Consonant Strength

Phonological Patterns and Phonetic Manifestations

Lisa M. Lavoie

• Publisher: Routledge
• Year: 2001
• eBook ISBN: 9781136794018
• Edition: 1
• Topic: Languages & Linguistics
• Subtopic: Linguistics
• Index: Languages & Linguistics

This book is a detailed examination of the phonetics and phonology of consonant strength, drawing data from parallel acoustic and articulatory studies of English and Spanish, as well as a cross linguistic survey of lenition and fortition.

CHAPTER ONE

Introduction

Consonant weakening, or lenition, is a very common occurrence. Although weakening is pervasive—in phonological alternations, fast speech, and language change—we do not have a clear understanding of it. General discussions of weakening seldom include precise characterization of the consonant’s environment. In particular, the role of position in word and lexical stress has not been fully analyzed. Building on an extensive cross-linguistic survey, this book tests phonetic hypotheses derived from theoretical views of lenition, and draws conclusions about the phonetics and phonology of consonant strength patterns.

1.1 FOCUS OF THE RESEARCH

This research focuses on American English and Mexican Spanish, languages which both exhibit phonological and phonetic weakening, but of different types and degrees. The consonants of English and Spanish are studied in two cross-cutting environments: initial or medial in the word and pre-stress or non-pre-stress. Through these phonetic studies, I show that various acoustic and articulatory parameters pattern either by manner of articulation or by position with respect to weakening. I show that the strength of a consonant’s realization is related to the strength of its prosodic position and that the primary acoustic correlate of weakening is shorter duration.

I predict that I will find phonetic parallels to phonological weakening in the various segments and environments that I investigate. Other researchers have found phonetic and phonological versions of the same general alternation. Cohn (1993) found both phonetic and phonological rules of nasalization and Tsuchida (1997) found phonetic and phonological vowel devoicing in Japanese. One of the differences between the phonetic and phonological versions of the same alternation is gradience: the phonological alternations are categorical while the phonetic are gradient (Keating 1996). My results provide evidence of both phonetic and phonological weakening.

1.2 THE PHENOMENON OF LENITION OR WEAKENING

In this section, I provide examples of lenition, define lenition for the purposes of the present study, and discuss the environments in which lenition takes place.

1.2.1 Examples

Lenition includes numerous kinds of consonant strength changes, particularly those in which a consonant increases in sonority. The most commonly cited examples of lenition are among those shown in (1) through (3), where the alternating segments are underlined. In (1), an American English alveolar stop becomes a flap in the middle of a foot.

(1)		‘write’
		‘writer’

In (2), an underlying voiceless uvular stop is realized as a fricative intervocalically in West Greenlandic, an Inuit language spoken in Greenland.

(2)		‘how much does it cost?’
		(Fortescue 1984:11)

In (3), an underlying voiceless stop is realized as a voiced stop intervocalically in Urubu-Kaapor, a Tupí-Guaraní language spoken in Brazil.

(3)		‘deer’
		(Kakumasu 1986:399)

Starting with work in historical linguistics (e.g., Sievers 1881, Jespersen 1904), lenition has been an object of study. Lenition can be seen particularly clearly in historical linguistics as a sequence of changes culminating in consonant loss. Although originally defined in historical linguistics, the term lenition is useful in modern generative phonology as there are many synchronic cases which parallel the diachronic.

1.2.2 Definition of Lenition

The consensus from a range of sources is that lenition represents a shift toward deletion. Bloomfield (1933) describes lenition as involving successive acoustic types. While it is easy to describe lenition as a shift, it is much more challenging to describe the steps that make up that shift. Trask (1996) defines lenition as:

any phonological process in which a segment becomes either less strongly occluded or more sonorous, such as k›x, x›h, k›g. Often the term is extended to various other processes, […], which represent ‘weakening’ in some intuitive sense.

Parallel to Trask’s decreased occlusion or increased sonority, Lass (1984) advocates treating lenition with a scale of sonority (increased output of periodic acoustic energy) and a scale of openness (decreased resistance to airflow). If lenition is a unified phenomenon, it is curious that two different kinds of scales are needed to describe it. In fact, Crowley (1987:26) states:

The concept of lenition is actually not very well defined, and linguists who use the term seem to rely more on intuition or guesswork than on detailed understanding of what lenition is.

A major goal of this book is to clarify what lenition actually represents, both phonetically and phonologically.

Determining if alternations represent increased sonority depends on the sonority hierarchy you use because not all segment types are included in most hierarchies. Looking at lenition as increased openness explains how stop bursts are lost or stop closures become incomplete. But there are other kinds of alternations that seem to involve strength, and they are much harder to understand under the heading of lenition. Vijayakrishnan (1999) separates weakening phenomena into two groups: loss of marked feature specifications and increase in sonority. Loss of marked feature specifications includes the loss of aspiration, place specification, or an entire segment. Increase in sonority includes voicing, fricativization, and sonorization.

The sonority approach works well for consonants that appear on a sonority hierarchy but not for consonants, like affricates, ejectives, and glottals, that typically do not. Crowley (1987) is one of few authors who include glottal sounds in a description of lenition. Glottals—consonants with no oral articulation—are clearly weaker than oral sounds. Affricates and ejectives do not have a clear interpretation in terms of strength unless we view them as marked segments, which are stronger than unmarked segments. We will see that affricates are ambiguous phonologically, because they play different phonetic roles, representing either weakening or strengthening depending on the environment.

Including both increases in sonority and decreases in marked structure brings a very broad-ranging group of changes under the heading of lenition, among them, voicing, deletion, fricativization, debuccalization, word-final devoicing, degemination, deaspiration, and vowel reduction. This group of changes ranges so widely that it is difficult to determine what they have in common, except that they are noted over and over by independent researchers who propose very similar weakening hierarchies, sharing many of the same steps and generalizations.

In developing my predictions and designing the word lists, I used the following working definition of lenition: any alternation which yields a consonant that is articulated with a more sonorous manner of articulation or with less marked structure. In the subsequent chapters of this book, I use the term weakening to avoid some of the diachronic notions that come with the term lenition. As a phenomenon defined within historical linguistics, lenition seems to apply more to the changes that can be seen as a step toward deletion of a segment. Of course, lenition extends to cover synchronic examples that parallel the diachrony.

Four major types of explanation have been put forth in the literature for lenition. These include lenition as a step on the way to deletion, lenition as an increase in sonority, lenition as a decrease in effort, and lenition as a decrease in the duration and magnitude of articulatory gestures.

While the sorts of alternations here have previously been discussed as lenition, they could be fortition or strengthening. The literature seems to view these kinds of alternations predominantly as weakening. I assume this may be due to a bias, probably rooted in recapitulating historical change in the synchronic phonology, toward choosing a strong underlying segment that weakens. This bias is particularly apparent in previous analyses of Spanish, since historically the Spanish voiced stops weakened to approximants in most positions. The current distribution of the stops is very restricted (utterance-initial and after homorganic sonorants) and a phonemic analysis that posits strengthening of underlying approximants can be justified just as easily as one with underlying stops. In this book, as I discuss in more detail later, I assume that the approximants are underlying in Spanish.

Consonant strengthening has recently been the subject of a number of studies (e.g. Hock 1992, Fougeron and Keating 1997, Fougeron 1998). Strengthening almost always occurs in the initial position of some constituent, either the word (e.g. Hooper 1976, Hock 1992) or a higher prosodic constituent (e.g. Fougeron and Keating 1997, Fougeron 1998). Fougeron and Keating (1997) show increased linguopalatal contact and Fougeron (1998) shows extremes in a variety of articulatory factors, including linguopalatal contact, in the initial positions of prosodic constituents. Though the term strengthening has seldom been used to describe alternations that are conditioned by the position of stress, my phonetic results include instances of considerable strengthening conditioned by stress.

1.2.3 Environments

Environments are an essential part of any discussion of lenition. Textbooks often describe lenition as occurring in the weak intervocalic or word-final environments. The canonical examples of lenition given earlier in (1) through (3) all occur either between vowels or between sonorants. I only examine intervocalic weakening in the present study. We cannot be sure if an alternation actually represents weakening unless we know its environment. Although word-final consonant alternations are often referred to as lenition, I do not consider them part of lenition proper. Coda position often licenses fewer contrasts than other positions (e.g., Ito 1989), and word-final alternations could be described as the loss of features, such as voicing, that are m...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Table of Contents
• Preface
• Acknowledgments
• List of Tables
• List of Figures
• Chapter 1. Introduction
• Chapter 2. Survey of Consonant Strength Alternations
• Chapter 3. Case Studies of American English and Mexican Spanish
• Chapter 4. Results Patterning by Manner of Articulation
• Chapter 5. Results Patterning by Position
• Chapter 6. Discussion and Conclusions
• Appendix A. English Durations
• Appendix B. Spanish Durations
• Appendix C. Data Base Language References
• Bibliography
• Index