This chapter looks at the acquisition of comparatives from formal, theoretical, and cross-linguistic perspectives. It begins by reviewing children’s aberrations from adults in the form of the comparative constructions that they produce through at least age 6, and then turns to theoretical accounts of comparatives and degree constructions across a range of languages to pinpoint specific areas in the construction of a comparative in which children’s representations and interpretations may go astray, or converge with adults. A range of studies and methodologies used over the years are reviewed in order to present a clear picture of what we currently know about children’s developing understanding of comparison and comparatives, and to clear a path for future research in this area.
In amphichronic phonology, synchronic and diachronic explanation feed each other. The architecture of grammar predicts the possible modes of implementation of phonological change (including neogrammarian regularity) and the life cycle of sound patterns. In turn, the life cycle accounts for synchronic phenomena such as scattered rules and the relative stratal affiliation of cognate processes.
Launched in 1989, when Elizabeth Gordon became aware of the existence of a set of recordings of early New Zealand English, the Origins of New Zealand English Project (ONZE) provides a unique opportunity to investigate sound change. These “Mobile Unit” (MU) recordings were made in the 1940s for radio broadcast, and included reminiscences from speakers born as early as the 1850s. The recordings are significant because they date to the first stages of large-scale immigration to New Zealand from the British Isles, representing the first generation of English speakers born in New Zealand. These recordings, which are in the possession of the University of Canterbury, form the core of the ONZE project. Interviews with more than 100 of these early speakers have now been compiled, digitized, transcribed, time-aligned, and then automatically segmented at the phoneme level. The Mobile Unit recordings have been a valuable tool for testing theories of the formation of new dialects.
This article provides details on human speech production involving a range of physical features, which may have evolved as specific adaptations for this purpose. All mammalian vocalizations are produced similarly, involving features that primarily evolved for respiration or ingestion. Sounds are produced using the flow of air inhaled through the nose or mouth, or expelled from the lungs. Unvoiced sounds are produced without the involvement of the vocal folds of the larynx. Mammalian vocalizations require coordination of the articulation of the supralaryngeal vocal tract with the flow of air, in or out. An extensive series of harmonics above a fundamental frequency, F0 for phonated sounds is produced by resonance. These series are filtered by the shape and size of the vocal tract, resulting in the retention of some parts of the series, and diminution or deletion of others, in the emitted vocalization. Human sound sequences are also much more rapid than those of non-human primates, except for very simple sequences such as repetitive trills or quavers. Human vocal tract articulation is much faster, and humans are able to produce multiple sounds on a single breath movement, inhalation or exhalation. The unique form of the tongue within the vocal tract in humans is considered to be a key factor in the speech-related flexibility of supralaryngeal vocal tract.
Empirical research often involves three activities: the systematic annotation of audiovisual media (coding), the management of the resulting data in a corpus, and various forms of statistical analysis. This chapter presents ANVIL, a highly generic and theory-independent research tool that supports all three activities in conjunction with audio, video, and 3D motion capture data. The tool allows to specify a formal coding scheme as a blueprint for project-specific coding. The process of coding is conducted on parallel time-aligned tracks. Instead of simple tags, ANVIL offers typed attributes for describing the basic annotation elements which can be used to hide away complexity and reduce visual clutter. Track types (interval vs. point) and logical track relationships (singleton, span, subdivision) further increase the clarity and consistency of the coding. Spatial mark-up on the video frame allows annotation of regions over time. For corpus management, ANVIL allows annotation files to be grouped into projects for browsing, export, and analysis across a corpus of data. For analysis, the tool supports simple descriptive analyses like transition diagrams or label frequency histograms and more complex operations like automatic inter-coder agreement computation (Cohen’s kappa).
Dina El Zarka
This overview of intonation in Arabic compares the intonational systems of selected Arabic dialects from Morocco in the West to Kuwait in the East. The formal comparison will mainly be carried out within the framework of autosegmental-metrical (AM) theory, taking the phonetic micro-prosody of the identified pitch accents as a tertium comparationis. Furthermore, the intonation systems will be compared with respect to prosodic phrasing. The second part of the overview is devoted to the functions of intonation in Arabic. In this section, the comparison will be based on a wider range of descriptions, including work carried out within other theoretical frameworks. The section will identify the role of metrical and tonal structures and the way they interact with syntax, information structure, and sentence mode in different varieties of Arabic. The concluding section will provide a preliminary typological picture of Arabic prosody with respect to the macro-rhythmic properties of Arabic.
Joshua Viau and Ann Bunger
Children acquiring any language must develop an understanding both of how event components are encoded in verb meanings and of the argument structure of those verbs, that is, how the participants of the event that each verb describes map onto linguistic arguments. This chapter begins with an overview of the major issues in the study of argument structure, including a consideration of the balance of power between verbs and constructions as it pertains to the encoding of thematic relations and a comparison of theoretical approaches with an eye toward learnability. The core of the chapter consists of a comprehensive synthesis of the current state of developmental research on argument structure.
Articulatory Analysis and Acoustic Modeling: Articulatory To Acoustic ModelingUltrasound As a Tool For Speech ResearchMethodologies Used to Investigate Laryngeal Function and Aerodynamic Properties of SpeechOn The Acoustics and Aerodynamics of Fricatives
Khalil Iskarous, Lisa Davidson, Helen M. Hanson, and Christine H. Shadle
This article describes theory and research methods employed for articulatory, acoustic, and aerodynamic analysis of speech. One of the theories, dispersion-focalization theory (DFT), combines two ideas that include focalization and contrast maximization. Focalization is a property that emerges from acoustic model nomograms and refers to points where constriction placement results in formants being close to each other (focal points). The theory distinguishes between independent and non-independent secondary contrasts. Independent secondary contrasts are secondary contrasts that do not interact with the primary vowel contrasts, while non-independent contrasts are secondary contrasts that affect the primary ones. The principle of distinctive region model (DRM) is that different regions of the vocal tract have uniform acoustic behavior. The results of DRM are based on how formation and release of constrictions affect the formants at different locations within the vocal tract. The DRM theory of linguistic contrast is based on the pseudo-orthogonality of the discrete regions and on a dynamic articulatory-acoustic principle stating that segments are preferred, which allow for the least motion from one segment to another, while maximizing contrast. Ultrasound imaging is increasingly been used to address specifically phonological questions. The technology is also used to examine the status of excrescent schwas in various phonological environments.
Joan L. Bybee
This chapter discusses the role of articulatory processing in sound change, emphasizing the tendency towards reduction and overlap of articulatory gestures, as well as explanations proposed for this tendency. The pattern of lexical diffusion proceeding from most to least frequent words and phrases is discussed as evidence for the important role of articulation in sound change.
Austronesian is the second largest language family on earth in number of languages, and had the widest geographical extent of any language family prior to the modern era. For these reasons it shows great typological diversity in word-formation processes, ranging from extremely elaborate systems of affixation in Philippine-type languages to systems that depend far more on compounding in some of the languages of Melanesia. Some of the more striking devices used in word formation include subtractive morphology in vocative forms, stress shifts as signalers of word-class change, and an extremely rich inventory of reduplication processes, including several that are theoretically unexpected, as well as the active use of triplication as a process that is distinct from serial reduplication.
This chapter surveys the main phonological phenomena found in Bantu languages. One such phenomenon is vowel harmony, where suffixes typically alternate in height depending on the preceding vowel. Combinations of nasal plus consonant are also frequently subject to various modifications; such clusters play an important role in debates over syllable structure, since Bantu languages seem to exhibit the peculiarity that the only onset clusters of consonantals are composed of a nasal followed by an obstruent, reversing the more general linguistic tendency for sonority to rise in the syllable onset. The tone systems of Bantu languages are especially well known for their complex systems of alternations in the form of spreading, dissimilation, and grammatically conditioned melodic alternation. A key feature of interest in the study of Bantu phonology is the considerable variability in the exact details of operation of a highly similar set of rules.
This chapter discusses the most fundamental types of phonological change. The first part is a presentation of the basic notions underlying virtually any discussion in historical phonology (conditioning of changes, the phonological levels affected, basic structural consequences, persistent rules vs. sound change). In the other part the major types of sound change (featural, segmental as well as prosodic) are presented under nine headings (assimilation, dissimilation, deletion, insertion, lenition, fortition, metathesis, lengthening and gemination, shortening and degemination). The goal of this chapter is to provide a theory-neutral presentation of these general notions.
The present article poses some fundamental questions related to bilingualism and to the acquisition of two phonological components, by very young children. It discusses different types of bilingualism and their outcomes. After a brief consideration of alleged pros and cons of bilingualism brought up in the past decades, two perspectives of bilingualism are sketched—psycholinguistic and sociolinguistic—and certain aspects of bilingual child phonology are presented from each of these points of view. The essential issue is whether different outcomes of bilingual child phonology are predictable, and to find the crucial criteria to support the predictions. Finally, the discussion addresses some basic questions about bilingual acquisition, and ends with a summary of various types of cross-linguistic interaction.
Although changes in stress systems are not as meticulously described and discussed as other aspects of phonology, they are recorded in the historical literature. Stress is related to quantity and weight and any change thereof may lead to an alteration of stress patterns. Usually, a change in the direction of stress assignment is rare, although abrupt changes from left-edge to right-edge stress or vice versa are known to happen. Causes for such changes are frequently assumed to be rooted in language contact. This chapter argues that stress patterns are surprisingly pertinacious, and that universal metrical preferences and constraints govern possible and impossible prosodic shifts. Rather than external influence alone, the chapter argues that acquisition and learnability can account for the data more coherently. It first considers general issues in the change of stress parameters and then focuses on the history of English, to exemplify some possible types of change on the basis of the prosodic changes that have occurred in that language.
Changes In Representations: The Nature of Historical ChangeThe Relationship Between Synchronic Variation and Diachronic ChangeModeling Exemplar-Based Phonologization
Ioana Chitoran, Jonathan Harrington, and Robert Kirchner
This article addresses the nature of historical change, focusing on experimental work presenting the motivations and mechanisms for language change. The two models of phonetic variation explaining sound change include Lindblom's H&H theory and Ohala's phonetic listener-based model. A smaller category of sound change falls under the scenario that Ohala calls hypercorrection, whereby the listener performs an unnecessary, inappropriate correction of the signal, and ends up producing a new form. Hypercorrection often results in dissimilation. Gestural reassignment captures the listener's failure to identify correctly the source of a particular property of the signal, as in Ohala's model. Gestural misparsing can also explain cases involving the apparent insertion or deletion of a gesture. The most widely cited formulated model of phonologization is by Hyman. The process involves two steps that include phonetic variation leading to phonological variation (phonologization), and phonological variation leading to distinctive variation (phonemicization). The speech recognition involves a calculation of distance in phonetic space between an auditory stimulus and the stored exemplars, and the application of a classification rule to these distances. Exemplar-based speech production involves generation of an output based on mean phonetic properties of the exemplars of the target category.
Janet B. Pierrehumbert, Mary E. Beckman, and D. Robert Ladd
This article explains the evolution of laboratory phonology over the years. The laboratory phonology community uses both discrete mathematics and continuous mathematics. It continually evaluates what type of formalism is most suitable and incisive for what types of linguistic phenomena. The work on the articulatory and acoustic nature of phonological categories uses a methodology adopted from physics, in which the behavior of the basic equations of the theory is explored with respect to issues such as stability, linearity, invertability, and effects of boundary conditions. The exact extent of the voicing, its statistical variation, and the dependence of these factors on structural position are involved in a laboratory experiment. One of the major contributions of laboratory phonology to the field of phonology has been the careful documentation of syndromes in language sound structure. Phonological categories are natural in the sense that the actual phonetic denotation of each category shapes its patterning in the sound system. The phonological rules that affect the stops reflect their actual phonetic character. The phonological categories are also natural in the sense that physical nonlinearities in both articulation and acoustics have the result that phonetics is already quasi-categorical.
Corpora, Databases, and Internet Resources: Corpus Phonology with Speech ResourcesUsing The Internet For Collecting Phonological DataSpeech Manipulation, Synthesis, and Automatic Recognition in Laboratory PhonologyPhonotactic Patterns in Lexical Corpora
Jennifer Cole, Mark Hasegawa-Johnson, Dan Loehr, Linda Van Guilder, Henning Reetz, and Stefan A. Frisch
This article introduces a wide range of approaches to using large bodies of data for linguistic research. Corpus analysis for phonological research involves the investigation of the phonetic, phonological, and lexical properties of speech for the purpose of understanding the patterns of variation in the phonetic expression of words, and the distributional patterns of sound elements in relation to the linguistic context. A speech corpus provides a basis for investigating variability in phonetic form and also provides a rich resource for studying the relationship between phonological form and other levels of linguistic structure. Linguistic metadata provides information about the speakers, such as sex, age, ethnicity, and region of residence. Metadata may also provide information about speaker recruitment and recording procedures. Forced alignment is done using algorithms from automatic speech recognition (ASR), and is most successful when each phone associated with the word in its dictionary form is actually fully pronounced. One of the easiest methods of manipulating natural speech is the splicing technique, where parts of a speech signal are cut out, repeated, or cross-spliced with another piece of the signal. The gating technique is another form of natural speech signal manipulation often applied in psycholinguistic experiments, where parts of a speech signal are cut off, and incrementally more of the signal is presented to a listener. Another speech signal manipulation is the mixing of two signals.
The aim of this chapter is to show, through the examination of a central area of French phonology, liaison, how corpora have contributed to a better understanding of the phenomena at play. The chapter adopts a historical perspective in going back to the influential work of S. A. Schane, who in the 1960s offered the first generative phonology treatment of French liaison. It is argued that a number of the issues debated within theoretical models since then are probably dead ends because they are based on data which is too scarce, occasionally spurious, and often uncritically adopted from previous treatments. Corpora have helped in firming up our database and allowing us to chart which areas are relatively stable across speakers and which are variable. There are, of course, cases where the data provided by corpora is neither rich enough nor controlled enough. But our understanding of French liaison is more liable to be advanced by further corpora construction, psycholinguistic testing, and phonetic experimentation than by devising new notational models on the basis of data inherited from tradition.
Elisabeth Delais-Roussarie and Hiyon Yoo
The aim of the chapter is twofold: presenting the different types of data that are used to carry research in phonetics and phonology, and giving some concrete examples of corpus-based research in phonetics and phonology. Despite the fact that both domains of linguistics do have a very tight relationship to data, since their objectives require a systematic study of sound patterning, the types of data they used may vary greatly, ranging from acoustic to articulatory data. Among the acoustic data, it is argued that a distinction has to be made between experimental or laboratory data on the one hand and naturally produced data on the other, even though these two types of data are complementary, and crucial to research in phonetics and phonology. As for the use of naturally produced data in phonology and phonetics, it is shown through some concrete examples (the analysis of liaison in French, the distribution of pauses in specific speech style) that these data can shed new light on the analysis of phenomena that have already been explored.
Elisabeth Delais-Roussarie and Brechtje Post
The aim of the chapter is twofold: explaining the prerequisites for providing a phonetic/phonological annotation of speech data; and presenting the different systems that can be used to encode the phonetic and phonological events present in the speech signal. Since phonetic/phonological annotation can be seen as the assignment of a label to a specific unit in the data, the segmentation of the speech signal and the assignment of labels are crucial tasks in the annotation process, regardless of the system chosen. As to the presentation of the systems, a distinction can be made between systems that are primarily designed to represent the segmental dimension of the speech signal and those that encode prosodic events such as stress, phrasing, and tonal or intonational patterns. In this chapter, we explore the advantages and limitations of the systems presented by considering the different types of speech data that one may want to annotate (standard data or non-standard data such as acquisition data, pathological speech, etc.) as well as the amount of knowledge about the language spoken that the annotator needs to have in order to successfully transcribe speech with the system in question (whether its phonology is known, etc.).