Download the Full Version of textbook for Fast Typing at textbookfull.

com

Phonological Typology 1st Edition Matthew K.

Gordon

https://textbookfull.com/product/phonological-typology-1st-
edition-matthew-k-gordon/

OR CLICK BUTTON

DOWNLOAD NOW

Download More textbook Instantly Today - Get Yours Now at textbookfull.com

 Recommended digital products (PDF, EPUB, MOBI) that
you can download immediately if you are interested.

Labov A Guide for the Perplexed Matthew J. Gordon

https://textbookfull.com/product/labov-a-guide-for-the-perplexed-
matthew-j-gordon/

textboxfull.com

Losing Hearts and Minds 2nd Edition Matthew K. Shannon

https://textbookfull.com/product/losing-hearts-and-minds-2nd-edition-
matthew-k-shannon/

textboxfull.com

A Theory of Phonological Features 1st Edition San Duanmu

https://textbookfull.com/product/a-theory-of-phonological-
features-1st-edition-san-duanmu/

textboxfull.com

A Phonological History of Chinese 1st Edition Zhongwei

Shen (■■■)

https://textbookfull.com/product/a-phonological-history-of-
chinese-1st-edition-zhongwei-shen-%e6%b2%88%e9%92%9f%e4%bc%9f/

textboxfull.com
Capsules Typology of Other Architecture Peter Šenk

https://textbookfull.com/product/capsules-typology-of-other-
architecture-peter-senk/

textboxfull.com

Commands : a cross-linguistic typology First Edition

Alexandra Y. Aikhenvald

https://textbookfull.com/product/commands-a-cross-linguistic-typology-
first-edition-alexandra-y-aikhenvald/

textboxfull.com

Always Mine Yours Everlasting Series Book 10 1st Edition

May Gordon Gordon May

https://textbookfull.com/product/always-mine-yours-everlasting-series-
book-10-1st-edition-may-gordon-gordon-may/

textboxfull.com

Urban Design A Typology of Procedures and Products 2nd

Edition Lang

https://textbookfull.com/product/urban-design-a-typology-of-
procedures-and-products-2nd-edition-lang/

textboxfull.com

Legally Loved: The Law Trilogy: Breaking the Law Book 5

1st Edition May Gordon [Gordon

https://textbookfull.com/product/legally-loved-the-law-trilogy-
breaking-the-law-book-5-1st-edition-may-gordon-gordon/

textboxfull.com
 OUP CORRECTED PROOF – FINAL, 24/3/2016, SPi

Phonological Typology
OUP CORRECTED PROOF – FINAL, 24/3/2016, SPi

OXFORD SURVEYS IN PHONOLOGY AND PHONETICS

GENERAL EDITOR: Bert Vaux, University of Cambridge
This series aims to provide a balanced account of leading approaches to and
debates in the most active and productive topics in phonology and phonetics.
Each volume examines current and past treatments of a specific topic and offers a
reasoned account of the theories and methods that lead to the best account for
the facts. The books provide students and practitioners of phonology, phonetics,
and related fields with a valuable source of instruction and reference, set within
the context of wider developments in the field, and where relevant in linguistics
and cognitive science more generally.

PUBLISHED
 Phonological Typology
Matthew K. Gordon

IN PREPARATION
The Syllable
Juliette Blevins
The Phonetics–Phonology Interface
Abigail C. Cohn and Marie Huffman
Intonation
Sónia Frota and Carlos Gussenhoven
OUP CORRECTED PROOF – FINAL, 24/3/2016, SPi

3
Great Clarendon Street, Oxford, OX DP,
United Kingdom
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide. Oxford is a registered trade mark of
Oxford University Press in the UK and in certain other countries
© Matthew K. Gordon 
The moral rights of the author have been asserted
First Edition published in 
Impression: 
All rights reserved. No part of this publication may be reproduced, stored in
a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by licence or under terms agreed with the appropriate reprographics
rights organization. Enquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above
You must not circulate this work in any other form
and you must impose this same condition on any acquirer
Published in the United States of America by Oxford University Press
 Madison Avenue, New York, NY , United States of America
British Library Cataloguing in Publication Data
Data available
Library of Congress Control Number: 
ISBN –––– (hbk)
–––– (pbk)
Printed in Great Britain by
Clays Ltd, St Ives plc
Links to third party websites are provided by Oxford in good faith and
for information only. Oxford disclaims any responsibility for the materials
contained in any third party website referenced in this work.
 OUP CORRECTED PROOF – FINAL, 24/3/2016, SPi

Phonological
Typology

MATTHEW K. GORDON

1
 Contents

Acknowledgments xi
List of abbreviations xiii

1 Introduction 1

i.1 Phonological typology exemplified: the case of sonority i

i.2 Frequency in phonology: phonology in typology 4

i.3 The present book 5
i.3.1 Cross-linguistic frequency 8
i.3.2 Language-internal frequency 9
i.3.3 Organization of the book 16

2 Theory and explanation in phonological typology 17

2.1 Types of explanations 17
2.i.1 Phonetic factors 17
2.i.2 Speech processing and phonological typology 18
2.i.3 Frequency in phonology 19
2.i.4 Analytic biases 21
2.2 Typology in phonology: incorporating explanation into the theory 22
2.2.1 The relationship between analytic bias and other functional
biases in typology: the case of laryngeal neutralization 22
2.2.2 Typological over- and under-prediction in phonetically
driven phonology 25
2.2.3 Typology as a reflex of diachronic change 27
2.2.4 Typology and learning biases: experimental approaches 29
2.2.5 Typological variation modeled: constraints o r rules 32
2.2.5.1 Steriade (1999) on laryngeal neutralization in
Optimality Theory 32
2.2.5.2 Factorial typology in phonology: the case of
syllable-contacts 34
2.2.5.3 Modeling frequency in a constraint-based grammar 37
2.2.5-4 Modeling phonological acquisition 39
2.3 Summary 41

3 Phoneme inventories 43
3.1 Cross-linguistic distribution of phonemes 44
3.2 Consonants 44
3.2.1 Plosives 45
3.2.2 Fricatives 46
3.2.3 Nasals 48
3.2-4 Liquids 49
3.2.5 Non-liquid approximants (glides) 49
vi CONTENTS

3.3 Vowels 49
3-4 Phonemic length 51
3.5 Explaining the typology of phoneme inventories 57
3.5.1 Perceptual and articulatory factors 58
3.5.i.1 (Adaptive) Dispersion Theory 58
3.5.i.2 Dispersion Focalization Theory 60
3.5.i.3 Articulatory complexity and perceptual saturation 62
3.5.i.4 Quanta! Theory 63
3.5.i.5 Feature enhancement 64
3.5.i.6 Feature economy 65
3.6 Frequency of sounds within languages 71
3.6.1 Explaining the frequency distributions within languages 77
3.7 Phoneme inventories: a summary 82

4 Syllables 83
4.1 Internal structure of the syllable 83
4.2 Syllable margins 84
4.2.1 Intralanguage frequency of syllable types 87
4.2.2 Relationship between onset and coda markedness 90
4.2.3 Final vs. non-final coda asymmetries 96
4.2.4 Sonority and place in syllabification 97
4.2.5 Syllable repair processes 104
4.2.6 Pseudo-syllable repair processes 106
4.3 Nucleus 108
4.3.1 Syllabic sounds 108
4-4 Representations of the syllable 111
4+1 The syllable rime as a constituent 111
4.{.l.l Prohibitions against CVVC syllables 111
4.{.1.2 Co-occurrence restrictions and the rime 114
4.{.2 Sonority sequencing violations: the syllable appendix
and perceptual salience 116
4-4·3 Syllable typology and perception 117
4.5 Correlations between syllable complexity and other properties 119
4.6 Summary 121

5 Segmental processes 123

5.1 Assimilation 123
5.i.1 Consonant-consonant assimilation 124
5.i.2 Consonant-vowel assimilation 126
5.i.3 Typological frequency of assimilation patterns 127
5.i.4 Implicational scales of assimilation: the phonetic grounding 129
5.i.5 The phonetic basis for assimilation: synchronic or diachronic 133
5. I.6 Long-distance assimilation: harmony systems 134
5.i.6.1 Vowel harmony 134
5.i.6.2 Consonant harmony 137
5.i.6.3 Directionality and morphology in harmony 138
 CONTENTS vii

5.i.6.4 The phonetic basis for harmony 139

5.i.6.5 Harmony as a local vs. long-distance phenomenon 139
5.2 Dissimilation 141
5.2.1 Local dissimilation 142
5.2.2 Long-distance dissimilation 143
5.2.2.1 Long-distance consonant dissimilation 143
5.2.2.2 Long-distance vowel dissimilation 145
5.2.3 Explaining dissimilation 146
5.3 The formal representation of assimilation and dissimilation 149
5-4 Fortition and lenition 151
5-4-1 Consonants 153
5-4-2 Vowels 155
5.5 Deletion 157
5.5.1 Deletion and compensatory lengthening 158
5.5.2 The representation of compensatory lengthening 159
5.5.3 Lenition and deletion as frequency effects 160
5.6 Epenthesis 161
5.6.1 Epenthesis as syllable repair 162
5.6.2 Other prosodic functions of epenthesis 162
5.6.3 Morphological constraints on epenthesis 164
5.6-4 Segmental constraints on epenthesis 164
5.6.5 The quality of epenthetic segments 165
5.6.6 The interaction between epenthesis and other
phonological phenomena 167
5.7 Metathesis 168
5-7-1 The phonetic source of metathesis 170
5.7.2 Metathesis as perceptual optimization 172
5.8 Summary 173

6 Stress
6.1 The descriptive typology of stress
6.2 Phonologically predictable stress
6.2.1 Weight-insensitive stress
6.2.2 Weight-sensitive stress
6.2.2.1 Syllable weight as a statistical bias
6.2.2.2 Representations of syllable weight
6.2.2.3 Phonetic underpinnings of syllable weight
6.2.2.4 Final vs. non-final weight asymmetries
6.3 Stress domains: the intonational basis for left-edge vs.
right-edge asymmetries 197
6-4 The phonetic basis for extrametricality 199
6.5 Representations of stress 201
6.5.1 Stress and metrical feet 201
6.5.2 Stress and the metrical grid 202
6.5.3 Factorial typology and metrical structure 204
6.5-4 Typological asymmetries as a reflex of foot structure 205
viii CONTENTS

6.6 Tone-sensitive stress 208

6.7 Lexkal and morphological stress 209
6.8 Summary 213

7 Tone and intonation 2 15

7. 1 Tone and the taxonomy of prosodic systems 215
7.2 The organization of tone languages 216
7.3 The relationship between tone and stress 216
7-4 Number of tones 219
7.5 Tonal complexity 223
7.5.1 Tonal complexity and language-internal frequency 224
7.5.2 Syllable weight and tonal complexity 227
7.5.3 The phonetic basis of tone restrictions 229
7·5-4 Weight-sensitive tone and language-internal frequency 232
7.6 Phonological characteristics of tone 233
7.7 Tonal processes 235
7.8 Tonogenesis and interactions between tone and other features 241
7.9 Intonation 243
7.10 Terminal contours 244
7.11 The typology of pitch accents 248
7.12 Prosodic constituency 250
7.13 Prosodic structure and syntax 254
7.14 When tones collide: responses to tonal crowding 255
7.14.1 Tonal crowding in the intonation system 256
7.14.2 Tonal crowding between intonational and lexical tones 257
7.14.3 Tonal crowding in intonation systems: a summary 259
7.15 Summary 260

8 Prosodic morphology 262

8.1 Minimality effects 262
8.i.1 The typological distribution of minimality constraints 263
8.i.2 Processes in response to minimality conditions 264
8.i.3 The source of minimality restrictions: independent
constraints and evolution 266
8.i.4 Minimality as a condition on mora population 267
8.2 Reduplication 269
8.2.1 Phonological characteristics of reduplication 270
8.2.i.1 Shape of the reduplicant 270
8.2.i.2 Fixed segmentism and reduplicant-base alternations:
markedness in reduplication 277
8.2.i.3 Location of the reduplicant 281
8.2.2 Overapplication and underapplication in reduplication 282
8.2.3 Cross-linguistic distribution of reduplication patterns 2 83
8.2.3.1 Overall frequency of reduplication 285
8.2.3.2 Position of the reduplicant 286
8.2.3.3 Shape of the reduplicant 287
 CONTENTS ix

8.2.3.4 Fixed segment(s) 290

8.2.3.5 Relationship between reduplication and other
prosodic properties 29 1
8.3 Non-reduplicative templatic morphology 296
BA Prosodic truncations 298
8.5 Subtractive morphology 299
8.6 Relationship between non-reduplicative templatic morphology
and other weight-sensitive phenomena 300
8.7 Summary 301

Conclusions 303
References 305
General index 355
Index of languages 360
 OUP CORRECTED PROOF – FINAL, 24/3/2016, SPi

Acknowledgments

Many people have contributed to this book in a multitude of ways including

conceptualizing the project, pointing out useful resources, discussing the ideas
presented, reading and commenting on earlier drafts, conducting the typological
surveys that figure prominently in the book, and proofreading and formatting the
manuscript. Thanks to the following people for providing insightful and very
useful feedback on prior versions of the manuscript: Eric Campbell, Germán
Coloma, Danny Hieber, Bob Kennedy, Don Killian, Jaye Padgett, Bert Vaux, Jie
Zhang, and an anonymous reviewer consulted by Oxford University Press. Thank
you to Jeff Mielke and two other anonymous reviewers of the book proposal for
providing invaluable aid in the conceptualization of the book. Caroline Crouch
and Hannah Yates were essential contributors to the construction of the typo-
logical database on phonological patterns, and Daniel Chui, Megan Lukaniec, and
Stefan Morse generously supplied some of the phoneme data frequency appearing
in Chapter . Kaveh Varjoy provided helpful formatting assistance and Karen Tsai
offered invaluable help with proofreading and indexing.
Many others have informed the contents of this book on various levels (my
apologies to anyone I’ve forgotten), including Adam Albright, Ayla Applebaum,
Wally Chafe, Bernard Comrie, Megan Crowhurst, Stuart Davis, Paul de Lacy,
Edward Flemming, Carol Genetti, Stefan Gries, Bruce Hayes, Jeff Heinz, Harry
van der Hulst, Larry Hyman, Junko Ito, Sun-Ah Jun, Pat Keating, Michael
Kenstowicz, the late Peter Ladefoged, Anja Lunden, Ian Maddieson, Jack Martin,
Armin Mester, Marianne Mithun, Pam Munro, Erich Round, Kevin Ryan, Russ
Schuh, Donca Steriade, Richard Wright, Kie Zuraw, as well as participants in my
 seminar on phonological typology (Daniel Chui, Megan Lukaniec, Brad
McDonnell, and Stefanie Morse), my advanced undergraduate phonology class
in  (Danny Epp, Yashua Ovando, Kaveh Varjoy, Meagan Vigus, and Alicia
York), and my graduate phonology class in  (Anna Bax, Kendra Calhoun,
Caroline Crouch, Kayla Eisman, Danny Hieber, Jessi Love-Nichols, Phill Rogers,
Kevin Schäfer, Nate Sims, Morgan Sleeper, and Brendon Yoder).
I gratefully acknowledge the financial support of the UCSB Academic Senate
for generously providing a grant for compiling the typological survey.
Thanks to Bill Idsardi and Bert Vaux for conceiving this book and offering me
the opportunity to write it, and to John Davey, Lisa Eaton, Jen Moore, Julia Steer,
and Victoria Sunter of Oxford University Press for providing crucial logistical
support in its preparation.
A heartfelt debt of gratitude is especially owed my wife, Rhonda, and two sons,
Eamon and Brendan, for their unflagging support throughout the lengthy process
of seeing this book through to completion.
OUP CORRECTED PROOF – FINAL, 24/3/2016, SPi
 OUP CORRECTED PROOF – FINAL, 24/3/2016, SPi

List of abbreviations

ATR advanced tongue root

CHT Co-articulation Hypercorrection Theory
CMU Carnegie Mellon University
GLA Gradual Learning Algorithm
HG Harmonic Grammar
IPA International Phonetic Association Alphabet
OCP Obligatory Contour Principle
OT Optimality Theory
PHOIBLE Phonetics Information Base and Lexicon
TBU tone-bearing unit
ULSID UCLA Lexical and Syllabic Inventory Database
UPSID UCLA Phonological Segment Inventory Database
VOT voice-onset-time
WALS World Atlas of Language Structures
OUP CORRECTED PROOF – FINAL, 24/3/2016, SPi
 OUP CORRECTED PROOF – FINAL, 24/3/2016, SPi

1
Introduction

Phonological typology is concerned with the study of the distribution and behavior
of sounds found in human languages of the world. One thread of typological
research in phonology involves defining the range of cross-linguistic variation and
the relative frequency of phonological patterns. Another line of investigation
attempts to couch these typological observations within theories designed to
model and explain the human knowledge of and capacity to acquire phonological
systems. Both of these research programs require a cross-linguistic database from
which to draw generalizations. They often differ, however, in the ultimate purpose
to which the typological data is put to use, a difference that has consequences for
the methodology employed by the researcher. Because phonological theory dating
back to work by Trubetzkoy (), Hockett (), Jakobson (), Jakobson
et al. () has characteristically been concerned with explaining and modeling
cross-linguistic variation, typology has become largely inseparable from most
research in phonology, a close bond that is obvious even in casual inspection of
the phonology literature (Hyman a). Most chapters in recent handbooks of
phonological theory explore particular phonological phenomena, e.g. phoneme
inventories, syllable structure, harmony processes, etc., providing an overview of the
typology of the relevant phenomenon and a summary of theories designed to account
for the range of patterns. One of the current dominant paradigms in phonological
theory, Optimality Theory, is well suited to capturing typological variation since it
employs a set of competing constraints on phonological well-formedness that can be
prioritized differently in different languages (see Chapter  for discussion).

1.1 Phonological typology exemplified: the case of sonority

To illustrate the role of typology in phonological theory, let us consider the
property of sonority, which, though difficult to pinpoint phonetically (see
Parker , , ), corresponds roughly to a measure of acoustic loudness.
Phonologically, sonority manifests itself through a number of phonological phe-
nomena that are sensitive to a prominence scale like the one in Figure .
(Clements , Parker , ).
One example of the sonority scale at work comes from the formation of
diphthongs in the Austronesian language Tahitian (Bickmore ). In Tahitian,
a sequence of vowels constitutes a diphthong if the first vowel is higher in sonority
than the second vowel where sonority is determined by the following scale /a/ >
/e,o/ > /i,u/ (a). If the second vowel is higher in sonority, the two vowels are

Phonological Typology. First edition. Matthew K. Gordon.

© Matthew K. Gordon . First Published  by Oxford University Press
OUP CORRECTED PROOF – FINAL, 24/3/2016, SPi

 

Low V Mid V High V Central V Liquids Nasals Fricatives Stops

Higher sonority Lower sonority

F .. Sonority hierarchy

parsed into separate syllables (b). If the two vowels are equivalent in sonority,
they are generally also parsed into separate syllables (c), though Bickmore
reports diphthongal pronunciations as an optional variant in such cases (p. ).1

() Dipthongs in Tahitian (Bickmore : –)

(a) ho.ˈroi ‘wash’
pa.ˈrau ‘speak’
ma.ˈhae ‘torn’
ʔa.ˈʔai ‘story’
piri.ˈpou ‘pants’
ˈʔae.to ‘eagle’
fa.ˈrao.a ‘bread’
(b) ti.ˈa.re ‘flower’
mo.ˈa.na ‘ocean’
te.ˈa.ta ‘theater’
ʔi.ˈo.re ‘rat’
hu.ˈe.ro ‘egg’
fe.pu.ˈa.re ‘February’
(c) no.ˈe.ma (ˈnoe.ma) ‘November’
ʔaˈpi.u (ʔaˈpiu) ‘sheet of purau leaves’

As the examples in () show, the distinction between vowels forming a diphthong
and vowels belonging to separate syllables is relevant for stress, which falls on the
final syllable if it contains a long vowel or diphthong, otherwise on the penult. The
entire diphthongal sequence in the forms in (a) thus carries stress, whereas stress
is localized to the second vowel in the vocalic sequences in (b).
The stress system of Armenian provides evidence for another section of the
sonority hierarchy: the lower sonority status of central vowels relative to all
peripheral vowels whether low, mid, or high. Stress in most varieties of Armenian
(Vaux , Gordon et al. ) falls on the final syllable (a) unless this syllable
contains schwa in which case stress shifts to the penult (b).2

() Armenian stress

(a) hɑˈsɑk ‘age’
səɾˈpʰɛl ‘to clean’
hiˈsun ‘fifty’
həˈʁi ‘pregnant’

1
The sequence /eu/ is an exception to these generalizations in that it is parsed as two syllables
despite having a falling sonority profile: ˈpe.u ‘custom’, pe.re.ˈu.e ‘coat’ (p. ).
2
Thanks to Bert Vaux for the forms in (b).
 OUP CORRECTED PROOF – FINAL, 24/3/2016, SPi

   

(b) ˈlikʰə ‘full’

ˈtɑsə ‘ten’
ˈsɑɾə ‘cold’
ˈinə ‘nine’

As the examples show, a low, mid, or high vowel all attract stress in final position
whereas a schwa does not.
Sonority is also relevant for consonants, as syllabification in the Afro-Asiatic
language Tashlhiyt Berber illustrates (Dell and Elmedlaoui , ). Syllabifi-
cation proceeds from left to right within a word selecting the leftmost sound of
highest sonority as a syllable nucleus and the immediately preceding sound to
form a core syllable consisting of an onset and nucleus. The parse continues
moving from higher to lower sonority sounds with the proviso that all non-initial
syllables must have a syllable onset. Leftover sounds that are neither syllabified as
syllable onsets or nuclei are adjoined as syllable codas. The syllabification of two
Tashlhiyt words, ħa.wl.tn ‘make them (masc.) plentiful’ and tf.tkt ‘you suffered a
sprain’, is illustrated in ().

() Syllabification of two Tashlhiyt words (Dell and Elmedlaoui 1985: 110, 113)
Core syllable 1 Core syllable 2 Core syllable 3
σ σ σ σ σ σ

/ħaUltn/ ħa Ultn ħa Ul tn ħa Ul tn ħa.wl.tn

Core syllable 1 Core syllable 2 Coda Adjunction
σ σ σ σσ

/tftkt/ tftkt tf tkt tf tkt tf.tkt

Looking first at ħa.wl.tn, its underlying form is /ħaUltn/ (Dell and Elmedlaoui
: ), where U stands for a not-yet-syllabified high front vocoid. During the
initial parse the sequence /ħa/ is grouped into a syllable (core syllable ) with the
highest sonority sound /a/ constituting the nucleus and the pharyngeal fricative
the onset. The scan continues to the right of /a/ grouping together /l/, the next
highest sonority segment that has an available onset preceding it, with the
immediately preceding /U/ (core syllable ). Finally, the sequence /tn/ is parsed
as a syllable (core syllable ) with /n/, the next most sonorous sound in the
sonority hierarchy, serving as the nucleus. The resulting form is ħa.wl.tn, where
[w] is the phonetic realization of /U/ in onset position.
The word tftkt (Dell and Elmedlaoui : ) illustrates typologically rarer
types of syllable nuclei. In this word, /f/ is the highest sonority segment and
accordingly is parsed as a nucleus with the preceding /t/ serving as its onset (core
syllable ). The next highest sonority segment to the right of /f/ that has an
available preceding onset is /k/; they together thus form a syllable (core syllable ).
Finally, the only available option for the word-final /t/ is to be adjoined as a coda of
the second syllable. The final parse is thus tf.tkt.
OUP CORRECTED PROOF – FINAL, 24/3/2016, SPi

 

As we have seen in our discussion of Tahitian, Armenian, and Tashlhiyt, not all
languages are sensitive to all distinctions projected along the sonority scale in
Figure .. A crucial prediction of the sonority hierarchy, however, is that no
language will display sonority reversals. For example, no stress system should
preferentially stress central vowels over high vowels, or high vowels over mid
vowels, or mid vowels over low vowels. Stated another way, stress on central
vowels in a given context implies stress on high vowels in the same context, which
in turn implies stress on mid vowels, which in turn implies stress on low vowels.
Similarly, the syllabification of a stop as a syllable peak implies that a nasal in the
same context also syllabifies as a nucleus (see Chapter  on syllables for more
discussion of sonority). All phonologists are interested in establishing implica-
tional relations of the type governing stress and syllabification. Since discovering
these implications crucially relies on a broad cross-linguistic database, one can say
that the vast majority of phonologists are also typologists.

1.2 Frequency in phonology: phonology in typology

There is another type of research that is an integral part of linguistic typology but
that has played a less prominent role in phonology: the investigation of frequency
distributions across languages. (Frequency can also be examined within lan-
guages, a point to which we return in section ...) One might thus ask whether
languages like Armenian that are sensitive to vowel quality in their stress systems
are common or not. (It will be shown in Chapter  that stress systems based on
vowel quality are moderately common though less common than other types of
stress systems.) Similarly, one might wonder whether languages like Tashlhiyt
Berber that permit fricatives and stops as syllable nuclei are widely attested in the
world. (Results presented in Chapter  indicate that they are quite rare.) A survey
designed to investigate cross-linguistic frequency must control for factors such as
genetic affiliation and geographic distribution in order to minimize confounds due
to language contact or inheritance of a feature from a proto-language (see Bakker
 for discussion of language sampling). For example, a survey designed to
establish whether syllabic obstruents are cross-linguistically common or not should
be based on a broad cross-section of languages that is not biased toward the Afro-
Asiatic family or the Berber sub-family of Afro-Asiatic to which Tashlhiyt belongs.
Nor should the survey be skewed toward languages spoken in North Africa.
The investigation of cross-linguistic frequency has received less attention in
phonology than in morphology or syntax (with some exceptions discussed in
section .). Because the investigation of frequency distributions plays such a
prominent role in the field of linguistics defined as typology, it is not surprising
that phonology is less visible in publications devoted to the study of typology.
As Hyman (a) observes, perusing recent issues of linguistic typology
journals and recent introductory textbooks on linguistic typology reveals only a
small portion of content devoted to phonological topics. Croft’s () introduc-
tion to typology does not have a single chapter that focuses on phonology. Whaley
 OUP CORRECTED PROOF – FINAL, 24/3/2016, SPi

   

() similarly does not allocate any chapters to phonology. Velupillai ()
devotes one chapter to phonology as opposed to seven that arguably fall under the
rubric of morphology and syntax. Song () contains a single chapter on
phonological typology by Ian Maddieson (Maddieson ). The online version
of The World Atlas of Language Structures (Dryer and Haspelmath ) contains
only  chapters devoted to phonological topics (chapters –) versus 
chapters (chapters –) focusing on morphosyntactic features. Moravcsik
() is more balanced in its coverage of phonology, allocating a single chapter
each to phonology, morphology, and syntax.
The impoverished position of phonology in typology extends to research
articles published in linguistic typology journals. In the five-year period from
 through , there were only six articles dealing with phonology in the 
issues of STUF: Language Typology and Universals. In the same five-year time
frame (abstracting away from an outlier  issue focusing on the relationship
between phoneme inventory complexity and the origin and migration of the
human species), there are only four research articles of  total issues of Linguistic
Typology, the flagship journal of the Association of Linguistic Typology, that are
devoted to phonology. Interestingly, this same journal published in  an article
by Larry Hyman “Where’s Phonology in Typology?” that examines the basis for
the paradoxical prominence of typological research in phonological theory along-
side its conspicuous rarity in venues devoted to typology (see Hyman a for
discussion). As Hyman’s paper suggests, surveying the fields of phonology and
typology gives the impression that most phonologists are typologists but most
typologists are not phonologists.

1.3 The present book

As primarily a typology work, the principal goal of this book is to provide a cross-
linguistic description of phonological properties, exploring both the range of
variation in these properties as well as their relative frequency. On the other
hand, as a phonology book, discussion of the typological patterns is accompanied
by an overview of the key assumptions, research questions, and relative merits and
weaknesses of various approaches to explaining these patterns in the theoretical
literature. This book thus represents an attempt to provide a synthesis of the fields
of typology and modern phonological theory.
In linking the theory with the typological observations serving as the target
of coverage by the theory, a practical distinction will be drawn between the
orthogonal issues of phonological representations (e.g. phonological features
and their geometry, models of the syllable, metrical structure, etc.) and the
paradigms employing those representations whether in a substantive or a more
tangential capacity. Chapter  is primarily devoted to overarching issues in
phonological theory that transcend the particular representations assumed by a
theory or the individual phenomena discussed in various chapters. These issues
include the architecture of the phonology as a rule-based or a constraint-based
system, the role of phonetic and other functional biases in phonology, the
OUP CORRECTED PROOF – FINAL, 24/3/2016, SPi

 

relationship between synchrony and diachrony, and the formal modeling of

probabilistic as opposed to categorical distributions. Representations, on the
other hand, will be introduced in the relevant sections devoted to the phenomena
that those representations have played a prominent role in treating, e.g. autoseg-
mental phonology in the discussion of assimilation and dissimilation in the
chapter on segmental processes (Chapter ), moraic theory in the course of
discussing compensatory lengthening processes in the segmental phonology
chapter and again in the chapter on stress (Chapter ), metrical grids and foot
structure also in the chapter on stress. Space constraints preclude a full consider-
ation of the relative merits of different types of representations proposed in the
literature or of the broader architectural or philosophical issues that are topical in
phonological theory.
Nevertheless, despite these practical constraints on the theoretical coverage
afforded by this book, it is important for a book on typology not to ignore the
theory since it has historically played a crucial role in making predictions that
guide the hypothesis space in typological inquiry, especially those relating to the
exploration of correlations between phenomena (see van der Hulst to appear for
discussion of the role of research on correlations in informing phonological
theory). This book contains data on a number of links between patterns and
phenomena that were sparked by predictions made by particular theories. To
name just a couple, the survey of the relationship between onset and coda
complexity in Chapter  was conducted in response to the hypothesized link
between onsets and codas advanced in the Split Margin Theory (Baertsch ,
Baertsch and Davis , , Davis and Baertsch ). Furthermore, the entire
conceptualization of Chapter  is grounded in the unified treatment of superfi-
cially diverse phenomena within the theory of prosodic morphology developed in
work by McCarthy and Prince (/).
Because phonological theory is inherently typological, a point made earlier in
this chapter and discussed at length in Hyman (a), there is overlap between
the content of this book and the content of other introductions to phonology.
However, the emphasis on quantitative cross-linguistic distributions likely differ-
entiates this book from others providing an overview of phonology less directly
focused on typology. At the same time, it is hoped that the scope of phonological
properties covered in this book distinguishes it from other introductions to
typology, which, as already discussed, characteristically devote only a small
portion of their content to phonology.
The book examines a wide range of phonological phenomena, including the
structure of phoneme inventories, positional restrictions on phonemes, phono-
logical processes, syllable structure, stress, tone, intonation, and prosodic morph-
ology. For some of these properties, there is already a well-developed typological
literature consisting of broad quantitative investigation of cross-linguistic distri-
butional properties. Most notably, phoneme inventories have been the subject of
intensive cross-linguistic study first as part of the Stanford Language Universals
project directed by Greenberg and Charles Ferguson between  and  and
then subsequently in Ian Maddieson’s seminal work Patterns of Sounds ()
and its expanded offshoot project the UCLA Phonological Segment Inventory
 OUP CORRECTED PROOF – FINAL, 24/3/2016, SPi

   

Database (Maddieson and Precoda ) with an online interface (<http://web.

phonetik.uni-frankfurt.de/upsid_info.html>). PHOIBLE (Moran et al. ) is a
considerably larger online database of phoneme inventories and their phono-
logical feature specifications containing over , languages. The World Phono-
tactics Database (Donohue et al. ) incorporates information on syllabification
in over , languages in addition to phoneme inventories for another ,
languages.
Stress has also been the target of several extensive cross-linguistic surveys
initiating with pioneering work by Larry Hyman () and pursued most
recently in the StressTyp databases: StressTyp (van der Hulst and Goedemans
) and StressTyp (Goedemans, Heinz, and van der Hulst ). The quan-
titative typological literature on other phenomena is sparser, consisting of isolated
studies of particular sub-patterns, e.g. Greenberg () on consonant phonotac-
tics in word-initial and word-final consonant clusters, Bell () on syllabic
consonants, Hyman (), Gordon (), and Zhang () on contour
tone restrictions, Zec () and Gordon (a) on various properties falling
under the rubric of syllable weight, Bolinger () on macro-intonational pat-
terns, etc.
Certain phenomena have been the subject of quantitative typologies that are
worth revisiting for various reasons. Phonological theory has advanced consider-
ably since the typological work conducted in the s under the auspices of the
Stanford Universals project, raising new research questions for typological inves-
tigation. A striking example of the theory spawning a new ___domain of typological
inquiry is provided by the moraic theory of syllable weight (Hyman , Hayes
a; see Chapter ), which has been claimed to unite a number of superficially
unrelated phenomena (e.g. stress, compensatory lengthening, tone, prosodic
morphology). Only with the theory of weight in place did it become possible to
formulate testable hypotheses fleshing out the relationship between all these
properties.
Other existing typologies of phonological phenomena are hampered by the
coarseness of their pattern categorization, which limits the range of generaliza-
tions that can be extracted from them. For example, the WALS sample of syllable
structure (Maddieson ; see also Maddieson ; see Chapter ) employs a
tripartite distinction of languages differing in the complexity of syllables that they
permit. According to this classification, languages with simple syllable structures
allow only open syllables and a single onset consonant (CV), those with moder-
ately complex syllable structure permit single coda consonants (CVC) and/or
onset clusters whose second member is either a liquid or glide (CLV, CWV), and
those with complex syllables permit coda clusters and/or onset clusters beyond
those consisting of two consonants the second of which is a liquid or glide. The
advantage of dividing the set of languages into only three categories is that it
allows for more robust statistical comparison of the relationship between syllable
structure complexity and other properties. Working with this categorization,
Maddieson () observes a correlation between syllable structure and the
number of consonants in the phoneme inventory of a language, whereby lan-
guages with more consonants characteristically tolerate more complex syllable
OUP CORRECTED PROOF – FINAL, 24/3/2016, SPi

 

structures. One of the disadvantages, however, of employing a coarse tripartite

division of the data is that it does not distinguish between sub-levels of complexity
within the moderately complex and complex categories. For example, it is not
sensitive to whether a language falls into the moderately complex category because
it allows single codas or because it permits onset clusters whose second member is
a liquid or glide. Similarly, the complex category encompasses a diverse set of
syllable structures, including complex onsets whose second member is not a liquid
or glide, codas consisting of two consonants, codas consisting of three consonants,
etc. The cross-linguistic distribution of each of these subtypes can profitably be
examined in order to draw an enriched set of generalizations about the typology of
syllable structure.
A similar issue of category conflation arises in the WALS chapter on redupli-
cation (Rubino ; see Chapter ), which divides languages into only two
groups: those with full reduplication, i.e. reduplication of entire words, and
those with both full and partial reduplication, the latter of which entails copying
of some substring of the word. This binary division obscures potentially interest-
ing divergences between languages in the type(s) of partial reduplication they
display. For example, a partial reduplicant could be a string of consonant–vowel–
consonant (CVC) or it might be a string of consonant–vowel (CV) or it may be a
single consonant (C). The Graz database on reduplication (Hurch ) provides
a more nuanced picture of reduplication. This volume aims to enrich the typo-
logical findings by employing a finer grained categorization of patterns for several
phenomena that might have previously been classified according to coarser
divisions.
This section’s overview of the current state of phonological typology should not
give the impression that there has been little research dealing with phonology on a
cross-linguistic basis. The theoretical literature is rife with work, especially in the
last  years, that explores the range of cross-linguistic variation for particular
phonological phenomena, along the lines of the research program dealing
with sonority that was discussed earlier. However, most of this literature is
primarily concerned with the discovery of the range of cross-linguistic variation.
Of only tangential relevance to much of this theory-oriented work is the relative
frequency of different patterns across and within languages, though interest in
frequency among theoreticians is gaining in traction and is continually being
facilitated by the introduction of new online databases (e.g. PHOIBLE, The Graz
Database on Reduplication, StressTyp, UPSID, WALS, The World Phonotactics
Database).

1.3.1 Cross-linguistic frequency

A primary goal of the present work is thus to examine the frequency distributions
for a wide range of phonological properties. Investigation of frequency potentially
offers insight into various biases and conditioning factors (articulatory, percep-
tual, and cognitive) that shape and constrain human languages both synchronic-
ally and diachronically. The study of frequency has a much wider scope than
the investigation of the limits of cross-linguistic variation since most once
 OUP CORRECTED PROOF – FINAL, 24/3/2016, SPi

   

purportedly universal generalizations of phonology have turned out to have

exceptions (at least for those phenomena that are sufficiently widely applicable
to allow for robust generalizations to even be formulated). For example, the claim
that every language has at least one nasal consonant (Ferguson ) has been
demonstrated to be false by Lakes Plains languages of Papua New Guinea, some of
which lack even allophonic nasals, e.g. Obokuitai (Jenison and Jenison ) and
Sikaritai (Martin ). The vulnerability of universal statements to refutation
indicates that the most productive line of study in typological research is dis-
covering which patterns are common and which ones are rare (and how rare or
common they are) and explaining their relative frequency.
The study of frequency employed in this book is approached from two angles:
language-internal frequency, which is discussed in section .., and typological
frequency, to which we now turn. The cross-linguistic distribution of various
phenomena is surveyed for the -language sample that contributors to the
World Atlas of Language Structures (WALS; Dryer and Haspelmath ) were
encouraged to include in their chapters. This set of languages is designed to
provide a genetically and geographically balanced set of languages for investigat-
ing linguistic features (see Comrie et al.’s introduction to WALS for discussion of
the sample). The -language WALS sample is fairly faithfully followed in the
present work with a few deviations. Following the suggestion of the WALS editors
in their discussion of the sample, one member of each of the three pairs of
languages in the -language sample (German and English, French and Spanish,
Modern Hebrew and Egyptian Arabic) that did not satisfy criteria for genetic
diversity but were nevertheless included in WALS due to their status as “major”
languages was excluded in the present survey, leaving a total of  sampled
languages. (Note that the survey will still be referred to as the WALS -language
sample.) From these three pairs, German, Spanish, and Egyptian Arabic were
included, an essentially arbitrary decision. In addition, in a few cases, languages in
the WALS sample were substituted with closely related languages for which more
complete phonological information was readily available either from published
sources or through scholars with extensive experience working on the language in
question. Kabardian was substituted for Abkhaz, Caddo for Wichita, Nuuchah-
nulth for Kw’akwala, and Seneca for Oneida. The list of languages (and their ISO
codes) sampled for this book is given in Table . along with sources consulted for
the survey and two levels of genetic classification provided in WALS. The family
reflects the highest generally accepted level of classification and the genus reflects a
lower level of classification that is intended to be roughly comparable across
genera in terms of time depth of separation (<, years) (see <http://wals.
info/languoid/genealogy> for further discussion of the genetic classification
adopted in WALS).

1.3.2 Language-internal frequency

The cross-linguistic survey of various phenomena is complemented by investiga-
tions of language-internal frequency for a subset of properties in order to deter-
mine whether features that are cross-linguistically common are also relatively
OUP CORRECTED PROOF – FINAL, 24/3/2016, SPi

 

T .. Languages included in the typology

ISO Language Genus Family Source(s)

cha Acoma Keresan Keresan Miller ()

ala Alamblak Sepik Hill Sepik Bruce ()
ame Amele Madang Trans-New Guinea Roberts ()
apu Apurinã Purus Arawakan Facundes ()
aeg Arabic (Egyptian) Semitic Afro-Asiatic Watson ()
arp Arapesh (Mountain) Kombio-Arapesh Torricelli Fortune (),
Conrad and Wogiga
(), Arapesh
grammar and digital
language archive
(<http://www.
arapesh.org/xml/
fortune/Entry>)
asm Asmat Asmat-Kamoro Trans-New Guinea Voorhoeve ()
bag Bagirmi Bongo-Bagirmi Nilo-Saharan Stevenson ()
brs Barasano Tucanoan Tucanoan Jones and Jones ()
bsq Basque Basque Basque Hualde and de
Urbina ()
shi Berber (Tashlhiyt) Berber Afro-Asiatic Dell and Elmedlaoui
(, , )
brm Burmese Burmese-Lolo Sino-Tibetan Okell (), Lay
()
bur Burushaski Burushaski Burushaski Anderson ()
cad Caddo Caddoan Caddoan Chafe ()
ckr Canela-Krahô Ge-Kaingang Macro-Ge Popjes and Popjes
()
cha Chamorro Chamorro Austronesian Topping ()
chk Chukchi Northern Chukotko- Bogoras (),
Chukotko- Kamchatkan Skorik (), Krause
Kamchatkan (), Dunn ()
cre Cree (Plains) Algonquian Algic Wolfart (, ),
Ahenakew and
Wolfart ()
dag Daga Dagan Dagan Murane ()
dni Dani (Lower Grand Dani Trans-New Guinea Bromley ()
Valley)
fij Fijian Oceanic Austronesian Milner (),
Schütz ()
fin Finnish Finnic Uralic Suomi et al. ()
 OUP CORRECTED PROOF – FINAL, 24/3/2016, SPi

   

geo Georgian Kartvelian Kartvelian Hewitt ()

ger German Germanic Indo-European Wiese (),
Kentner ()
goo Gooniyandi Bunuban Australian McGregor ()
grb Grebo Kru Niger-Congo Innes ()
grk Greek (Modern) Greek Indo-European Joseph and
Philippaki-
Warburton ()
grw Greenlandic (West) Eskimo Eskimo-Aleut Fortescue ()
gua Guaraní Tupi-Guaraní Tupian Bridgeman (),
Hamidzadeh ()
hau Hausa West Chadic Afro-Asiatic Newman ()
hin Hindi Indic Indo-European Kachru ()
hix Hixkaryana Cariban Cariban Derbyshire (,
)
hmo Hmong Njua Hmong-Mien Hmong-Mien Lyman ()
imo Imonda Border Border Seiler ()
ind Indonesian Malayo- Austronesian MacDonald ()
Sumbawan
jak Jakaltek Mayan Mayan Day ()
jpn Japanese Japanese Japanese Venditti (),
Ichikawa and
Kobayashi ()
kab Kabardian Northwest Northwest Colarusso (,
Caucasian Caucasian ), Applebaum
and Gordon (),
Gordon and
Applebaum (,
a, b)
knd Kannada Southern Dravidian Sridhar ()
Dravidian
krk Karok Karok Karok Bright (),
Macaulay ()
kay Kayardild Tangkic Australian Evans (), Round
()
kew Kewa Engan Trans-New Guinea Franklin ()
kha Khalkha Mongolian Mongolic Altaic Svantesson et al.
()
kho Khoekhoe Central Khoisan Khoisan Hagman ()
kio Kiowa Kiowa-Tanoan Kiowa-Tanoan Watkins ()
koa Koasati Muskogean Muskogean Kimball (, )

(continued )
OUP CORRECTED PROOF – FINAL, 24/3/2016, SPi

 

T .. Continued

ISO Language Genus Family Source(s)

kor Korean Korean Korean Lee (), Jun

(, a), Lee
and Ramsey ()
kse Koyraboro Senni Songhay Nilo-Saharan Prost (), Heath
()
kro Krongo Kadugli Kadugli Reh ()
kut Kutenai Kutenai Kutenai Garvin (),
Morgan ()
lkt Lakhota Siouan Siouan Boas and Deloria
(), Rood and
Taylor (),
Albright ()
lan Lango Nilotic Nilo-Saharan Noonan ()
lav Lavukaleve Lavukaleve Solomons East Terrill ()
Papuan
lez Lezgian Lezgic Nakh- Haspelmath ()
Daghestanian
luv Luvale Bantoid Niger-Congo Horton ()
mal Malagasy Barito Austronesian Rajaonarimanana
(), Martin ()
mnd Mandarin Chinese Sino-Tibetan Li and Thompson
(), Lin ()
myi Mangarrayi Mangarrayi Australian Merlan ()
map Mapudungun Araucanian Araucanian Smeets ()
mar Maricopa Yuman Hokan Gordon ()
mrt Martuthunira Pama-Nyungan Australian Dench ()
mau Maung Iwaidjan Australian Capell and Hinch
()
may Maybrat North-Central West Papuan Dol ()
Bird’s Head
mei Meithei Kuki-Chin Sino-Tibetan Chelliah ()
mxc Mixtec Mixtecan Oto-Manguean Macaulay ()
(Chalcatongo)
ngi Ngiyambaa Pama-Nyungan Australian Donaldson ()
noo Nuuchahnulth Southern Wakashan Stonham (),
Wakashan Kim ()
orh Oromo (Harar) Eastern Cushitic Afro-Asiatic Owens ()
otm Otomí (Mezquital) Otomian Oto-Manguean Sinclair and Pike
(), Hensey (),
Blight and Pike (),
Wallis ()
 OUP CORRECTED PROOF – FINAL, 24/3/2016, SPi

   

pai Paiwan Paiwanic Austronesian Ferrel (), Egli

()
prs Persian Iranian Indo-European Perry (),
Mahootian ()
prh Pirahã Mura Mura Everett and Everett
(), Everett (,
)
qim Quechua Quechuan Quechuan Cole ()
(Imbabura)
ram Rama Rama Chibchan Grinevald-Craig
()
rap Rapanui Oceanic Austronesian Du Feu ()
rus Russian Slavic Indo-European Jones and Ward
()
san Sango Ubangi Niger-Congo Samarin ()
snm Sanuma Yanomam Yanomam Borgman ()
see Seneca Northern Iroquoian Chafe (, )
Iroquoian
sla Slave Athapaskan Na-Dene Rice ()
spa Spanish Romance Indo-European Alarcos (),
Harris ()
sup Supyire Gur Niger-Congo Carlson ()
swa Swahili Bantoid Niger-Congo Ashton (),
Polomé (),
Mohammed ()
tag Tagalog Greater Central Austronesian Schachter and Otanes
Philippine ()
tha Thai Kam-Tai Tai-Kadai Iwasaki and
Ingkaphirom ()
tiw Tiwi Tiwian Australian Osborne ()
tuk Tukang Besi Celebic Austronesian Donohue ()
tur Turkish Turkic Altaic Clements and Sezer
(), Kornfilt
(), Demircan
(), Wedel ()
vie Vietnamese Viet-Muong Austro-Asiatic Thompson ()
wra Warao Warao Warao Osborn ()
war Wari' Chapacura- Chapacura- Everett and Kern
Wanham Wanham ()
wch Wichí Matacoan Matacoan Viñas Urquiza (,
), Claesson
(), Avram ()
yag Yagua Peba-Yaguan Peba-Yaguan Payne and Payne
()
(continued )
OUP CORRECTED PROOF – FINAL, 24/3/2016, SPi

 

T .. Continued

ISO Language Genus Family Source(s)

yaq Yaqui Cahita Uto-Aztecan Dedrick and Casad

()
yor Yoruba Defoid Niger-Congo Bamgbose ()
zqc Zoque (Copainalá) Mixe-Zoque Mixe-Zoque Wonderly ()
zul Zulu Bantoid Niger-Congo Doke (), Poulos
and Msimang (),
Thomas-Vilakati
()

common even in languages that tolerate them. There are several reasons to believe
that this hypothesis is worthy of study and that the quantitative investigation of
distributions is likely to be fruitful in furthering our understanding of linguistic
knowledge. First, given the largely shared physiological and cognitive capacities
across humans, it is plausible that the same factors that contribute to categorical
constraints on the occurrence of properties in certain languages might also render
them statistically dispreferred in other languages. Furthermore, evidence con-
tinues to mount that language learners even under a year old are sensitive to
distributional patterns in the ambient language and use these distributions to
construct generalizations (e.g. Coleman and Pierrehumbert , Zuraw ,
Ernestus and Baayen , Albright and Hayes , Eddington , Hayes and
Londe ; see Diessel  for an overview of the many ways in which
frequency is relevant in shaping language). Finally, there is ample evidence that
has corroborated the link between categorical phonological properties and statis-
tical biases.
To take a compelling example of this link, consider the case of onset-sensitive
stress, which is discussed further in Chapter . The crucial phonological observa-
tion is that certain languages preferentially stress syllables with an onset conson-
ant over those lacking one (Davis , Goedemans , Gordon a,
Topintzi ). For example, in the Australian language Arrernte (Strehlow
, Davis , Breen and Pensalfini , Gordon a), stress falls on the
first syllable of a trisyllabic or longer word but only if that syllable begins with
an onset consonant. If the word begins with a vowel, stress instead falls on
the second syllable. Thus, we have initial stress in words like ˈtukura ‘ulcer’ and
ˈworaˌtara (place name) but second syllable stress in words like erˈguma ‘to
seize’ and arˈtanama ‘to run’ (Davis : ). (Stress falls on the first syllable of
▭

disyllabic words regardless of whether they begin with a consonant or not, e.g.
ˈkama ‘to cut’, ˈilba ‘ear’.) In Arrernte, the attraction of stress by syllables with an
onset consonant reflects a categorical feature that is predictable across most
(if not all) of the vocabulary.
Interestingly, recent research by Ryan () has shown that the preference for
positioning stress on syllables with an onset is reflected in gradient but statistically
 OUP CORRECTED PROOF – FINAL, 24/3/2016, SPi

   

robust biases in other languages (e.g. English and Russian). Ryan finds that this
bias against stress on onsetless syllables is observed both in the statistical distri-
bution of lexical stress and in productivity experiments in which the felicity of
stress on onsetless syllables is judged by speakers.
In this book, the language samples from which the frequency data are calculated
stem from various sources. Existing sources containing frequency calculations
were consulted whenever available, supplemented with my own values for certain
languages in the interest of broadening the diversity of the data set. The consulted
sources employ different types of corpora and different methods for calculating
frequency. With respect to the latter dimension, a broad distinction can be drawn
between type frequency and token frequency counts. Type frequency refers to the
frequency of a pattern. In type frequency counts, a single item is counted only
once regardless of the number of times that it occurs in the corpus. In a token
frequency count, on the other hand, each occurrence of an item contributes to the
aggregate count for that item. The corpora from which the frequency values are
calculated are either dictionaries (or other types of word or root lists) or other
written or spoken corpora.
Corpora other than dictionaries potentially provide either type or token fre-
quency counts depending on whether duplicate entries have been eliminated or
not. Even for type frequency data, methodologies may vary. Sources may differ in
terms of their level of morphological redundancy, including or excluding mor-
phologically derived forms containing the same root. For example, if one were
determining the type frequency of [ks] clusters in English, the English words
comple[ks] and comple[ks]ity could either count as one or two instances depend-
ing on whether duplication was evaluated at the level of the root or the word.
Multiple examples of either of the two words in a corpus would not increase the
type frequency of [ks], although they would be counted toward token frequency.
Token (or type) frequency counts may also vary as a function of the genre in
which the words containing those phonemes occur.
Despite the methodological variation between data samples it is hoped that the
frequency data in this book will provide some useful confirmation of (or diver-
gences from) the categorical patterns discussed. In support of this optimistic
outlook, the frequency data considered in this work are for the most part quite
similar across languages (with some divergences of course) regardless of the
nature of the source. Indeed, several of the sources consulted present both type
and token frequency data that line up closely in their distributions both on a
casual level and (for those sources that quantitatively compare the different
frequency counts) on a statistical level (see, for example, Shin et al.  for
English and Korean, Leung et al.  for Cantonese, Duanmu  for Mandarin,
Tamaoka and Makioka  for Japanese). Admittedly, type and token frequency
data may diverge due to the numerical boost awarded to phonemes that occur in
particularly high frequency items. For example, the voiced dental fricative /ð/ in
English, which is otherwise rare in content words, occurs in a few highly frequent
function words, e.g. the, that, this, which inflates its token frequency relative to its
type frequency. It is the rd most frequent consonant in type frequency but
is ranked th in token frequency (Shin et al. ). The promotion of English /ð/
Exploring the Variety of Random
Documents with Different Content
 “But how did he fall off?” queried the old lady. “Jim was a good
rider, Doc. The saddle never turned with him.”
The doctor shook his head.
“I’m sure I don’t know, Mrs. Wheeler. I am not a detective. His leg
was broken from being hung in the stirrup, I suppose.”
“He wasn’t hung to the stirrup when Joe found him.”
“Wasn’t he? Perhaps Joe Rich knows more about it than we do,
Mrs. Wheeler.”
“Sure—but where’s Joe?”
“If I knew I’d be a thousand dollars better off than I am.”
But few, if any, of the men thought that it had been anything but an
accident. A sudden dizziness, perhaps caused by indigestion, might
have made him fall. And the horse, even if it was well broken, might
have got frightened and dragged him. But there was no question
about his being robbed.
It was the evening of the fifth day since Joe Rich had left Pinnacle
City when a long train of dusty cattle-cars drew into the town of Kelo.
Dusty, wild-eyed animals peered out through the barred sides of the
cars, bawling their displeasure.
The wind was blowing a gale, and to the north an electric storm
was coming down the valley. But there was no rain; only wind and a
depressed atmosphere which presaged the coming storm. The
engine clanked in past the depot and stopped with a jerk that
shortened every draw-bar in the long line of cars.
In the caboose of the cattle-train sat a cowboy, humped over on a
bench, holding his face in his hands. His broad shoulders twisted
painfully and he gave vent to a withering curse when the caboose
almost jerked him off the bench.
On the opposite side of the car sat a tall, lean-faced cowboy, his
sad gray eyes contemplating the sufferer, who lifted his head,
disclosing a swollen jaw. Two other cowboys were seated on the
floor of the car, resting their backs against the side-seats, while they
industriously shot craps for dimes.
“Hurt yuh pretty bad, Sleepy?” asked the tall cowboy.
The sufferer lifted his head, nodded slowly and inserted a big
forefinger inside his mouth.
 “Wursh a glew har glog daged dantist libed.”
He removed the finger, spat painfully and took his face in both
hands again.
“Sleepy” Stevens was suffering the pangs of an aching molar.
“Hashknife” Hartley, the tall, lean cowboy, nodded understandingly.
“It’s worse than I thought, Sleepy,” he said, his voice full of
sympathy. “You’ve got what they call a Eskimo abscess.”
“Huh? How do yuh know?”
“I can tell by yore talk—pure Eskimo.”
“A-a-a-aw, —-! If you had this ⸺ tooth—”
“We’re goin’ to water these animals at Pinnacle City,” offered one
of the crap-shooters. “You’ll have time to have that tooth pulled.”
“Hadn’t ought to be far now,” observed Hashknife.
He bent his long nose against the dirty window glass and peered
out. The wind whistled past, and the sand sifted through the window.
A lightning flash illuminated things and a rumble of thunder came to
their ears.
A few minutes later a brakeman, carrying a lighted lantern, swung
aboard.
“Wires down,” he said shortly.
“What’ll that do to us?” queried Hashknife.
“Not much. We’re late and we ought to lay out here and let
Number 4 pass us, goin’ north; but we can’t get any orders, and the
sidin’ is blocked with a freight that broke an axle. We’ll go on to
Pinnacle City, and the passenger will have to foller us on a slow
order.”
“Quite a storm, eh?” remarked a crapshooter.
“⸺ of a storm ahead of us,” declared the brakeman, going out
again.
Finally the engine sent out its shrill blasts, calling in the flagman,
and in a few moments the draw-bars jerked shudderingly. The cattle-
train was on its way again, picking up the conductor at the station.
Sleepy groaned and hunched down in his chair. The tooth had
been thumping for eight hours. And there was a question in Sleepy’s
mind about finding a dentist in Pinnacle City. Few of the old cow-
towns boasted a dentist, and the local doctor was usually more or
less of a failure with forceps.
The long cattle-train moved slowly. There was considerable of a
grade between Kelo and Pinnacle City, and the terrific head wind
held them back. The conductor and brakeman got into the crap
game, trying to kill time over the dreary eighteen-mile stretch.
The train rumbled and clanked along, unable to make much
headway.
“Likely blow all the hair off them cow critters,” observed one of the
cowboys.
The caboose was foggy with dust, and the oil lamps hardly made
light enough for them to see the spots on the worn dice.
Suddenly the draw-bars clanked together and the caboose began
stopping by jerks. Sleepy swore painfully, when it jerked him upright.
The engine whistled shrilly, and the train ground to a stop. The
conductor peered out, swore softly and picked up his lantern.
“Must be just about to the Tumbling River bridge,” he said.
“How far is it from town?” asked Sleepy.
“Couple of miles,” said the brakeman.
He too had picked up his lantern, and they went outside. A
moment later the brakeman sprang back onto the steps.
“Bridge on fire,” he said. “Lightnin’ must have struck it.”
He lifted the top off a seat and took out several fuses which he
tucked under his arm, picked up a red lantern and hurried out to flag
down the track. Hashknife put on his sombrero and climbed off the
caboose. It was a long way to the front end of the train, and the wind
threatened to blow him off the side of the fill at any time.
The Tumbling River bridge was about a hundred and fifty feet
across, built high above the stream. It was mostly of timber
construction and one span of it was burning merrily.
Hashknife found the conductor and engineer looking over, both
decided that it would be folly to try to run it. It had evidently been
burning for quite a while.
“That shore hangs us high and dry, don’t it?” asked Hashknife.
The conductor nodded grimly.
 “We’re here for a while,” he said. “Can’t take a chance on that
thing, and we’ve got a passenger coming in behind us. They’ll be
running slow, and won’t be hard to flag. The best thing for you boys
to do is to go to bed. That span is sure to burn out in this wind.”
The wind was so strong that they had to yell in order to converse.
“Might as well be comfortable!” yelled the engineer.
The conductor nodded and followed Hashknife back to the
caboose, where he broke the news to the rest of the boys.
“Ain’t that ⸺?” wailed Sleepy. “Two miles from a dentist, and the
road on fire!”
“Better go to bed, Sleepy,” said Hashknife. “Mebbe yuh can sleep
it off.”
But Sleepy told them in no uncertain terms that sleep was out of
the question. One of the cowboys produced a pint of liquor, and this
served to put Sleepy in better spirits. No one denied him any of it.
Hashknife was curious about the passenger train which was
following them, and went on to the rear platform.
Possibly they had been stopped for thirty minutes when Hashknife
saw the beams of the passenger engine. The road was fairly
crooked for several miles, and he could see the beams of the
headlight, as it swung around the curves, throwing streamers of light
off across the hills. It was not traveling fast. It came closer and
closer, and Hashknife wondered why it did not seem to pay any
attention to the rear flagman. Of course he was out of sight around a
curve, but the speed of the passenger had not diminished.
It swung to the straight track, the beams of the headlight
illuminating the rear of the stalled train. It was then that the whistle
shrieked and the train quickly ground to a stop about a hundred
yards short of the caboose.
A man dropped from the engine and came up to the caboose. It
was a uniformed brakeman.
“What’s that ahead—a fire?” he asked, swinging up on the steps.
“Bridge on fire,” said Hashknife. “Looks like we’re here for a
while.”
“Pshaw! Some wind, eh? Say, I wonder why nobody was flaggin’
the rear of this train?”
 “They did,” declared Hashknife. “I saw the brakeman start back
with his fuses and lantern.”
“You did? That’s funny, we never seen him.”
The conductor came out and corroborated Hashknife. In a few
minutes the conductor of the passenger came along. He was a fussy
little fat man, very important. He wheezed his profanity.
“Can’t get across, eh? ⸺! Wires down behind us. Nothing to do
but wait. How did it happen you didn’t send out a flag? We might
have rammed you.”
“Flag went out!” snapped the freight conductor.
“We didn’t see it,” said the brakeman. “I was in the cab.”
“Anyway, he went back,” declared the freight conductor. “It’s no
fault of mine if you fellows can’t see.”
“Any chance of putting the fire out?” asked the passenger
conductor.
“Not a chance. One whole span on fire and this wind is like a
blow-torch. Looks like a complete tie-up for this division. There’s a
section crew at Pinnacle City, but this will be a job for bridge
builders.”
Hashknife went back in the caboose where Sleepy was lying on a
seat, still caressing a sore jaw.
“Stuck completely,” said Hashknife. “No dentist for you tonight,
cowboy.”
The brakeman came in to light a cigaret, and Hashknife
questioned him about Pinnacle City.
“South of here is the wagon-bridge,” said the brakeman. “I ain’t
familiar with this country, so I can’t tell yuh how far it is, but it can’t be
a mile—not over that, anyway.”
He went out, and Hashknife turned to Sleepy.
“How about yuh, cowboy? It ain’t over three miles to town.
Suppose we walk over and find a dentist?”
“⸺, I’d do anythin’ to stop this ache, Hashknife!”
“All right.”
Hashknife went down the car, where he picked up their war-bags
and brought them back.
 “You ain’t pullin’ out for keeps, are yuh?” asked one of the crap-
shooting cowboys.
“Nope,” grinned Hashknife. “We’ll meet yuh in Pinnacle City. Only
a fool walks away and leaves his war-bag. Yuh never know what’s
ahead of yuh.”
He dug down in his bag and drew out a well-worn cartridge belt to
which was attached a scarred holster containing a heavy Colt
revolver. He looped the belt around his lean hips, yanked the buckle
together and proceeded to fill the cylinder with .45 cartridges.
Sleepy released his jaw long enough to buckle on his own
armament, and swung the bag over his shoulder and they went out
into the night. The train crew had left the caboose steps as the two
cowboys swung down off the fill and stumbled their way to the barb-
wire fence of the right-of-way.
“Blacker ’n the inside of a cat,” declared Sleepy, after they were
away from the lights of the train. “Look out yuh don’t fall off the river
bank.”
“It shore is kinda vague,” said Hashknife. “Jist take it easy.”
“Ain’t nobody breakin’ into a gallop,” retorted Sleepy.
They were traveling through a thicket of jack-pines, which
whipped them across the face and tangled their feet. The wind was
still blowing furiously, and there was a spit of rain in the air.
Hashknife was surging ahead, one hand flung up to protect his
face from the whipping branches, when he almost ran into some
object. It flashed into his mind that it was a range animal, perhaps a
horse. Sleepy bumped into Hashknife and stopped with a grunt.
Then came the flash of a gun, a streak of flame that licked out into
the wind not over fifteen feet from them. The wind seemed fairly to
blow the report away from them. It was little more than a sharp pop.
Hashknife stumbled over a little jack-pine and went to his knees
while Sleepy unceremoniously sat down. And then the animal was
gone. Evidently it had borne a rider. The wind prevented them from
hearing which way it went.
Hashknife crawled back and found one of Sleepy’s boots.
“Didn’t hit yuh, did it?” yelled Hashknife.
“No! What do yuh make of it?”
 “Queer thing to do, Sleepy.”
They got back to their feet.
“How’s the tooth?” asked Hashknife.
“Tooth? Oh, yeah. Say, I forgot it. Let’s go.”
They went ahead again, stumbling along, while the rain increased,
and they began to be very uncomfortable. Added to their discomfort
was the knowledge that they had lost all sense of direction.
Hashknife knew they were traveling parallel to the river until they
were shot at, and from that time on he wasn’t sure of anything.
He felt they had traveled more than a mile, but they found no
wagon-road. There were no stars to guide them, and the wind had
shifted several times.
“‘We’re lost, the captain shouted,’” declared Sleepy, as they halted
against the bank of a washout, where the wind and rain did not strike
them so heavily.
“That wind was blowin’ from the north when we started, and we
tried to foller the wind,” laughed Hashknife. “Is yore tobacco wet?”
They rolled a smoke and considered things.
“I wish we was back in that nice warm caboose,” said Sleepy.
“Gosh, that shore was a comfortable place. But this is jist my luck. It
makes five times we’ve started East with a train of cows—and never
got out of the sagebrush.”
“Aw, we’ll pick ’em up in Pinnacle City, Sleepy.”
“Yeah, that’s great. But where’s Pinnacle City?”
“Two miles from the railroad bridge.”
“Good guesser.”
“It can’t be more than nine o’clock, Sleepy. By golly, there ought to
be somebody livin’ in this place-where-the-wind-comes-from.”
“If they’re all like that jigger we ran into back there, I don’t care
about meetin’ ’em,” declared Sleepy. “Anyway, the tooth has quit
hurtin’. I think the swellin’ busted when we stopped at the bridge.
That engineer shore knows how to spike his mount’s tail to the
earth!”
“There’s only three things that are botherin’ me,” said Hashknife.
“One is: Why did that party take a shot at us? And the other two are
my boots full of water.”
 “And there’s another small matter,” said Sleepy flapping his arms
dismally. “We ain’t taken any nourishment since this mornin’,
Hashknife.”
“Yeah, there’s that small matter,” agreed Hashknife. “Oh, if yuh
ever stop to check up on things, Sleepy, the world is all wrong. But
never stop grinnin’ and look back. The only place yuh ever see
ghosts is behind yuh.”
“Well, that wasn’t no ghost that snapped his gun at us.”
“He shore wasn’t, cowboy. That jigger was plumb alive. Well, I
dunno but what we might as well keep circlin’. Eventually we’ll wear
a trail, if we keep goin’ long enough. I wish I knew which was south.”
They sloshed away from the brush and headed down a slope.
“There’s a light!” exclaimed Sleepy. “Straight ahead.”
A flurry of rain obliterated the light, but it flickered again.
“Light in a winder,” said Sleepy. “Must be a house.”
“Must be,” agreed Hashknife dryly. “Windows don’t usually occur
without a house in connection.”
They struck a corral fence, followed it around to the stable and
then headed for the house. It was the HJ ranch. But these two
cowboys were far too wise to walk right up to a strange house in the
dark, especially after having been shot at so recently; so they sidled
up to the house and took a look through the window.
It was a side window of the living-room, and in the room were
Peggy Wheeler, Laura Hatton and Honey Bee. It was evident to
Hashknife and Sleepy that the living-room roof had sprung a leak
and the three people were making an earnest endeavor to catch the
water in a wash-tub, dishpan and numerous other receptacles.
A long dry period had warped the old shingles of the ranch-house
to such an extent that they leaked like a sieve.
“Looks like a harmless place,” observed Hashknife.
“And not a ⸺ of a lot of advantage over bein’ outside,” said
Sleepy. “Anyway, they look awful human.”
They walked around to the front door, clumped up the steps and
knocked on the door. Honey Bee answered the knock by opening the
door about six inches and peering out.
 “We just wondered if yuh didn’t need a couple of good men to fix
yore roof,” said Hashknife seriously.
Honey opened the door a little and peered out at them. He had
never seen either of them before, but the lamplight illuminated their
faces enough to show their grins.
“Fix the roof?” he said slowly. “Oh, yeah. Well, I’ll bet we do need
help.”
He opened the door.
“C’mon in out of the wet.”
They shuffled the mud off their boots and came in. The two girls
stood near the dining-room doorway, each of them holding a
receptacle, looking curiously at Hashknife, who removed his dripping
hat and grinned widely at them. Hashknife’s grin was irresistible.
Honey grinned foolishly and shuffled his feet.
“My name’s Hartley,” said Hashknife. “This soakin’ wet object with
me is named Stevens. He was sufferin’ from a bad tooth, and we
went huntin’ a dentist in the rain.”
“Yuh went huntin’ a dentist?” queried Honey foolishly. “Wh-where
didja expect to find one?”
“Sounds kinda queer,” grinned Hashknife. “Yuh see, we was actin’
as a couple of chambermaids to a train of cows, but the bridge
caught on fire and we got stalled. Sleepy’s tooth shore needed help;
so we started out to find the wagon-bridge, figurin’ to find this
Pinnacle City. But we didn’t find the bridge.”
“Oh, yeah,” Honey scratched his head. “The railroad bridge
caught fire. Uh-huh. Ho-o-o-old on!”
He ran across the room, grabbed up a wash-basin and placed it
under a fresh leak. Then he came back and introduced the girls to
Hashknife and Sleepy.
“My name’s Bee,” he said. “B-e-e.”
“Last or first?” asked Hashknife.
“Last. Say, I better rustle some wood for that fireplace. Kinda take
the chill off the air. Gosh, you fellers shore are wet.”
Honey hurried away for some wood, while Hashknife moved some
of the containers to more advantageous spots. There seemed to be
no end to the leaks in the HJ ranch-house.
 “Terrible, isn’t it?” smiled Peggy.
It seemed to her that these two strange cowboys, even with their
wet garments and muddy boots, had brought a warmth and cheer to
the ranch that was sorely needed.
“Oh, not so bad,” said Hashknife, squinting at a leak. “Didja ever
stop to think how much worse it would be if them few little spots were
the only place where it didn’t leak?”
“That would be terrible,” declared Laura.
“Yeah, it would. But suppose it leaked everywhere. That would be
worse, eh?”
“Do you always look at things that way?” asked Peggy.
“Mostly,” said Hashknife seriously. “Why not, Miss Wheeler?
Sunlight is brighter than shadows; and it’s a lot easier to find, if yuh
look for it. Bright things are easier to see than dark ones.”
“You listen to him a while and he’ll prove to yuh that a leaky roof is
a godsend,” laughed Sleepy.
“Well, ain’t it?” asked Hashknife. “If this roof hadn’t leaked, you
folks would probably have been in bed—and we wouldn’t have seen
their light, Sleepy.”
“That is true,” said Laura. “Oh, it was way past bedtime at the HJ
ranch!”
Honey came in with an armful of wood, which he threw in the big
fireplace.
“I’m makin’ a bet you fellers are hungry,” he said.
“Never mind that,” grinned Hashknife. “Point us the way to
Pinnacle City, and we’ll be on our way.”
“Not in that rain,” declared Peggy quickly.
She went into the kitchen, where she called Wong Lee.
“Aw, don’t bother the cook,” begged Hashknife. “Pshaw, it ain’t
worth it.”
“It’s no bother to Wong Lee,” said Peggy. “You boys get over by
that fire and dry out a little. Wong Lee will get you a meal, and Honey
will show you where to sleep. Laura and I will go to bed. Good night,
everybody.”
“Good night, and thank yuh a thousand times.”
 Hashknife and Sleepy crossed the room and shook hands with the
two girls. Peggy smiled at Hashknife.
“Thank you for coming,” she said.
The two cowboys went back to the fire and removed some of their
wet garments, after which Hashknife went back to the porch and got
their water-proof war-bags, which contained some dry clothing. They
could hear Wong Lee shuffling about the kitchen, preparing them a
meal.
He came to the door and looked in on them. He was a little,
wizen-faced Celestial.
“Yo’ like some ham-egg?” he asked.
Hashknife grinned at him, but did not reply. A smile slowly stole
across the Chinaman’s face and he bobbed his head.
“Yessa, velly good,” he said. “No tlouble.”
“You kinda got the Injun sign on Wong Lee,” grunted Honey.
“Darned old rascal almost laughed. I tell yuh, he ain’t even smiled
since Jim Wheeler was killed.”
“Thasso?” Hashknife borrowed Sleepy’s tobacco and rolled a
cigaret. “What happened to Jim Wheeler?”
“Horse dragged him to death the other day.”
Hashknife shuddered. The thought of a man’s hanging by one foot
to a stirrup never failed to rasp his nerves. He had seen men die that
way, and once when he was but a youngster he had been thrown
from a wild horse and had hung from a stirrup. Luckily the horse had
whirled into a fence corner, where another cowboy was able to hold
the animal and extricate Hashknife.
“Tough way to die,” said Hashknife.
“Y’betcha,” nodded Honey. “Head all busted up on the rocks, and
his leg twisted. Golly, it shore was awful! He owned this HJ outfit. I
work for the Flyin’ H, but I’m down here kinda helpin’ out. Hozie,
Jim’s brother, owns the Flyin’ H.”
“Miss Wheeler is Jim’s daughter, eh?”
“Uh-huh. It’s shore been a hard time for her, Hartley,” Honey
lowered his voice. “She was engaged to marry Joe Rich, and he got
drunk on his weddin’ night. Didn’t show up. Then Peggy aims to go
East with Laura Hatton. Yuh see, Jim wasn’t awful well heeled with
money. He owes the Pinnacle bank quite a lot; so he borrows five
thousand from Ed Merrick, who owns the Circle M, and gives Ed his
note.
“Ed gives him the money, and Jim starts home with it. And that’s
the last anybody ever seen of the money. Joe Rich was aimin’ to pull
out of the country; so he comes out to tell Peggy good-by. And Joe
was the one who found Jim Wheeler. Hozie Wheeler and Lonnie
Myers comes ridin’ along just a little later, and found Joe with Jim.
“And when the sheriff finds out about the missin’ money, he tries
to make Joe wait for an investigation, and Joe pops him through the
gun-arm. That’s the last we saw of Joe. There’s a reward for him,
and the sheriff has been ridin’ the hocks off his horse, but ain’t found
nothin’. So yuh can see it’s been awful tough for Peggy.”
Hashknife had been standing on one foot like a stork, holding the
other foot out to the blazing fire, while Honey sketched his story.
Sleepy hunched down, his back to the fire, his damp hair straggling
down over his forehead.
“I wonder,” he said, “if it ain’t stopped rainin’ enough for us to go
on to town? We don’t want to miss that train, Hashknife.”
“Joe Rich was the sheriff,” said Honey, as an afterthought. “But he
resigned the mornin’ after he got drunk. They made a sheriff out of
his deputy. Jim Wheeler knocked Joe down that mornin’, but Joe
didn’t do anythin’, they say.”
“And it hadn’t ought to take long to fix that bridge,” said Sleepy.
“This rain would put the fire out.”
“What kind of a jigger was this Joe Rich?” asked Hashknife
curiously.
“Jist salt of the earth, Hartley.”
“Uh-huh,” thoughtfully. “And got so drunk he forgot to get married,
eh?”
“Yeah, that’s true,” sighed Honey. “I dunno why he did; and he
never said.”
“Didn’t have no quarrel with the girl?”
“⸺, no! Aw, it was to be a big marriage. I was to be best man.
My ⸺, I almost crippled myself for life, tryin’ to wear number six
shoes.”
 “You come eat now?” asked Wong Lee.
Honey sat down with them. Sleepy looked gloomily at Hashknife
and reminded him gently that sugar was for the coffee, and not for
the eggs.
Hashknife chuckled, but sobered quickly. The rain still pattered on
the old roof and dripped off the eaves. It was warm in the kitchen.
“Five thousand dollars is a lot of money,” mused Hashknife,
stirring his coffee with a fork. He had used the same fork to dip sugar
from the bowl and did not seem to realize that it had all leaked out.
Sleepy knew the symptoms and groaned inwardly. Years of
association with Hashknife had taught Sleepy to recognize the
sudden moods of the tall cowboy. Trouble and mystery affected
Hashknife as the scent of upland fowl affects a pointer.
Hashknife, in the days of his callow youth, had been known as
George. His father, an itinerant minister in the Milk River country and
head of a big family, had had little time or money to do more than
just let this boy grow up. As soon as he was able to sit in a saddle he
lived with the cowboys and became one of them.
Blessed with a balanced mind, possibly inherited from his father,
who surely needed a balanced mind to make both ends meet, the
boy struck out for himself, absorbing all kinds of knowledge, studying
human nature. Eventually he drifted to the ranch, which gave him his
nickname, and here he met the grinning Sleepy Stevens, whose
baptismal name was David.
From the Hashknife ranch their trail led to many places. Soldiers
of fortune they became, although Hashknife referred to themselves
as cowpunchers of disaster. From the wide lands of Alberta to the
Mexican Border they had left their mark. They did not stay long in
any place, unless fate decreed that a certain time must elapse
before their work was finished. And then they would go on, possibly
poorer in pocket. Their life had made them fatalists, had made them
very human. To salve their own consciences they declared that they
were looking for the right spot to settle down; a place to live out the
rest of their life in peaceable pursuits.
But down in their hearts they knew that this place did not exist.
They wanted to see the other side of the hill. Hashknife’s brain
rebelled against a mystery. It seemed to challenge him to combat.
Where range detectives had failed utterly because they were unable
to see beyond actual facts, Hashknife’s analytical mind had enabled
him to build up chains of evidence that had cleared up mystery after
mystery.
But solving mysteries was not a business with them. They did not
pose as detectives. It merely happened that fate threw them into
contact with these things. Sleepy’s mind did not function with any
more rapidity than that of any average man, but he was blessed with
a vast sense of humor, bulldog tenacity and a faculty for using a gun
when a gun was most needed.
Whether it was merely a pose or not, Sleepy always tried to
prevent Hashknife from getting interested in these mysteries of the
range country. He argued often and loud, but to no avail. But once
started, Sleepy worked as diligently as Hashknife. Neither of them
were wizards with their guns. No amount of persuasion would induce
them to compete with others in marksmanship, nor did they ever
practise drawing a gun.
“Leave that to the gun-men,” Hashknife had said. “We’re not gun-
men.”
Which was something that many men would take great pains to
disprove, along the back-trail of Hashknife and Sleepy.
And right now, while he ate heavily of the HJ food, Sleepy Stevens
knew he was being dragged into the whirlpool of the Tumbling River
range. He could tell by the twitch of Hashknife’s nose, by the
calculating squint of his gray eyes; and if that was not enough—
Hashknife was cutting a biscuit with a knife and fork.
“Five thousand is a lot of money for the HJ to lose,” agreed Honey.
“Take that along with the seven thousand owin’ to the Pinnacle City
bank and it jist about nails the HJ hide to the floor and leaves it there
to starve.”
“Was Jim Wheeler a sickly man?” asked Hashknife.
“Sickly? Not a bit; he was built like a bull.”
“Drink much?”
“Hardly ever took a drink.”
“Ride a bad horse?”
 “Been ridin’ the same one three years, and it never made a
bobble. Jim’s broncscratchin’ days was over, Hartley.”
“Uh-huh,” Hashknife rubbed his chin with the fork. “Was it goin’ to
take five thousand dollars for to ship that girl back East?”
“Probably not.”
“What kind of a feller is Ed Merrick?”
“Good cow-man. He’s one of the county commissioners. Owned
the Circle M about five years, and is kind of a big man in the county.
Mostly horse outfit.”
“Yuh say they made a sheriff out of the deputy?”
“Yeah; Len Kelsey.”
Honey described the trouble on the street between Kelsey and
Rich, in which Kelsey was wounded. He also told them how the
cowboys hid out to keep from being sworn in to follow the fugitive.
This interested Sleepy.
“Sounds like there was some reg’lar boys around here,” he said.
“Oh, the boys like Joe,” grinned Honey. “You’d like him.”
“I dunno. Any man that ain’t got no more sense than to get drunk
and miss a chance of a wife like that dark-haired girl ain’t very much
of a feller. Or the blonde one.”
“The blonde one is my girl,” said Honey softly.
Sleepy reached impulsively across the table and shook hands
with Honey, who looked foolish.
“I’m glad yuh told me,” said Sleepy seriously. “Prob’ly save me a
lot of heartaches. She’s a dinger.”
Hashknife shoved back from the table, thanking Wong Lee for his
hospitality.
“Velly good,” Wong Lee bobbed his head. “No tlouble. You come
some mo’.”
“Mebbe we will, Wong.”
“All lite; I cook plenty.”
The rain had increased again, and Honey advised them against
attempting to go to Pinnacle City. It was not difficult to convince
them. Sleepy’s tooth did not ache any more, and their clothes were
beginning to dry; so they followed Honey down to the dry bunk-
house and went to bed.
 It did not take the rain long to extinguish the fire at the bridge, and
after an examination the train crews decided that it was still safe.
Many of the timbers were badly charred, and but for the heavy rain
which followed the wind, the whole bridge would have been doomed.
The cattle-train, minus two of the cowhands, proceeded slowly to
Pinnacle City, where it took the siding. It would spend several hours
there, watering stock, and the man in charge expected Hashknife
and Sleepy to put in an appearance before leaving time.
The passenger train drew in at the station, possibly an hour late.
The wires being down, it was impossible for them to get orders. The
heavy rain swept the wooden platform, but the depot agent trundled
out some express packages. The express car door was partly open,
but there was no messenger.
The agent climbed into the car, and the first thing that greeted his
eye was the through safe, almost in the center of the car, its door
torn open. A single car light burned in the upper end of the car, and it
was there that the agent found the messenger, bound hand and foot.
Running back to the depot, the agent told what he had found, and
the train crew hurried to the car, while another man went to get an
officer. In the waiting room of the depot the express messenger told
what he knew of the robbery. A man had struck him over the head,
and he was a trifle hazy about what had happened.
The man had boarded the car at Kelo. The messenger said he
had received several packages from the agent at Kelo, and had
gone to place them before closing the door. The wind was blowing a
gale, and he did not hear the man come in. In fact he merely
surmised that the man got on at Kelo, because as far as he knew
there was no other man than himself on the car when they stopped
at Kelo.
At any rate, the man had forced him at the point of a revolver to
close and lock the door, and had made him sit down and wait for the
train to pull out. There was quite a long delay, and the bandit
seemed rather nervous.
In fact he grew so nervous that he knocked the messenger
unconscious with his gun, and the messenger didn’t know that the
safe had been blown open. He dimly remembered a loud noise, but
was in no shape to find out what it was. Anyway, the robber had
bound and placed him behind some trunks out of the way of the
explosion.
He was just a little sick all over, yet he gave Len Kelsey a fairly
good description of the robber—as good as usually is given. A
masked man of medium height. Might have been tall, or possibly
short. Wore black sombrero, striped shirt, overalls and boots. No
vest. The shirt might have been blue and white—or red and green.
The messenger wasn’t sure. He noted particularly that the robber
had a six-shooter in his right hand, and that he wore leather cuffs—
black leather, with silver stars in a circle around the upper edge of
the cuffs.
“Was there any money in the safe?” asked Len.
“A lot of it,” declared the messenger. “I don’t know how much. I’d
like to see a doctor about my head.”
Slim Coleman, of the Lazy B, happened to be there at the depot,
and he walked back with Len Kelsey.
“What do yuh think about it, Len?” he asked.
“I dunno,” lied Len.
Slim had noted the expression of Len’s face when the messenger
told about the leather cuffs.
When Joe Rich had left Pinnacle City he was wearing a blue and
white striped shirt, black sombrero, overalls and a pair of black
leather cuffs, on which were riveted a lot of small, silver stars. Joe
had done the decorating himself, and Slim knew that no other
cowboy in the Tumbling River country wore a cuff like Joe’s.
Len did not seem inclined to talk about it, so Slim went back to the
depot, where old Doctor Curzon was bandaging up the messenger’s
head. A drink of raw liquor had helped to make the messenger more
sociable and willing to talk.
“You got a good look at his gun, didn’t yuh?” asked Slim.
“I felt it,” smiled the messenger, wincing slightly from Doctor
Curzon’s ministrations.
“What did it look like?”
“Very large caliber—about six inches in diameter.” The man
laughed at his description. “Weighed a ton. Seriously, I can’t describe
it, but it seems to me that it had a white handle. Perhaps it was
yellow, like bone. You know what I mean—not pearl. It was a Colt, I
am sure.”
Slim sighed deeply.
“Man wear any rings on his fingers?”
“I didn’t see any.”
Slim went back uptown. Joe Rich carried a Colt .45 with a yellow
bone handle. Slim remembered when Joe had carved out those
pieces of bone, working for days, at odd times, shaping the grip to fit
his hand. Slim didn’t know of another cowpuncher in the country that
carried a bone-handled gun.
The news spread quickly around the town that the safe of the
passenger train had been blown by a lone bandit who wore silver
stars on his cuffs and carried a bone-handled gun. Joe Rich’s name
did not need to be mentioned. Len Kelsey did nothing, because there
was nothing to be done. The telegraph wires were down and there
was no use in his riding out into the storm. Even if the robber did get
out at the river bridge, the storm would wipe out any tracks he might
make, and even if there were no storm, how could he track one
man?
Len Kelsey was very wise. He stayed at home where it was warm
and dry, and went to bed. He had sufficient description to prove who
had pulled the job, and he had already worn out two perfectly good
horses trying to find this elusive young man.
 CHAPTER VI: HASHKNIFE SMELLS A RAT
Sometime during that night the trouble shooters for the telegraph
company had repaired the break, and this enabled the despatchers
to straighten out the trains. The cattle-train headed out of Pinnacle
City the following morning, minus two cowboys.
The depot agent knew about this, and told Len Kelsey that there
were two lost cowpunchers somewhere on the east side of the river.
The agent knew from what he had heard the crew of the cattle-train
say that these men had left the train, intending to walk down to the
wagon-bridge. But he also knew they had taken their war-bags with
them and had buckled on their belts and guns before leaving the
train.
“Kinda looks as though they intended missin’ the train,” said
Kelsey.
“Might be worth investigating, Sheriff. The passenger was close
behind the cattle-train for a long time out there by the bridge. And
that express messenger had been hit so hard on the head that he
wasn’t sure of anything.”
“Sure—I’ll look into it,” agreed Len. “I won’t leave any stone
unturned.”
He had read this in a book, and it sounded like the proper thing for
a sheriff to say.

Hashknife and Sleepy did not mention to Peggy that Honey Bee
had told them about her troubles. She was in good spirits that
morning, and even Wong Lee sang at his work. Laura told Honey
that Peggy had talked quite a while about the tall cowboy and his
wonderful grin—and Honey told Hashknife about it.
“Didn’t either of ’em mention me?” asked Sleepy. “No? That’s
tough. But how could I grin, with my jaw all swelled? But that’s jist
my luck!”
 Honey offered to take them to Pinnacle City in the buggy. They
were hitching up the horses when Len Kelsey and Jack Ralston rode
in.
“Now, what do them ⸺ whippoorwills want?” growled Honey.
“That’s the sheriff and deputy.”
“What had we ought to do—put up our hands?” asked Sleepy.
The two officers dismounted and spoke to Honey.
“Howdy,” growled Honey.
Hashknife could plainly see that Honey Bee did not care for these
two officers of the law.
Len Kelsey studied Hashknife and Sleepy for a moment.
“I reckon you boys are the two missin’ members of the cattle-train
outfit, eh?”
“If there’s two missin’—we’re both of ’em,” said Hashknife gravely.
“Has the train left Pinnacle City?”
“Before daylight.”
“Stranded again,” groaned Sleepy. “I’ll never see the East, that’s a
cinch.”
Hashknife hitched up his belt and leaned against the buggy.
“Yuh wasn’t exactly lookin’ for us, was yuh?” he asked.
“I don’t hardly think so,” replied Kelsey. “The safe on the express
car of the passenger train that stopped back of yuh at the bridge last
night was dynamited somewhere between Kelo and Pinnacle City.”
Hashknife and Sleepy exchanged a quick glance. That might
explain why a shot had been fired at them in the dark. They had
blundered into the bandit who was making his getaway.
“For gosh sake!” snorted Honey. “Did they get much, Len?”
“Dunno how much. One man pulled the job, Honey—a man who
wore black leather cuffs with silver stars, and a bone handled six-
shooter.”
“Leather cuffs with silver stars and bone—” Honey stopped and
came in closer to the sheriff.
“Are yuh sure of that, Len?”
“That’s the messenger’s description.”
“Well, for gosh sake!”
Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade

Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.

Let us accompany you on the journey of exploring knowledge and

personal growth!

textbookfull.com