Samians
took part by sea against the Argives[912], and it is natural to connect
this action of theirs with their repeated attacks on Aegina in the days
of the Samian King Amphikrates, at some period indefinitely before
the reign of Polycrates. The Samians were certainly a naval power in
the first half of the seventh century. The four triremes built for them
in 704 B.C. marked for Thucydides an epoch in naval history[913].
About 668 B.C. Kolaios the Samian made his famous voyage beyond
the Straits of Gibraltar to the Spanish seaport of Tartessus, a voyage
that implies much previous naval enterprise on the Samians’ part.
The rivalry with Aegina was probably commercial. Kolaios and his
crew returned from the “silver rooted streams” of the Tartessus
river[914], having “made the greatest profits from cargoes of all Greeks
of whom we have accurate information, excepting Sostratos the
Aeginetan: for it is impossible for anyone else to rival him[915].”
Samian attacks on Aegina are thus particularly likely to have
happened about the time of the second Messenian war.
A century ago C. Mueller[916] argued that some event or other
connecting Samos with Aegina must have been closely connected
with the revolt of Aegina from Epidaurus, since the revolt was
described in the History of Samos of the Samian historian Duris
(born about 340 B.C.)[917]. From this he proceeds to advocate a date
for the revolt not very long before the war between Samos and
Aegina of 520 B.C. Arguments based on the laws of digression
observed by a writer whose works are known to us only in a few
fragments need to be used with caution. If Duris is any indication
whatever for the date of the revolt, he leaves an open choice between
the time of the war of 520 B.C. and that of the days of King
Amphikrates; and as between these two the evidence shows that the
earlier is probable while the latter is almost impossible.
As independent evidence these hints would be of scarcely any
value. As confirmation of a definite but disputable statement their
value is considerable.
The recovery of ancestral domains is a Summary of
favourite euphemism among military Pheidon’s
conquerors for their policy of annexation. activities
according to
The chronology, both relative and absolute,
Strabo (=
of Strabo’s summary of Pheidon’s career Ephorus).
has every appearance of authenticity.
Pheidon first recovers the lot of Temenos, then “invents” his
measures and coinage, and after that attempts to expand eastwards
and southwards to secure the whole inheritance of Heracles, or in
other words aims at the suzerainty of the whole Peloponnese, and to
that end celebrates the Olympian games. This last event is probably
to be dated 668 B.C. The coinage must be put indefinitely earlier in
his reign, a perfectly reasonable date on numismatic and historical
grounds, and the recovery of the lot of Temenos a few years earlier
still.
The date thus reached is confirmed by the histories of the two
other leading cities of this part of the Peloponnese, Sicyon and
Corinth.
Sicyon formed part of the lot of Temenos, Pheidon and
[918]
and was held by his son Phalkes . About other parts of the
670 B.C. the city fell under the tyranny of lot of Temenos:
(i) Sicyon.
the able and powerful family of Orthagoras,
whose policy was marked by extreme hostility to Argos[919]. Pheidon
plainly can have had no footing in Sicyon during the rule of the
Orthagorids. But the unusual stability and popularity of the tyranny
at Sicyon have often been explained, not without reason, as due to its
popular anti-Dorian policy. During the second Messenian war, which
Pausanias dates 686–668 B.C.[920], so that the rise of Orthagoras
coincides with its conclusion, the Sicyonians appear to have acted in
close co-operation with the Argives[921]. The position and policy of the
Sicyonian tyrants becomes particularly comprehensible if they had
risen to power as leaders of a racial uprising that put an end in the
city to a Dorian ascendancy that dated originally from the days of
Temenos[922] and had been revived by Pheidon[923].
Whether Corinth formed part of the lot of (ii) Corinth.
Temenos is uncertain. Probably it did.
Strabo and Ptolemy exclude the city from the Argolid[924]. But on the
other hand Homer speaks of it as being “in a corner of horse rearing
Argos[925],” and Pausanias states that “the district of Corinth is part of
Argolis[926],” and that he believes it to have been so in Homeric
times[927]. The conflicting statements of these excellent authorities
are best reconciled by supposing them to be referring to different
periods. If this is so, and if, as well might be, all the domains of
Homeric Argos passed to its first Dorian lord, then Corinth formed
part of the lot of Temenos. A Temenid Corinth is perhaps implied in
Apollodorus[928], where Temenos, the two sons of Aristodemus, and
Kresphontes “when they had conquered the Peloponnese, set up
three altars of Zeus Patroos and sacrificed on them and drew lots for
the cities. The first lot was Argos, the second Sparta, the third
Messene.”
For connexions between Pheidon and Corinth we have only a story
told by Plutarch and a Scholiast on Apollonius Rhodius[929] of which
the salient points are that (a) Pheidon tries to annex Corinth; (b) the
Bacchiads and Archias are the pro-Argive party; (c) the fall of the
Bacchiads (which led to the rise of the tyrant Cypselus) meant the
overthrow of Argive influence.
So far the story is all of a piece, and supports the view that the
simultaneous establishment of Cypselus in Corinth and Orthagoras
in Sicyon may have been part cause and part result of the fall of
Pheidon and the breaking up again of the lot of Temenos. Such a
suggestion harmonizes well with the friendship that existed between
the Corinthian and Sicyon tyrants[930].
There are however chronological difficulties in this interpretation
of the Pheidon Archias story. In the story (i) the fall of the Bacchiads
is made contemporary with the foundation of Syracuse, i.e. it must
presumably be dated about 734 B.C.[931]; (ii) Pheidon is put some
time before this event, his contemporary Habron being grandfather
of Archias’ favourite Actaeon: the Marmor Parium enters Pheidon
before Archias.
In a highly romantic narrative like that of Archias and Actaeon the
last thing to be looked for is a reliable and exact chronology.
Impossible dates may mean impossible statements; but on the other
hand they may mean merely a confusion of facts of different dates, or
again, the facts may be coherent, and the dates just simply wrong.
In the present case, except for the relative dating of Archias and
Pheidon, the historic background is perfectly coherent, if the events
are put early in the seventh century. To accept the 750 date for
Pheidon sets him right in relation to Archias, but leaves the rest of
the story in the air. There is indeed always the refuge of assuming a
double banishment of the Bacchiads. But the idea of a double
banishment, traces of which might easily be discovered by the
reduplicating school of historians, is deservedly suspect, and may
have arisen from a double dating due to double dating of Pheidon. If
there really were two banishments, the story better suits the second.
 Neither Plutarch nor the Scholiast on Apollonius gives any
absolute dates; and those of the Parian Marble, which does, are
impossible. The Marble dates Pheidon 895 B.C. and Archias 758.
Pheidon is also indeed made contemporary with an Athenian who
according to Castor held the office of king from 864 to 846 B.C.[932]
From 846 to 758 is a possible, though improbably long interval
between Pheidon and Archias, if as the story tells us, the latter had as
favourite the grandson of one of Pheidon’s contemporaries; but even
so the dating is so unsatisfactory, that the latest editor of the Parian
Marble[933] has suggested transposing Archias and Pheidon. But,
apart from other difficulties, the resultant early date for Archias is
altogether against the evidence. There is no need to put him back
into the ninth century merely because it is not unlikely that Greeks at
that period were already making their way to Sicily. The antedating
of Pheidon has already been accounted for, and he appears to have
taken back Archias with him part of the way.
The date of Archias is a problem any way. But it is not difficult to
suggest a possible chronology. Pheidon’s fall[934] was probably rapid
(a proof of his hubris). His rise was probably slow. Being a hereditary
monarch, he may well have ruled for fifty years, from about 715 to
665 B.C. It was early in his career that he began to carry out his
designs on Corinth. Archias, who had founded Syracuse in 734, gave
him support. We are told no details, but the alliances of the period of
the second Messenian war and the naval struggle in the Saronic Gulf
must have supplied abundant motives and inducements. Bacchiad
government under Argive protection continued till Pheidon fell,
which meant also the fall of the Bacchiads themselves. They
withdrew to the far west. Demaratus penetrated as far as Tarquinii.
Large numbers doubtless settled at Syracuse. The order of events just
outlined coincides entirely with the extant narratives, except in the
one matter of Syracuse, and there the divergence is very
comprehensible. The founder of Syracuse had supported Pheidon.
Pheidon’s fall had led to a great influx of pro-Argive Corinthians into
Syracuse, and threw Archias back entirely onto his Sicilian colony. If
this is what really happened, it would not be surprising if the fall of
Pheidon came to be regarded as having led to the original foundation
of Syracuse.
 The chief doubt however as to the Pheidon of
historical truth of the Argive tyrant’s Corinth: is he
interference in Corinth is caused by certain identical with the
Argive?
references to a Corinthian Pheidon,
described by Aristotle as “one of the earliest lawgivers[935].” When an
Argive Pheidon is reported as making his appearance in Corinthian
history, is it a mistake due to the confusing of two separate
personalities? If two existed, they were unquestionably confused. A
Pindar Scholiast says that “a certain Pheidon, a man of Corinth,
invented measures and weights[936].”
But there is an alternative possibility. The Corinthian Pheidon may
be only one aspect of the Argive: this is suggested by the Pindar
Scholiast later in the same ode, where he says that “the Pheidon who
first struck their measures (κόψας τὸ μέτρον) for the Corinthians
was an Argive[937].” Too much stress must not be laid on such very
confused statements[938]. At best they can only corroborate other and
better evidence. This however is not altogether lacking. When
Karanos, the kinsman of Pheidon, went to Macedonia and occupied
Edessa and the lands of the Argeadae[939], Bacchiads from Corinth
settled near by among the Lynkestai[940].
A lawgiver who was “one of the earliest” can have arisen in Corinth
only before the establishment of the tyranny in 657 B.C. On the other
hand lawgivers seem to have been mainly a seventh century
phenomenon in Greece, and the most natural time for one to have
been appointed in Corinth is when the Bacchiad nobility was losing
its ascendancy, a process which may be imagined as beginning early
in the seventh century or at the end of the eighth. Plutarch describes
Pheidon’s designs on Corinth as formed at the beginning of his
career. Everything points to the Argive tyrant having had a long
reign. There is nothing improbable in the supposition that the rival
factions in Corinth invited to act as their lawgiver a young sovereign
of unusual ability who ruled a city of great traditions but not at the
time particularly powerful[941]. I have already suggested the course
taken by events in Corinth after Pheidon had once secured a position
in the city. One passage remains to be quoted that makes it still more
probable both that the Corinthian lawgiver was the Argive tyrant,
and that events in Corinth took something like the course that I have
suggested. According to Nicholas of Damascus[942] Pheidon out of
friendship went to the help of the Corinthians during a civil war: an
attack was made by his supporters, and he was killed[943]. An intimate
connexion from the beginning of his career with the great trading
and manufacturing city of the Isthmus would go far to explain the
commercial and financial inventiveness that was the distinguishing
feature of this royal tyrant[944].
 Chapter VII. Corinth

ἡ μὲν δὴ πόλις ἡ τῶν Κορινθίων μεγάλη τε καὶ πλουσία

διὰ παντὸς ὑπῆρξεν, ἀνδρῶν τε ηὐπόρησεν ἀγαθῶν εἴς τε
τὰ πολιτικὰ καὶ εἰς τὰς τέχνας τὰς δημιουργικάς.
Strabo VIII. 382.
In the passage of Thucydides[945] in which Mercantile and
he associates the origin of tyranny with the marine
acquisition of wealth, one other developments at
Corinth quoted
development is mentioned as characteristic
by Thucydides to
of the age. “Greece began to fit out fleets illustrate the
and took more to the sea.” conditions that
If the views expressed in the last chapter led to the rise of
are not entirely mistaken, then in Greece tyrannies.
Proper the earliest phases of all these developments, in politics, in
industry, and in commerce by land and sea, are all to be associated
with Pheidon of Argos. But on the same showing Pheidon was a man
born rather before his time and not quite in the right place.
The town marked out by its situation[946] to develop the new
tendencies to the fullest was Corinth, and it is in fact from Corinth
that Thucydides draws his illustration, mentioning in this connexion
the shipbuilding of the Corinthian Ameinokles about 704 B.C. and
the naval battle between Corinth and Corcyra of about 664 B.C.[947]
He says nothing about Corinthian tyrants, but the description of the
situation at Corinth is simply a paraphrase of that of the general
situation that led to tyrannies[948]. Corinth is chosen to exemplify the
normal course of things in a seventh century Greek town, and it may
be taken as certain that Thucydides regards the tyranny at Corinth as
the outcome of the mercantile and maritime developments described
in the passage just quoted.
 Only, what was the personal relationship of the tyrants to the new
developments?

Fig. 22. Corinthian vase

found at Corinth.

Fig. 23. Corinthian terra cotta tablet depicting a

potter at his wheel.
 Fig. 24. Corinthian terra cotta tablet depicting the
interior of a kiln.

Before attempting to answer this Seventh century

question, one important addition may be Corinth was also
made to Thucydides’ picture of the state of a great industrial
centre, especially
things in the city when the tyranny arose. for pottery.
Corinth was not engaged only in commerce
and shipping. She was also a great industrial centre. In the chapter
on Argos reasons have been given for thinking that the tyrant
Pheidon flourished just at the time when pottery of the style called
for want of a better name proto-Corinthian was enjoying a great
vogue in a great part of the Greek world, that much at least of this
ware was made in Pheidon’s dominions, and that Pheidon took
political measures to crush or cripple rival centres. Towards the
middle of the seventh century proto-Corinthian ware was eclipsed by
a new style, which with good reason has been named Corinthian[949].
This new style became so popular that the invention of the potter’s
wheel was ascribed to a Corinthian[950]. Corinthian vases of this
period show one of the most decorative and distinctive styles of
pottery that has ever been invented. The style of decoration
somewhat recalls oriental carpets, and it was long ago plausibly
suggested that oriental carpets or tapestries furnished the models for
the Corinthian vase painters[951]. Two jugs in this style, one from
Corinth itself[952], the other from Corneto (Tarquinii), are shown in
figs. 22 and 34. Votive tablets of the sixth century B.C. have been
found at Corinth that depict various stages of the manufacture. Two
are here reproduced (figs. 23, 24). This very distinctive pottery made
Welcome to Our Bookstore - The Ultimate Destination for Book Lovers
Are you passionate about books and eager to explore new worlds of
knowledge? At our website, we offer a vast collection of books that
cater to every interest and age group. From classic literature to
specialized publications, self-help books, and children’s stories, we
have it all! Each book is a gateway to new adventures, helping you
expand your knowledge and nourish your soul
Experience Convenient and Enjoyable Book Shopping Our website is more
than just an online bookstore—it’s a bridge connecting readers to the
timeless values of culture and wisdom. With a sleek and user-friendly
interface and a smart search system, you can find your favorite books
quickly and easily. Enjoy special promotions, fast home delivery, and
a seamless shopping experience that saves you time and enhances your
love for reading.
Let us accompany you on the journey of exploring knowledge and
personal growth!

ebookgate.com