Non-final Position Research Articles

Levels of prominence in Spanish infant-directed speech

Research on Spanish prosody in adult-directed speech (ADS) (Ortega-Llebaria & Prieto 2005, Ortega-Llebaria & Prieto 2011) shows evidence for three acoustic levels of prominence: stressed and focused, stressed and unfocused, and unstressed and unfocused. We tested whether Spanish infant-directed speech (IDS) contains the same levels of prosodic prominence as ADS. In order to test this hypothesis, 30-minute play sessions between adult Spanish-speaking females of Latin American origin (n = 12) and 12-month old infants were recorded. We extracted target words from the recordings containing stressed and unstressed /a/, /e/, /i/, and /o/ in non-final position. Voicesauce (Shue et al 2011) was then used to obtain acoustic correlates of stress in Spanish: pitch, duration, intensity. We also labeled pitch accents using SpToBI conventions (Estebas Vilaplana & Prieto 2009). Latent class analyses to be performed on the data will determine the number of distinct prominence levels present in Spanish IDS as realized by Latin American speakers.

Coronal stop deletion in Hawai'i English

Coronal Stop Deletion (CSD) is believed to be constrained by the phonological environment and the word’s morphological make-up (i.e., monomorphemic, semi-weak past tense, regular past tense) (Guy 1991; Hazen 2011). In this talk, we present results from an analysis of CSD in the speech of four speakers of Hawai'i English. In addition to using a traditional approach (i.e., auditory analysis of CSD in word-final position), we also analyzed CSD in non-final position (e.g., followed by plural -s) and conducted acoustic phonetic analysis of /t/ in tokens where it was realized. The results demonstrate highly significant effects of environment. Additionally, there is no effect of morpheme type; instead, the results provide evidence that deletion is favored in words with a greater number of morphemes, regardless of whether the coronal stop is found in the stem or the affix. Additionally, the acoustic analysis suggests that the effect of following environment is acoustically gradient: tokens followed by a voiceless, non-alveolar consonant (i.e., the environment that most strongly disfavors deletion in our data) are most likely to be realized with complete occlusion in the vocal tract. Implications for models of speech production are discussed.

A preliminary analysis of stop codas in Kandze Khams Tibetan

Research on Tibetan linguistics is scarce. The sounds in Tibetic languages are seldom described systematically, and acoustic-phonetic analyses are especially rare. Much of the existing literature has focused on the Lhasa dialect; other varieties have received far less attention. This study presents data from the under-documented Kandze dialect of Khams Tibetan. The phonetic realization of stop codas and their interaction with the preceding vowel were analyzed, and a sketch of the Kandze vowel space was made. Historically, Tibetan allowed up to two consonants in coda position, as reflected in the written language. In most modern dialects, however, many of the codas have been lost or simplified in some way. Production results from one male native speaker of Kandze showed all three historical stop codas, /b/, /d/, and /g/, being neutralized to a glottal stop word-finally and deleted in non-final positions. The historical /d/ coda, in particular, was almost always associated with fronting of the preceding vowel. The fronted vowels are likely to have originated as allophonic variants of back vowels occurring before coronal codas but became phonologized over time. The historical process of fronting may be attributed to the upward shifts in F2 found in vowels with neighboring coronal segments.

Narrow presentational focus in Mexican Spanish: Experimental evidence

AbstractIt is most often claimed that in Spanish constituents in narrow presentational or information focus appear rightmost, where they also receive main sentence stress, while shifting the stress to the focus in its canonical position is infelicitous. Some, however, claim that Spanish in fact has recourse to both strategies for making the focus prominent, and some recent quantitative work has shown support for this alternative view. The present paper contributes to this debate by experimentally testing the realization of presentational focus in Mexican Spanish using an acceptability judgment task. The results of the experiment reveal that, for these speakers, focused constituents need not be rightmost and can in fact be stressed in non-final position, contra the consensus view. These findings expand the database on focus in Spanish and indicate that theories of the prosody/syntax interface may need to be revised, especially those theories that motivate discourse-related syntactic movement based on the requirements of the prosody.

A phonetic inquiry into Dagbani vowel neutralisations

AbstractPrevious researchers differ on what the feature specifications of Dagbani vowels are, what the vowel inventory is, which vowels are contrastive and how phonological processes affect the surface realisation of Dagbani vowels. This study provides experimental results on Dagbani vowel features from acoustic and perceptual psycholinguistic inquiries with the goal of better defining the vowel features and inventory. Basic formant frequencies of the vowels in different phonological contexts support the view that (1) vowels of all height specifications (including /a/) have [ATR] variants; (2) /i/ and /ɨ/ neutralise in harmonic contexts; and (3) mid vowels surface as low in non-final positions. Results of a perceptual psycholinguistic investigation of the neutralisation between mid and low vowels also show that in certain contexts, speakers are unable to perceive any distinction between different words with these vowels. The paper discusses the implications of these results on the Dagbani vowel inventory, evaluating the merits of different possible approaches in the process.

On the syntax of multi-focused fragment answers in Korean: An oblique merge analysis

Chung, Daeho. 2015. On the syntax of multi-focused fragment answers in Korean: An oblique merge analysis. Linguistic Research 32(3), 573-606. As observed in the literature (Park 2005, 2013, Park and Shin 2014, Ku and Cho 2014, among others), focus elements in a multi-focused fragment answer in Korean behave asymmetrically depending on their positions: Focus elements in a non-final position retain their dependent markers, while those in the final position optionally delete them. This paper first shows that the restriction is merely a subcase of a superordinate constraint, which states that a focus element in a non-final position pied-pipes the minimal node that dominates it and c-commands the following focus element(s), whereas a focus element in the final position optionally pied-pipes its dominating node(s). It is argued then that the pied-piping for non-final focus elements is caused by an economy-driven oblique merge operation that clusters elements with a focus feature before the complex focus element formed in this way moves to the focus licensing functional phrase. Focus clustering due to the same oblique merge operation is attested in other constructions containing multiple focus elements. (Hanyang University)

Production of L3 Vowels: Is it Possible to Separate Them From L1 and L2 Sounds?

It is incontrovertible that acquisition of a sound system of a second language is always a complex phenomenon and presents a great challenge for L2 learners (e.g. Rojczyk, 2010a). There are numerous studies (e.g. Nowacka, 2010; Flege, 1991) which show that L2 learners whose first language has a scarce number of sounds, have problems to distinguish L2 sound categories and tend to apply their L1 segments to new contexts. It may be easily detectable in the case of vowels. There is abundance of studies examining L2 learners’ successes and failures in production of L1 and L2 vowels (e.g. Flege, 1992; Nowacka, 2010; Rojczyk, 2010a). Usually such projects show how difficult it is for L2 learners to separate “old” and “new” vowel categories. However, the situation becomes much more complicated when we think of third language (L3) production. While in the case of L2 segmental production the number of factors affecting L2 sounds is rather limited (either interference from learners’ L1 or some kind of L2 intralingual influence), in the case of L3 segmental production we may encounter L1®L3, L2®L3 or L3 intralingual interference. This makes separation of L3 sounds a much more complex process. The aim of this study is to examine whether speakers of L1 Polish, L2 English and L3 German are able to separate new, L3 vowel categories from their native and L2 categories. Being a part of a larger project, this time the focus is on German /oe/. This vowel was chosen since it is regarded as especially difficult for Polish learners of German and it is frequently substituted with some other sounds. A group of English philology (Polish-English-German translation and interpretation programme) students was chosen to participate in this project. They were advanced speakers of English who did not encounter any difficulties in communication with native speakers of this language and upperintermediate users of German. They had been taught both English and German pronunciation/practical phonetics during their studies at the University of Silesia. The subjects were asked to produce words containing analysed vowels, namely: P /u/, P /ɔ/, P /ɛ/, E /uː/, E /ɔː/ and G /oe/. All examined vowels were embedded in a /bVt/ context. The target /bVt/ words were then embedded in carrier sentences I said /bVt/ this time in English, Ich sag’ /bVt/ diesmal in German and Mówię /bVt/ teraz in Polish, in a non-final position. The sentences were presented to subjects on a computer screen and the produced chunks were stored in a notebook’s memory as .wav files ready for inspection. The Praat 5.3.12 speech-analysis software package (Boersma, 2001) was used to scroll through the audio files in order to locate an onset and offset of target vowels, measure the F1 and F2 frequencies and plot vowels on the plane. All analyses were also performed using Praat. The obtained results shed new light on L3 segmental production and L1 and L2 interference.

Dynamic localization in second language English and German

Marking new and given constituents requires speakers to use morphosyntactic and phonological cues within a discourse context. The current study uses a dynamic localization paradigm whereby German and English native speakers, with the other language as a second language (L2), describe constellations of pictures. In each picture a new or reintroduced animal is localized relative to other animals, thereby allowing for control of newness vs. givenness of animals. Participants completed the task in their native language (L1) and L2. English native speakers use predominantly canonical word order and often mark the new object with a falling pitch accent. German native speakers use a given-before-new word order, even when this is non-canonical, and they use a rising pitch accent in non-final position. The results indicate that speakers easily transfer unmarked grammatical structures – both word order and pitch accents – from their L1 to their L2.

The weak vowels of South African English: A critical review and comparative acoustic analysis

This article firstly provides a thorough critical review of the existing literature on the status of the weak vowels of South African English (SAfE), both within the context of developments within SAfE as well as within the context of other relevant accents of English. This review of the literature is accompanied by the presentation of new acoustic data, specifically in the form of a comparative acoustic analysis of the General SAfE lettER, commA and happY vowels, as well as weak-vowels in non-final position. This analysis draws on contemporary data from both SAfE and comparable data collected by Fleming and Johnson (2007) for American varieties. The results of the combined review and acoustic analysis indicate that happY has the same phonetic quality as FLEECE (via happY-Tensing); that the Weak Vowel Merger is characteristic of SAfE (although at a higher-than-schwa quality); and that the prototypical value for commA and lettER in SAfE is ‘low schwa’, i.e., [3]. The article ends with some speculation concerning the possible historical origins of these various features of SAfE.

The prosody of focus in the Spanish of Quechua-Spanish bilinguals: A case study on noun phrases

This study examines the prosody of focus in the Spanish of 16 Quechua-Spanish bilinguals near Cusco, Peru. Data come from a dialogue game that involved noun phrases consisting of a noun and an adjective. The questions in the game elicited broad focus, contrastive focus on the noun (non-final position) and contrastive focus on the adjective (final position). The phonetic analysis in Praat included peak alignment, peak height, local range and duration of the stressed syllable and word. The study revealed that Cusco Spanish differs from other Spanish varieties. In other Spanish varieties, contrastive focus is marked by early peak alignment, whereas broad focus involves a late peak on the non-final word. Furthermore, in other Spanish varieties contrastive focus is indicated by a higher F0 maximum, a wider local range, post-focal pitch reduction and a longer duration of the stressed syllable/word. For Cusco Spanish no phonological contrast between early and late peak alignment was found. However, peak alignment on the adjective in contrastive focus was significantly earlier than in the two other contexts. For women, similar results were found for the noun in contrastive focus. An additional prominence-lending feature marking contrastive focus concerned duration of the final word. Furthermore, the results revealed a higher F0 maximum for broad focus than for contrastive focus. The findings suggest a prosodic change, which is possibly due to contact with Quechua. The study contributes to research on information structure, prosody and contact-induced language change.

Acquisition of articulatory control for sentential focus in children

The present study examined articulator movement changes associated with the production of sentential focus in 4, 7 and 11-year-old children and adults. A facial capture system was used to track jaw, lower lip and upper lip movements while words were produced with and without focus in phrase-final and non-final positions. Adults produced focused words with significantly greater movement duration, displacement and velocity than unfocused words. Significant differences in duration, displacement and velocity between focused and unfocused productions were seen in the 7 and 11 year olds, although these results were not as consistent as those seen in the adults. Articulator movement differences between focused and unfocused words were seen in both the phrase-final and non-final positions in adults yet only in the phrase-final position in children. Findings suggest that children begin to modify articulator movement to mark sentential focus between 7 and 11 years of age and are likely to continue to refine movement patterns for this prosodic task throughout adolescence.

The nature and use of Icelandic prenuclear and nuclear pitch accents: Evidence from F0 alignment and syllable/segment duration

Two production studies and one perception study were designed to systematically test F0 alignment and segment duration in Icelandic pitch accents with a view to investigating previous claims about the inventory of distinct intonational categories. Four different conditions were tested: (i) prenuclear pitch accents, (ii) nuclear accents in sentence-final position in sentences with either broad focus or (iii) with final narrow focus, and (iv) nuclear narrow focus accents in non-final position. The alignment results are such that (i) prenuclear accents are signalled by a late rise (L*H), while final nuclear accents are signalled by an early rise; (ii) F0 peaks in prefinal nuclear accents are aligned earlier than in prenuclear accents, but later than in final nuclear accents, suggesting a prosodic boundary effect. The duration measurements suggest a positional, but no focus, effect on the duration of the accented syllable and its vowel, such that syllables/vowels earlier in the sentence are longer than later ones.

Syllable structure and extrametricality

This paper proposes a functional basis for final consonant extrametricality, the asymmetric status of CVC syllables as stress-attracting in non-final position of a word but stress-rejecting in final position. A typological study of phonemic vowel length pattern in 10 languages with this final vs. non-final stress asymmetry and 30 languages in which CVC attracts stress in final position indicates a robust asymmetry between languages differing in their stress system’s treatment of final CVC. Languages that asymmetrically allow stress on non-final but not on final CVC all lack phonemic vowel length contrast in final position, whereas those lacking the stress asymmetry often have contrastive length in final vowels. It is claimed that the absence of phonemic length in languages that do not stress final CVC facilitates the nearly universal pattern of phonetic final lengthening, which threatens to obscure the perception of phonemic length. The enhanced lengthening of final vowels in languages with final phonemic vowel length reduces the duration ratio of CVC relative to CV, thereby reducing CVC’s perceptual prominence and thus its propensity to attract stress in keeping with Lunden’s (2006) proportional duration theory of weight. A phonetic study of two languages differing in the stress-attracting ability of final CVC offers support for the proposed account. Arabic, which displays consonant extrametricality and largely lacks phonemic vowel length in final position, has substantial final vowel lengthening, whereas Kabardian, which stresses final CVC and contrasts vowel length in final position, lacks substantial final lengthening.

Reconsidering the Role of Syntactic "Heaviness" in Old English Split Coordination

Reconsidering the Role of Syntactic "Heaviness" in Old English Split Coordination The splitting of coordinate structures in Old English has traditionally been attributed to structural size or "heaviness", assuming that long, complex coordinate constructions required increased parsing and processing effort. In addition to this, most texts on Old English syntax take for granted that the split elements always appear in clause-final position. The conclusions in this paper - drawn from the analysis of a large corpus of Old English texts - imply a radical revision of these assumptions. They suggest that the role of syntactic heaviness should be reconsidered, and its importance minimised in favour of other considerations of a pragmatic and discoursive nature. The analysis of the position of the split elements confirms that they appear in non-final position much more often than has been assumed, producing syntactic discontinuity.

Production characteristics of accent in four dialects of Nahuatl

Most descriptions of Nahuatl (Uto‐Aztecan, Mexico) describe regular penultimate accent, although production characteristics are unstudied. Recently, however, dialects from Guerrero state have been described with pitch‐accent from historical coda breathy /h/ in nonfinal position [Guion and Amith, J. Acoust. Soc. Am. 117, S2490 (2005)]. In the current study, the acoustic correlates of this new pitch‐accent, as well as penultimate accent, are investigated. Speakers (n=6) from four dialects–two with innovated pitch‐accent and two with only penultimate accent–were recorded. The production characteristics of fundamental frequency (F0), duration, and voice quality, as indexed by the difference between the first two harmonics (H1‐H2), were investigated. Innovated pitch‐accent was found to be cued solely by F0. Penultimate accent varied in its production by dialect. In dialects with only penultimate accent, production exhibited characteristics typical of stress‐accent: Accented syllables had higher F0, longer durations, and more vocal effort (smaller H1‐H2 differences) than unaccented syllables. Dialects with innovated pitch‐accent had less stresslike penultimate accents, being cued primarily by F0, or with a relatively small durational difference, but no voice quality difference. This indicates that the innovation of a pitch‐accent has changed the formerly stresslike penultimate accent into a more pitch‐type accent.

Difficulties in the control of irrelevant visuospatial information in children with visuospatial learning disabilities

This research tested the hypothesis that children’s difficulties in visuospatial working memory (VSWM) may mirror difficulties found with verbal working memory tasks in other categories of children. Two experiments compared the number of correct responses and errors in groups of visuospatial learning disabled children (VSLD) and Controls who were engaged in an active task testing visuospatial working memory. Children were presented with sequences of positions on a 4×4 matrix and were subsequently asked to remember only the last position of each series. In the first Experiment, VSLD children showed greater difficulty in both recalling the last positions and avoiding the irrelevant non-final positions compared with Controls. In the second experiment children of different age groups (second-graders and fifth-graders) were also required to stress, by tapping on the table, the irrelevant positions whenever the experimenter pointed to a coloured cell. Results showed that the number of errors was greater in the VSLD children, and the pattern of errors differed with their grade. In particular, the increased activation of stressed locations produced an increase of correct responses, and a decrease of intrusion errors, except in the case of VSLD second-graders, who made a higher number of intrusions for stressed than for unstressed locations.Results confirm that children with VSLD show a specific deficit in active VSWM, and in particular, in the ability to avoid intrusion errors. In general, the control of irrelevant information appears critical for a successful use of VSWM.

Final lowering in non-final position

This article argues that a factor contributing to final lowering in intonation is that H tones preceding downstep are scaled higher than H tones not preceding downstep, all else being equal. Final lowering is argued to be (in part) the absence of this effect with the last element of a sequence of downstepped tones. The evidence comes from intonation data recorded with speakers of German from Southern Germany and Austria. In two structures, a sequence of downstepped prenuclear pitch accents PA 1!PA 2…!PA n−1 is continued with a nuclear pitch accent PA n that is not downstepped. In both structures, final lowering is found on the last downstepped pitch accent PA n−1 , crucially in penultimate position of the intonation phrase. Here potential triggers for final lowering in the environment are not present (phrase-final position) or present only in one structure (a following L% boundary tone). The application of final lowering in these cases shows that there is a factor contributing to final lowering that affects accentual tones not followed by downstep, regardless of the environment.

Temporal characteristics of Chinese-accented English: Preliminary findings

This study investigated several temporal features of English to determine the extent of their occurrence in the speech of talkers of Chinese-accented English who had relatively limited experience with spoken English. Specifically, the extent to which these speakers produced the following temporal contrasts was examined: (1) tense versus lax vowel duration, (2) vowel duration before voiced versus voiceless consonants, and (3) vowel and consonant duration in sentence-final versus nonfinal position. Preliminary data from sentences produced by eight non-native and eight native talkers indicates that the native English speakers and the Chinese-accented talkers did not differ in the extent to which they realized the inherent duration difference between tense and lax vowels. However, the native English subjects tended to show substantially greater vowel lengthening before voiced versus before voiceless consonants than the Chinese-accented talkers. In addition, while the two groups did not differ significantly in the extent to which they lengthened sentence-final consonants relative to nonfinal consonants, the native English talkers showed greater sentence-final vowel lengthening than the Chinese-accented talkers. When group differences for a given temporal parameter were found, 1 or 2 of the non-native subjects typically fell within the range of performance shown by the native speakers.

Adaptations in spoken corrections: Implications for models of conversational speech

Miscommunication in spoken human–computer interaction is unavoidable. Ironically, the user's attempts to repair these miscommunications are even more likely to result in recognition failures, leading to frustrating error “spirals”. In this paper we investigate users' adaptations to recognition errors made by a spoken language system and the impact of these adaptations on models for speech recognition. In analyzing over 300 pairs of original and repeat correction utterances, matched on speaker and lexical content, we found overall increases in utterance and pause duration from original to correction. Here we focus on those adaptations – phonological and durational – that are most likely to adversely impact the accuracy of speech recognizers. We identify several phonological shifts from conversational to clear speech style. We determine that the observed durations of spoken user corrections from a field trial represent increases over, and divergences from, those derived from a speech recognizer's underlying model. Furthermore, words in final position diverge significantly more than those in non-final position, due to the additional effects of phrase-final lengthening. These systematic changes argue for a general model of pronunciation and duration, extending beyond the sentence level to incorporate higher-level dialog features, and illustrate important features for such a model to capture.

Production constraints on utterance-final consonant characteristics in babbling.

In order to evaluate hypotheses regarding production constraints on final consonants in babbling, 721 utterance-final consonants produced by 6 infants in consonant-vowel-consonant (CVC) syllables were examined and compared with the preceding consonant in the CVC. Consistent with earlier studies, major patterns were observed for each of the three main consonantal properties--place and manner of articulation and voicing. These patterns included a strong tendency for final consonants to repeat the place of articulation of nonfinal consonants and a tendency for relatively more fricative, nasal and voiceless consonants to occur in final position than in nonfinal position. The high frequency with which final consonants shared place of articulation with the preceding consonant was considered to reflect 'frame dominance' or the tendency of a relatively constant mandibular cycle (the frame) to determine the structure of utterances with very little contribution from other active articulators. The manner and voicing effects were attributed to an overall terminal energy decrease in the vocal production system.