Prosodic Model Research Articles

A prosodic account of complex predicate acquisition in Mam: A Mayan language

The Mayan language Mam uses complex predicates to express events. Complex predicates map multiple semantic elements onto a single word, and consequently have a blend of lexical and phrasal features. The chameleon-like nature of complex predicates provides a window on children’s ability to express phrasal combinations at the one-word stage of language development. The ubiquity of complex predicates in the adult language insures that children will produce complex predicates as some of their first words. The verb complex in Mam has obligatory inflections for aspect, person, and to a degree direction. The inflections vary in degree of attachment between syllable segments, affixes, and clitics. Inflections with vowels are phonologically free, while inflections without vowels attach as either syllable segments or affixes. The Mam verb complex requires the addition of a phrasal layer to prosodic models of lexical acquisition. The paper used this extended version of prosodic theory to make five predictions for the acquisition of the verb complex. The paper analyzes production data for three children between 2;0 and 2;8 acquiring the northern variety of Mam spoken in San Ildefonso Ixtahuacán, Guatemala. The children’s production data for both the intransitive and transitive verb complexes support all five predictions to some degree. The children produced prefixes more frequently on vowel-initial stems than on consonant-initial stems, and they produced imperative suffixes more frequently than prefixes on consonant-initial stems. The children exhibited developmental differences and produced phrasal contractions that the prosodic theory did not predict. The results underline the need to integrate prosody into models of morphosyntactic development, and highlight the significance of complex predicates for theories of language acquisition.

PHONETIC TYPOLOGY

In spite of many works in the field of phonetics some problems of typology still want systematisation both on segmental and suprasegmental levels. Experimental research is one of them. The research is aimed at determining phonetic typologi- cal characteristics in English and Ukrainian on segmental and suprasegmental levels. The results of the research demonstrate the significant amount of typo- logically similar characteristics in English and Ukrainian. The presence of fore lingual and back lingual vowels, plosives, fricatives, voiced and voiceless conso- nants, such phenomena as reduction and assimilation in both languages under investigation testify to it. Similar mechanisms of word stress production that is the total energy data (dynamic characteristics and duration time) is another mark of typological similarity of English and Ukrainian. The intonation models, the degree of stability of intonation models, sentence stress variability, the same functions of intonation in the process of speech produc- tion show similar mechanisms of prosody elements’ interaction. The most stable in both languages are the prosodic models of attitudinal and emotional functions of intonation. The status of different functions (the attitudinal function, sense-group delimitation, sense-groups relation, communicative type of utterances, sentence stress) of intonation changes from the subordinate one (as in the function of di- vision into sense groups) up to the dominating one (as in the function reflect- ing the speaker’s emotional state). In correspondence with this status and the poly-functional character of intonation the resulting intonation contour mainly preserves the relevant prosodic features of the dominating intonation function in English and Ukrainian. The qualitative and quantitative differences of segmental (phonematic length of vowels in English, phonemic length of consonants in Ukrainian, some difference in articulation) and suprasegmental units (variety of melody contour structure, prevalence of dynamic characteristics in English and temporal in Ukrainian) in the languages under investigation are explained by linguacultural peculiarities of the communities, their historical development

Toward enriched decoding of mandarin spontaneous speech

A deep neural network (DNN)-based automatic speech recognition (ASR) method for enriched decoding of Mandarin spontaneous speech is proposed. It adopts an enhanced approach over the baseline model built with factored time delay neural networks (TDNN-f) and rescored with RNNLM to first building a baseline system composed of a TDNN-f acoustic model (AM), a trigram language model (LM), and a recurrent neural network language model (RNNLM) to generate a word lattice. It then sequentially incorporates a multi-task Part-of-Speech-RNNLM (POS-RNNLM), a hierarchical prosodic model (HPM), and a reduplication-word LM (RLM) into the decoding process by expanding the word lattice and performing rescoring to improve recognition performance and enrich the decoding output with syntactic parameters of POS and punctuation (PM), prosodic tags of word-juncture break types and syllable prosodic states, and an edited recognition text with reduplication words being eliminated. Experimental results on the Mandarin conversational dialogue corpus (MCDC) showed that SER, CER, and WER of 13.2 %, 13.9 %, and 19.1 % were achieved when incorporating the POS-RNNLM and HPM into the baseline system. They represented relative SER, CER, and WER reductions of 7.7 %, 7.9 % and 5.0 % as comparing with those of the baseline system. Futhermore, the use of the RLM resulted in additional 3 %, 4.6 %, and 4.5 % relative SER, CER, and WER reductions through eliminating reduplication words.

Establishing Relationship between Emotional Prosody and Affective Pragmatics

This work aims to provide theoretical grounds to establish a relationship between emotional prosody and affective pragmatics. Affective pragmatics is a significant yet least studied area of research. This theoretical framework focuses how emotional expressions are constructive in channelizing the pragmatic meaning under the umbrella of affective pragmatics and also encompasses the speech engineering that conveys unabridged abstract emotions in the phenomenal process of emotion recognition. Since speech act theory focuses meaning at utterance level and not at emotional level, thus, there is a need to reflect on emotional expressions that function as paralinguistic features. There are several studies carried out on identification of emotions using prosodic modeling; however, there is no meticulous study that shows relationship between emotions and their pragmatic meanings. The current study will be an application of Theory of Affective Pragmatics (TAP) proposed by Andrea Scarantino in 2017, a theory analogous to speech act theory. The objectives of the study can be achieved through analysis of quantitative data procured taking advantage of the emotional recordings from Emotional Prosody Speech and Transcripts (EPST), a worldwide database used in emotion recognition processes, along with employment of appropriate prosodic features using a Hidden Markov Model (HMM), a popular emotion recognition statistical model. The outcomes of this study will contribute to benefit new researchers in the field of linguistics to understand affective pragmatics at a profound level as a novel area of research.

Das Verhältnis geschriebener und gesprochener Sprache im 18. Jahrhundert am Beispiel der Interpunktion in Sprache und Musik

Abstract The oral realization of written punctuation basically brings two parameters into play: rhythm and intonation. The former is visualized in writing the typeface by a space, while the latter is determined by conventions and depends on the speaker’s personal means of expression. In the 18th century, the oral realization of punctuation was only marginally addressed in grammars. Additional sources can be found in the area of music history, because the music theorists of the 18th century discussed the connection between language and music with great interest. In this essay, prosodic models for the oral (audible) realization of punctuation marks are examined by comparing the current prosodic description of German with the musical analyses of Johann Mattheson and the remarks in the grammars of his time. Then baroque music examples are examined from the point of view of the possibility of refining the theoretical models. The aim of this interdisciplinary approach is to make use of ancient sources, both theoretical and practical, for the reconstruction of the prosody of 18th century German.

Paraguayan Guaraní and the typology of free affix order

Cross-linguistically, affix order is commonly determined by semantic scope (Rice 2006) or a morphological template. Less frequently, affix order is free, which means that suffixes can be reordered without a concomitant change in scope. To address the question of what gives rise to and constrains free affix order (FAO), I present a case study of Paraguayan Guaraní (or PG, Tupí-Guaraní, Paraguay, ISO 639-3: gug). I argue that FAO in PG should be analyzed as driven by prosodic factors. The prosodic analysis has previously been proposed only for Chintang (Bickel et al. 2007). Two major analyses of FAO see the phenomenon as driven by either morphology (e.g. Ryan 2010) or prosody (Bickel et al. 2007). The morphological analysis proposes that FAO is a consequence of free variation within the morphological template. The prosodic analysis models FAO with prosodic subcategorization for phonologically prominent positions. I argue that the two analyses make different predictions as to the preconditions for and the extent of FAO. I show that both the morphological and the prosodic profile of FAO are attested. I propose that FAO in PG is an instance of the latter. Thus, I demonstrate that FAO is not a unified phenomenon, but rather should be typologized as driven by either morphological or prosodic factors.

Performance analysis of different intonation models in Kannada speech synthesis

Text <span lang="EN-US">to speech (TTS) is a system that generates artificial speech from text input. The prosodic models used improve the quality of the synthesized speech especially naturalness and intelligibility. The prosody involves intonation, intonation refers to the variations in the pitch frequency (F0) with respect to time in an utterance. This work mainly concentrates on building feedback neural network model to predict F0 contour in the utterances using Fujisaki intonation model parameters as the input features to the network since the Fujisaki intonation model is data driven and not a rule based one. In this work we have built 4-layer feedback neural network in the festival framework. Finally, the synthetically generated Kannada speech using the neural network model, is compared for its performance with the classification and regression tree (CART) model and Tilt model. Database of simple declarative Kannada sentences created by Carnegie Mellon University have been deployed in this work. From the study it is very clear that F0 contours can be accurately predicted using CART and neural network models, whereas naturalness and intelligibility is high in CART model rather than neural network model.</span>

The American English spontaneous speech minicorpus. Architecture and comparability

This paper presents the American English (AE) minicorpus, a spontaneous speech resource created within the auspices of the C-ORAL-BRASIL project consisting of texts selected from the Santa Barbara Corpus of Spoken American English. We focus on the sampling strategy that guided the selection of texts, the transcription criteria that were implemented and the prosodic and informational annotation carried on the AE minicorpus. The minicorpus was designed to be comparable to the minicorpora of the C-ORAL projects for Italian and Brazilian Portuguese, which were conceived to allow the study of information structure in spontaneous speech in accordance with the principles of the Language into Act Theory. This theory comprises a pragmatic framework for the study of spontaneous speech and it integrates the IPO approach into its prosodic model. The IPO approach consists of a perception-based model for the study of intonation, providing an apparatus for the description and classification of melodic contours observed in spontaneous speech.

THE VARIANT NATURE OF THE BRITISH WOMEN'S SPEECH BEHAVIOR

The question of prosodic variation in women’s speech has been analysed in the present article. Women’s of different social position (age and profession) speech behavior is in the centre of the given investigation. Women’s speech, being the subject of the analysis, allowed to reveal their specific peculiarities and arrange their speech behavior. The aim of the investigation is to define prosodic markers of women`s speech behavior. The subject of investigation is variation of women`s speech prosodic models. The object is unprepared monological speech of TV announcers and university lecturers as representatives of the analysed professions. The following tasks have been set: to distinguish main trends in gender linguistics, with the aim of defining speech behavior; to conduct phonetic analysis; to reveal differentiating prosodic characteristics, identifying women's speech behavior. In accordance with the aim and tasks of investigation, perceptive auditory analysis was applied which allowed to identify and interpret the analysed phonetic data and method of intoning was used to describe speech intonation. The concluded investigation allowed to analysed contemporary British women`s speech behavior with the help of prosodic peculiarities of their speech. The auditory analyses allowed to come to the following conclusions: speech behavior of successful women-announcer is characterized by prevalence of Falling Scale and Falling terminal tone, which identify them as confidant, categorical in speech women, who can convince interlocutors, influence them, as the main aim of announcers speech is to reach the listener, to make information maximum available and understandable. Speech behavior of university lecturers is characterised by more variable character. Their speech is dominated by Rising, Wavy scales, which is explained by peculiarities of the profession: they masterfully use their voice tember, and high tonality of speech testify their skills to adopt speech behavior to communication conditions. Thus prosodic means, which form during the act of communication may characterise the speaker`s originality, different by age and profession.

Prosodic correlation between the focusing adverb ozik ‘only’ and focus/givenness in korean

This study investigates the prosodic correlation between the focusing adverb ozik ‘only’ and focus/GIVENness in Korean. The goal is to test the issue of whether or not the element associated with ozik can be dephrased in Korean. This question is answered by a perceptual experiment in which the pitch contours of target sentences (i.e. Jeoneun ozik gichareul tamnida ‘I only take the train’ and Jeoneun ozik daehanhanggongeul iyonghamnida ‘I only take Korean Air.’) are modified using Praat. The modified sentences are embedded in two competing contexts in which focus/GIVENness on the element associated with ozik (i.e. gichareul ‘the train’, daehanhanggongeul ‘Korean Air’) is elicited. The results reveal that DoubleH (where both the focusing adverb and the object have prosodic prominence) is shown to be the most favored prosodic model when the element associated with ozik is focused in discourse. Conversely, OzikH (where the focusing adverb has prosodic prominence) is shown to be most favored when the element associated with ozik is GIVEN. Therefore, we argue that the element associated with ozik can be either prosodically prominent or dephrased depending on the discourse. This finding demonstrates that discourse functions such as focus and GIVENness are not directly contingent on the lexical semantics of the focusing adverb but instead are constrained by the discourse.

Towards a prosodic model for Tiberian Hebrew: An intonation-based analysis

This study advances a preliminary framework for conceptualising the prosodic nature and structure of Tiberian Hebrew (TH) represented by the ṭaʿămê hammiqrāʾ through an analysis of an extant Ashkenazi cantillation tradition. The ṭaʿămê hammiqrāʾ (lit. “the senses of the reading [viz. Scripture]”) are notations added by medieval scribes to the written text of the Hebrew Bible to preserve and transmit its oral performance. Modern prosodic theory and the musical concept of conjunct and disjunct melodic motion are used to demonstrate that the ṭaʿămê hammiqrāʾ have a highly structured iconic and intonational basis that organises the system and conforms substantially to Selkirk’s (2000, 2009) cross-linguistic prosodic hierarchy. The intonation-based prosodic model proposed in this study offers a solution to the limitation Dresher (1994, 2013; see also Dresher and DeCaen 2018; DeCaen and Dresher 2020) encounters with the intonational phrase domain of his prosodic model, permitting an alternative analysis of so-called pausal forms as lengthened forms, which can occur at prosodic phrase boundaries regardless of pause. The intonation-based model is tested by assessing how accurately it reflects the cross-linguistic prosodic features of restrictive and nonrestrictive relative clauses. The results indicate that TH distinguishes three prosodic classes of relatives—prosodically marked restrictives, prosodically marked nonrestrictives, and prosodically undifferentiated relatives—findings that accord with Birkner’s (2012) intonation-based study of the prosodic structure of German relative clauses.

Взаємодія просодичних характеристик у реалізації висловлювань-здивувань

The aim of the article is to define the role of prosodic variables play in actualizing emotional utterances. Sometimes they can be the only ones that indicate the speaker’s pragmatic intention. The purpose of the study is to determine the interaction level of prosodic means that helps to achieve maximum effect on the recipient as well as to establish the invariant of the intonation model of utterances expressing surprise and its variable modifications. The correlation of the intonation components and parts of the intonation contour is able to provide an adequate description of positive, negative, indefinite and complex emotions. Utterances expressing surprise of different kinds are characterized by the variability of a pitch-tone level, tempo, timbre, rhythm, each of which carries certain semantic meaning. At the same time, the type of a nuclear tone, scale, pre-scale, core elements of the melodic contour, interacting with the components of intonation define the modal meaning of an utterance. To determine the common and distinct perceptual prosodic characteristics of utterances expressing surprise, as the most frequent emotional speech units of business and cultural communication, with a view to define the prosodic model of the utterances explored, the auditory and acoustic analyses were performed. The perceptive and acoustic analyses comparison allowed us to formulate the general structure of the intonation patterns of utterances expressing surprise. It is concluded that the leading prosodic means, which differentiate the types of utterances expressing surprise, should include: the tone level of the beginning of the syntagm, the tone of the first stressed syllable, the configuration of the terminal tone and its range, the rate of pitch frequency change in the terminal rhythm group, the frequency range of the intonation group, the tone intervals on the pre-scale, scale, nuclear tone as well as tempo and timbre. Linguistic interpretation of the results of the perceptual and acoustic analyses allows us to consider the established intonation features sufficiently complete to describe their prosodic model and define them as appropriate for the use in teaching proper intonation of the English emotional utterances. The achieved data can be used in teaching students the most common intonation patterns of emotional utterances and to make them realize semantic variabilities of utterances expressing surprise.

MELODIC PECULIARITIES OF REGIONAL DIALECT REPRESENTATIVES’ SPEECH

The article is performed in the framework of modern sociophonetics. Functioning of English in Scotland is characterized by correlation of regional pronouncing norm, dialects of Scots and also Gaelic language. In bilingual situation, interference of language system led to appearance of the majority of mixed language forms. The given article deals with the attempt to reveal in Glasgo informants nonformal spontaneous speech of the two age group representatives prosodic markers of their speech behavior. Emphasized during the analysis age prosodic peculiarities of the Scots speech behavior revealed characteristic differentiating features of each investigated group. Differentiating features of the Glasgo speech dialect age group 55-65 years, male representatives are the following: fractional syntagmatic devision, prevalence of High Head and Rising terminal tone of the final intonational group. Typical speech behavior age group 19-20 years, male college students is a highly fractional delimitation of speech and usage of emphatic terminal tones, which on the background of narrow voice range create the effect of emotionally neutral speech. Specific peculiarity of Scotish dialect speakers is the usage of low unstressed pre head in combination with high rising nuclear syllable and rising level of unstressed tail. Thus, taking into account the conducted analysis we may state that elements of urban speech style most vividly presented in the youth speech, on the background of emotionally colored prosodic models usage. The results of the perceptive analyses will support further study of sociophonetic variation of the Scottish speech.

Integrating lexical and prosodic features for automatic paragraph segmentation

Spoken documents, such as podcasts or lectures, are a growing presence in everyday life. Being able to automatically identify their discourse structure is an important step to understanding what a spoken document is about. Moreover, finer-grained units, such as paragraphs, are highly desirable for presenting and analyzing spoken content. However, little work has been done on discourse based speech segmentation below the level of broad topics. In order to examine how discourse transitions are cued in speech, we investigate automatic paragraph segmentation of TED talks using lexical and prosodic features. Experiments using Support Vector Machines, AdaBoost, and Neural Networks show that models using supra-sentential prosodic features and induced cue words perform better than those based on the type of lexical cohesion measures often used in broad topic segmentation. Moreover, combining a wide range of individually weak lexical and prosodic predictors improves performance, and modelling contextual information using recurrent neural networks outperforms other approaches by a large margin. Our best results come from using late fusion methods that integrate representations generated by separate lexical and prosodic models while allowing interactions between these features streams rather than treating them as independent information sources. Application to ASR outputs shows that adding prosodic features, particularly using late fusion, can significantly ameliorate decreases in performance due to transcription errors.

ОСОБЛИВОСТІ СЛУХОВОГО СПРИЙНЯТТЯ ІНТОНАЦІЇ ІНТЕРОГАТИВІВ У ТЕАТРАЛЬНОМУ ПУБЛІЧНОМУ ВИСТУПІ ЯК РІЗНОВИДІ ДРАМАТИЧНОГО МОНОЛОГУ

The article deals with specific features of perception of public speaking prosody. Speech intonation is an important component of oral text, a carrier of semantic meanings, and at the same time a means of demonstrating the emotionally expressive nature of expression. The nature and influence of public speaking depends not only on the adequately disclosed facts, but also on the very speech form of their presentation. The leading role in this process belongs to prosodic means, in which the peculiarities of public broadcasting are clearly revealed. The article examines the interrogatives of quasi-spontaneous public speakingmonologues from Shakespeare's plays, gathered from audio recordings of different periods of time, dating back to 30 years of the last century. This approach made it possible to compare the main features of interrogative rhetorical questions in the chronological plan. In the analysis of prosodic features of question structures, the inventory of their universal characteristics was distinguished,since there are certain normative, invariant prosodic models for the main types of questions. On the other hand, the melody should be attributed to the most variable parameters of interoactive constructions, so the most typical variants of intonational deviations were also analyzed.

Study on French Baroque singing by accent and intonation segmentation

The baroque music is modelled on the language, so the claim of the composers of the time and the report of the musicians and musicologists of today. It is thus national because it always refers to a specific language. Typically, we consider as a proof the work of the composer Jean-Baptiste Lully who, as the saying goes, would have formed his recitatives according to the declamation of the actress Champmeslé. Starting from a specific linguistic material, Baroque music is national: it is French, Italian, German, English ... Numerous studies, in particular musicological, have shown the treatment of language rhythm among composers. In this article we will discuss the subject from a wider angle and try to show by an accentuated and intonational segmentation of the text and the melody composed by the musicians, to what extent the prosody of French and the French Baroque music correspond.We use the prosodic segmentation model proposed by Piet Mertens for French, and we base our study on a body of recitatives, because this genre is considered extremely close to declamation, and therefore to the spoken language. Despite the differences between the language used in the compositions and the standard Parisian French of the 21st century, due to the style and evolution of French, Mertens' approach proves to be convincing in describing the compositional setting of the text by composers French Baroque. The research thus confirms the close proximity between textual and musical prosody for the French Baroque recitative.

Complex illocutive units in language into act theory

This work presents a preliminary analysis for a prosodic description of two different spoken structures in spoken language within the theoretical framework of the Language into Act Theory (L-AcT): (i) chains of two or more Bound Comments (COB) that do not form a compositional informative and prosodic unit; (ii) compositional Information Units formed by two or more Multiple Comments (CMM) of the List type, linked together by a conventional prosodic model that implements a specific meta-illocutive structure . The goal of this study is to underline specific features of the COB units and the List-type CMM units, detecting prosodic properties of Italian and Brazilian Portuguese spoken language. Through a specific script for Praat software, different parameters are automatically calculated: f0 reset, slope and variation rate, pause duration, spectral emphasis. Our results highlighted a common prosodic behavior in COB-units in terms of f0 movement (rising in the stressed syllable before the break and falling in the unstressed one just before the break), and high similarity between the two COBs and Lists, but also the need to distinguish the effects connected to the position of the stress from the specific features of the unit as detectable Textual Unit.

Génération automatique de la structure prosodique en français

An automated process for building a prosodic structure form transcribed speech recordings in French is presented, based on the incremental prosodic model [1, 2, 3]. In this model, the prosodic structure is defined incrementally by dependency relations instantiated by melodic contours located on the last syllable of the last word of stress groups, subject to a rhythmic constrain limiting the gap between successive stressed syllable to a 250-1250 ms range. Although they frequently contain lexical words (noun, verb, adverb, adjective), stress groups in French can also include only grammatical words (pronoun, conjunction, preposition). Melodic contours are phonologically defined from their melodic rise or fall and their glissando value ensuring their function as dependency markers between stress groups. The algorithm proceeds from an orthographic transcription as follows: 1. Automatic segmentation of the orthographic text into IPA and word tiers 2. Automatic annotation of stressed vowels in three classes (followed by 250 ms silence, above the glissando threshold and lexical category based) 3. Assignment of melodic contours from fundamental frequency values at stressed vowels boundaries. Comparisons with automatic and manual stressed syllable annotation on existing corpora are given, showing the validity of the phonological rules implemented in the algorithm.

FOCALIZATION AS LINGUISTIC PHENOMENON

The article reveals the principal notions of focalization, its functions and features (emphasis and contrast) and their influence on transforming the model of the prosodic structure of the actual division of the utterance. Most frequently, focalization of a text fragment is prosodically expressed with a high register and extension of the tonal range, a raised loudness and a distinct pronunciation. It is established that one of the most well-known factors of the transformation of the prototype prosodic model is the listener’s focus on the most important aspects of the message, the conveyance of the semantic interaction of the message’s specific aspects. Focalization is interwoven with the both possible manifestations of this function: logical stress (in case of the opposition of words) or emphatic stress is implemented by its means. Regarding the changes in the tone direction, it should be noted that this tonal feature does not participate directly in forming an intonation model of the discourse’s focalized structures. It has been observed that there exists a certain modification of the tonal contour formed under the influence of the illocutionary and phase factors and certain types of the subjective-modal meanings. Key words: emphasis, opposition, focalization, actual division of the utterance, theme, rheme.

Role of Prosodic Reading in Listening Comprehension

This study aims to determine how listening comprehension levels of students are affected by listening to prosodic and non-prosodic readings vocalized by a computer and human. Third-grade students of four different classes at a primary school were randomly selected in a city center in the Western Black Sea Region of Turkey to participate in the study (n=91). Four equivalent classes formed the listening groups of the research in listening comprehension. The groups were (1) the group listening to the model prosodic reading, (2) the group listening to the computer prosodic reading, (3) the group listening to the model non-prosodic reading, and (4) the group listening to the computer non-prosodic reading. Two stories were used in the measurement of listening comprehension, and comprehension skills were measured with open-ended questions. The data obtained were analyzed with the Kruskal Wallis and Conover–Iman tests. Logistic Regression Analysis (LRA) was performed to reinforce the results and increase distinctiveness. According to the results, inferential comprehension scores of the students who listened to the prosodically-vocalized texts differed from other groups significantly. Meanwhile, literal comprehension scores of all students in the listening groups did not differ. The LRA results indicated that the inferential comprehension scores were a significant predictor of the listening groups. Consequently, the relationship between prosody and inferential comprehension was found to be significant in this study. The results also showed that it is necessary to attach particular importance to prosody in listening activities and to use prosodic models suitable for students in reading activities in the early grades of the primary school.