The Western dialect of Naxi has four lexical tones: High, Mid, Low and Rising; the latter is rare in the lexicon. Rising contours on monosyllables are frequent in connected speech, however, as a result of a process of syllable reduction: reduction of a morpheme carrying the High tone results in re-association of its tone to the syllable that precedes it in the sentence, creating a rising contour. An experiment (with one speaker and five listeners) establishes that there is not only one rising contour that originates in tonal reassociation, as reported in earlier descriptions, but two: Low-to-High and Mid-to-High – as could be expected by analogy with phenomena observed in Niger-Congo languages and elsewhere. A second set of experiments (same speaker; six listeners) investigates the reduction of Mid- and Low-tone syllables: they reduce to [] and [], respectively, and coalesce with the preceding syllable (in Naxi, syllabic structure is simply consonant + glide + vowel). Unlike High-tone syllable reduction, this process stops short of complete tonal de-linking. These experiments aim to provide a complete picture of syllable reduction patterns in Naxi. It is argued that the notions of floating tones and tonal reassociation can be usefully applied to the Naxi data.

Tashlhiyt Berber contrasts singletons and geminates in intervocalic as well as in initial and final positions. This study presents results of an investigation of the phonetic correlates that distinguish these types of segments in these three positions. A claim has been made that Tashlhiyt Berber geminates are simple segments that are distinguished from their singleton counterparts primarily by tenseness. This proposal is not supported by our data. The examination of ten temporal and non-temporal parameters shows that consonant duration is the most systematic correlate across all types of consonants and positions, even for voiceless stops after pause. Preceding vowel duration and, to a lesser degree, amplitude of the release were also found to be significantly affected by gemination. These additional characteristics, interpreted as enhancing correlates, may be present in order to increase the perceptual distance between the two phonemic categories.

Previous studies of the Arabic pharyngeal ‘ayn have reported stop, fricative and approximant realisations within and across dialects. This paper presents an approximant variant that has not been specifically identified before. Described as a ‘tight approximant’, it is characterised acoustically by a pattern of filtering in which the lowest half dozen or so harmonics, including the fundamental, are markedly reduced in amplitude. Auditory responses to this kind of spectrum are characterised in terms of residue pitch. The variant can be thought of as compressed or ‘squeezed’ in the articulatory, acoustic and auditory domains in a chain of phonetic cause and effect. Laryngopharyngeal tension creates a low F0 and a compressed bandpass-filtered spectrum, which is processed auditorily by adjacent harmonics being squeezed through the same auditory filter. Acoustic and laryngographic data from selected tokens of a corpus of 436 realisations of ‘ayn, produced by 21 speakers from 11 different North African and Middle Eastern countries, are presented in support of the description of this variant. Its frequency of occurrence in the corpus suggests it is a common variant in Arabic speech. Some suggestions concerning the diachrony of ‘ayn are made in light of the reported observations.

The present study is an electropalatographic and acoustic investigation of the fricatives /s, f/ and the affricates /ts, dz, tf, d[zcy ]/ based on data from five speakers of Majorcan Catalan and five speakers of Valencian Catalan. Results show that the articulatory characteristics of fricatives and affricates agree in several respects: the sounds traditionally labeled /f/ and /tf, d[zcy ]/ are alveolopalatal, and are articulated at a less anterior location, are less constricted and show more dorsopalatal contact than the alveolars /s/ and /ts, dz/; the two place categories are closer to each other in Valencian than in Majorcan. Compared to voiceless affricates, voiced affricates are more anterior and more constricted, and show less dorsopalatal contact. Data also show that closure location for /tf, d[zcy ]/ occurs at the alveolar zone, and that articulatory differences among affricates are better specified at frication than at closure. Strict homorganicity between the stop and frication components of affricates appears to hold provided that constriction location at frication is compared with place of articulation at closure offset. In comparison to voiceless affricates, voiced affricates were shorter, and exhibited a longer closure and a shorter frication period, in Majorcan; in Valencian, on the other hand, closures were shortest for /d[zcy ]/, and frication was systematically longer for voiceless vs. voiced affricates. These duration data appear to conform to a universal trend in Valencian but not in Majorcan where voiced affricates are lengthened intentionally. In both Catalan dialects, vowel duration varies inversely with the duration of the affricate and of its closure and frication components. The implications of these articulatory and duration characteristics for the interpretation of sound changes affecting affricates, i.e. place merging, lenition and devoicing, are discussed.

This paper examines the acoustic characteristics of the Cibaeño Dominican Spanish ‘pre-aspirated’ trill. The data are based on narrative speech of ten native speakers from the Cibao region. The findings reveal that the segment is most often produced as a period of pre-breathy voice followed by a single tap, [hr]. The second most frequently occurring production involves pre-breathy voice and a trill. The term pre-breathy-voiced vibrant is proposed as a more appropriate characterization of the Cibaeño Dominican Spanish trill.

Auditory and acoustic data were produced from recordings of a Glaswegian English speaker in conversational and reading modes. Clearly different intonational systems were used in the two modes. The reading style used an intonation similar to that used in standard British intonation (the intonation of ‘Received Pronunciation’ (RPI)). The conversational style was an example of the type of intonation used in a number of cities in the north of the UK (Urban North British Intonation (UNBI)), characterised by a default intonation involving rising or rising-slumping nuclear pitch patterns. This speaker illustrates a clear-cut case of intonational diglossia with a falling default tune in the one mode and a rising(-falling) default tune in the other.

Three American English diphthongs involving onsets in the back of the vowel space were investigated by means of acoustic measurements, discriminant analysis, and perceptual experiments. These diphthongs, termed DOOR/WAR, TOY, and STAR, were observed in the speech of ten young women residing in southern Michigan. The diphthongs were elicited in a variety of /b/-, /ɡ/-, and /h/-initial words (e.g. board, Boyd, and barred). Words with similar consonantal margins but different nuclei (e.g. bode, baud, and bod) were elicited, as well. The DOOR/WAR and TOY words were compared with bode and baud (and similar words with initial /ɡ/ and /h/), while the STAR words were compared with baud and bod (and similar words with initial /ɡ/ and /h/). All three methods of investigation showed that the initial part of the DOOR/WAR and TOY diphthongs more closely resembles that of bode than baud. On the other hand, the initial part of the STAR diphthong was shown to be intermediate between that of baud and that of bod.

Komi-Permyak (Finno-Ugric) has a typologically rare contrast between palatal stops and palatal affricates; the latter are also contrastive with postalveolar affricates. In addition, the language has a relatively uncommon three-way place contrast among sibilant fricatives. This paper investigates acoustic properties of the Komi-Permyak voiceless coronal obstruents in different positional contexts. An acoustic analysis of data from six native speakers of the language revealed that the contrasts were well differentiated by a combination of duration, noise intensity, spectral shapes, and vowel transitions – consistently with previous studies of similar consonants in other languages. The results also showed some positional differences in the realization of the consonants. Overall, the study contributes to a better understanding of phonetic implementation of marked phonological contrasts and positional variation; it also provides an acoustic record of the endangered language.

Different types of non-pulmonic sound production found in read and spontaneous German are described and exemplified. Stop releases driven by a glottalic airstream are found in sequences of final plosive plus glottalized vowel onset. Both ingressive and egressive velaric stop releases are considered to arise from double occlusions involving partial articulatory overlap in sequences of dorsal stop followed by an apical or bilabial stop. Certain aspects of the articulatory and aerodynamic mechanisms involved remain unclear. However, it is argued that an identification of these patterns in the acoustic record alone provides an invaluable insight into aspects of articulatory synchronization in types of spoken language, such as conversation, which are not generally amenable to the more intrusive methods of articulatory analysis.

The present study examines evidence for change in real time within the short vowel subsystem of the RP accent of English over the course of the twentieth century. It compares plots of average formant positions for the short vowels, stemming from several data corpora. It furthermore describes a change over time in the juxtaposition of the trap and strut vowels as captured in the calculated angle and distance between the two, using trap as a fixed point. This representation of a relationship in a single measurement by means of angle calculation is a methodological innovation for the sociophonetic enterprise. A value specifying the geometric relationship between two vowel positions is precise and replicable, as well as abstract enough to be comparable across data sets. Differences between ‘phonetic’ and ‘sociolinguistic’ stances on the interpretation of acoustic vowel data in formant plots and the issue of suitable vowel normalisation procedures for sociophonetics will also be discussed.

In this paper we present a production study designed to explore the relationship between three observations which have previously been made about liquids in British English: first, that laterals have prosodically-determined ‘clear’ (syllable-initial) and ‘dark’ (syllable-final) variants; second, that some varieties of English have either clear [1] in all positions or dark [l] in all positions; third, that some varieties with clear [1] have dark [r] while some varieties with dark [1] have clear [r] (in broad phonetic transcription). We take F2 as an acoustic correlate of clearness/darkness and report on F2 variation in two representative varieties of British English, one which has clear initial [1] (Newcastle-upon-Tyne) and one with dark initial [1] (Leeds). We show that Newcastle English has higher F2 frequencies in [1] than in [r] and that the reverse pattern is found in Leeds English. These patterns can also be found in adjacent unstressed vowels but not in adjacent stressed vowels. Final [1] in both varieties has a lower F2 than initial [1]. In intervocalic contexts, these F2 distinctions in the liquids are observed in iambic words for both varieties. In trochaic words they are observed for Leeds only, though the vowel effects can be observed in both varieties. We discuss some phonological consequences of these findings.

Beginning phonetics students are taught that some varieties of American English have two contrasting reduced vowels, transcribed as [e] and [i], illustrated by the unstressed vowels in the minimal pair Rosa's vs. roses (e.g. Ladefoged 2001, 2005). However, little seems to be known about the precise nature or distribution of these vowels. This study explores these questions through acoustic analysis of reduced vowels in the speech of nine American English speakers. The results show that there is a fundamental distinction between the mid central [e] vowel that can occur in unstressed word-final position (e.g. in Rosa), and high reduced vowels that occur in most other unstressed positions, and might be transcribed as [i]. The contrast between pairs like Rosa's and roses derives from this difference because the word-final [e] is preserved when an inflectional suffix is added, so the schwa of Rosa's is similar to the final vowel of Rosa, whereas the unstressed vowel of roses is the high [i] reduced vowel quality found elsewhere. So the standard transcription of the reduced vowel contrast is justified, but the widespread use of [e] to transcribe word-internal reduced vowels is misleading – mid reduced vowels are generally only found in stem-final position.

In this article, I will argue for the use of the epsilon symbol in the lexical dress set (which includes words like step, ready, said, shelf etc.) for RP. The need for this paper arises from the fact that many, but by no means all, dictionaries and linguistic treatises employ the [ē] symbol and that this symbol is neither the most accurate nor a particularly useful one, especially for foreign learners of English. An examination of current usage and its historical rationale (or lack thereof) is followed by articulatory and perceptual evidence for the dress vowel being close to the third cardinal vowel, more practical arguments, and a discussion of the issues raised.

In this illustration, we present a transcription and discussion of the variety of English spoken by people of European descent in New Zealand. Locally, such people are referred to as being Pākehā, using the Māori word. Pākehā speech is illustrated here in the speech of three educated women. Thus some of the features which might be expected from New Zealanders do not occur in these recordings, although the speakers are all clearly New Zealanders.

Chistabino is the variety of Aragonese, a Romance dialect descended from Latin, like Spanish, still spoken in the Valley of Gistau, bordering on France and formed by the River Cinqueta, in the central part of the Spanish Pyrenees (Province of Huesca). Although it is spoken mainly in the village of Gistaín (c. 240 inhabitants; c. 1400 m above sea-level, near latitude 42°35′), and is fairly well preserved by the older generations there, fluent speakers can still be found also in the nearby locality of San Juan. The other inhabited villages in the valley (Plan, Serveto, Sin and Saravillo) have suffered much greater degrees of erosion by Castilian (Standard Spanish), which has increased its influence through much improved communications, leading to the development of tourism in the area and emigration by the young in search of greater prosperity. The remaining village, Señes, has been uninhabited since 1970.