Experimental phonetics. Changes in the location of vowels

Research Paper (postgraduate), 2017

26 Pages, Grade: B




Background of the Language (Akan)

The Vowel System of Akan

Statement of the Problem

Objectives of the Research

Related Studies



Discussion / Results



The positions of individual vowels in the acoustics vowels space often change over time in languages of which Asante-Twi a dialect of Akan is not an exception. Sound changes in languages are typically associated with social factors like age, gender, social class. This study will explore the changes in the location of vowels as produced by four generational of speakers of Asante-Twi speakers. Huber et al (1999) argues that the acoustic structure of vowels, as a consequence of articulation is influence by numerous factors such as the age, physical status or gender of the speaker. Fant (1966) also posits that since the length of the vocal track determines the overall patterns change, age and depend on gender. The current study will therefore investigate whether generational differences can affect vowel quality by comparing speakers of Asante-Twi with ages 6-14, 19-24, 40-50 and 60+..

Background of the Language (Akan)

Akan is a language spoken in Ghana, West Africa, of which Asante Twi is part. It belongs to the Kwa group of the Niger-Congo language family. It is spoken as a first language in six regions (Ashanti, Eastern, Central, Brong Ahafo, Western and Volta out of the ten (10) of Ghana.

The Akan people are found mainly in the southern half of Ghana and in neighbouring Cote d’Ivoire. Westermann and Bryan (1952) have used the name "Akan" to refer to the group of languages spoken in the geographical area referred to above, so that their use of the name “Akan" more or less coincides with the ethnographic use of the name. The Akan language is one of the most widely spoken languages of the Kwa sub branch of the Niger-Congo family, being the first language of approximately 49.1% of the population of Ghana (GSS 2002).It is also estimated that 44% of the population of Ghana use Akan as a second language(GSS, 2002). In terms of language use, Akan is virtually used in almost every sector of the Ghanaian society, from education to trade and commerce, politics, judiciary, arts, religion and the media and so on.

Dolphyne (2006) posits that the Akan language has several dialects and three of its dialects, Asante, Akuapem and Fante have achieved literacy status. The Asante and Akuapem dialects are called Twi. Each has a written form which reflects the peculiarities of the particular dialect, so that it is not easy for an Asante speaker who does not speak Fante to read a text written in Fante and vice versa, even though the two dialects are mutually intelligible. The dialect chosen for this study is Asante Twi.

The Vowel System of Akan

The Akan language is said to have fifteen vowel phonemes. It is made up of ten oral vowels and five nasalized vowel. Dolphyne (2006) argues that Twi dialect of Akan has ten oral vowels and Fante has nine. [i, u, o, a, e, ɛ, ɔ, ɪ, ʊ, æ ] are the oral vowels, whiles [i͂, ẽ, a͂, u͂] are the nasalized vowels.

illustration not visible in this excerpt

Akan vowel chart by Dolphyne (2006 ).

Statement of the Problem

Description of vowels qualities have been mostly impressionistic and based on traditional tongue description and this has led the use of similar symbols to represent different vowel sound in different languages. The description of Asante-Twi vowels has also been impressionistic. Some researchers have been done on Akan vowels qualities. Prominent among these include Ladefoged (1964) who worked on sixty-one West African languages including Asante-Twi, his focus was on vowel quality and vowel harmony. Lomotey (2008) also did a cross-dialectal study of Akan vowels using acoustic measurement. However, her research was not centered on how generational difference affects vowel quality although she recorded different ages for the study. This creates a vacuum which needs to be filled. It is upon this that the researcher seeks to delve into this topic to strategically and systematically come out with empirical data to address the issue of how generational difference affects vowel quality by comparing speakers of Asante-Twi with ages 6-14, 19-24, 40+ and 60+..

Objectives of the Research

The purpose of this study is to gain a better understanding of how generational difference affects vowel quality among the Asante- -Twi speakers .

Research Question

How does generational difference affects vowel quality among the Asante-Twi dialects of Akan?

Related Studies

Some acoustics phoneticians have carried out some research on how generational differences can affect vowel quality. Prominent among them include : Jacewicz and Fox (2012), who researched on the effects of cross-generational and cross-dialectal variation on vowel identification and classification, Auszmann and Neuberger (2014) also studied age-and gender-related differences in formant structure during the stabilization process of vowels, Mayr and Siddika (2016) also did a study Inter-generational transmission in a minority language setting: Stop consonant production by Bangladeshi heritage children and adults.

Dihingia and Choudhury (2016) did a preliminary study of cross-generational vowel variation in Barpetiya Assamese. The aim of their paper this study was to presents a socio-phonetic account of vowel variation across two age groups in the variety of Assamese spoken in Barpeta district. This study also attempted to present an account of variation across two age groups of the speakers of the Barpetiya dialect and to see if the standard variety has had any effect on the vowel sounds of this variety. This variety, also known as “Barpetiya Assamese” falls under the larger Western variety of Assamese and is regarded to be a distinct variety owing to its phonological, morphological and lexical variations from other varieties. This study focuses on the acoustic variation of Barpetiya Assamese vowels across two age groups: below 30 years and above 40 years. For the preliminary study, data was collected from two female native speakers from Barpeta as well as from Jorhat each from the age group below of 30 years and above 40 years. The Assamese vowels /i, e, ɛ, a, ɔ, o, ʊ, u/. The recordings were done in a noiseless environment using a Shure unidirectional head-worn microphone connected to a Tascam linear PCM recorder via xlr jack. For this particular study the words in sentence frame and produced and were further analyzed. The data recorded was segmented phoneme wise in Praat after which a Burg algorithm was used to calculate the first three formants (F1, F2, and F3) in Hertz at the midpoint of the steady state. The values were normalized with Lobanov method for speaker effects using NORM and plots of vowel group means with 1 standard deviation ellipses were generated on F1-F2 space. While the Jorhat variety showed merger of the two vowels /ʊ, u/ in both the age groups, the Barpeta variety showed merger of the vowels /ʊ, o/ in the age group of above 40 years and the vowels /ʊ, u/ in the age group of below 30 years. Additionally, the vowels /e/ and /ɛ/ also showed merger in both the age groups of the Barpetiya dialect. The results suggest that while the older generation has maintained the vowel system, a gradual change can be observed whereby the vowel system of the younger generation has been seen to be influenced by the standard variety. The current research will also examined how age difference affects the position of vowels on the vowels space as these researchers did, though this research will not compare how dialectal difference affects vowels position on vowel space unlike what these researchers did.

Grama (2015) presents an acoustic phonetic examination of the vowel systems of 32 Hawaiʻi Creole speakers with special attention paid to how these vowel realizations have changed across time, gender, phonological context, and the number of Hawaiʻi Creole morpho-syntactic features exhibited by speakers. This research was motivated by an interest in two questions in creole and variationist linguistics: how does Hawaiʻi Creole differ from its main lexifier language, English; and how has the language changed over time? To address these questions, vowel data was taken from existing sociolinguistic interviews archived in Kaipuleohone at the University of Hawaiʻi. The analyzed speakers come from two corpora conducted at different points in time: one conducted in the 1970s, and one conducted in the 2000s; 16 speakers from each corpus were analyzed, and these speakers were evenly distributed across age and gender. The first two formants and the duration of 11,191 vowels in fourteen vowel classes were analyzed from spontaneous speech produced during these interviews. Analysis revealed that the vowel spaces of speakers recorded in the 1970s vary significantly with respect to the vowel spaces of speakers recorded in the 2000s. 1970s speakers show substantial spectral overlap between high front vowels /i/ and /ɪ/, and overlap between the high back vowels /u/ and /ʊ/. 1970s speakers are also more likely to realize low vowels /a/ and /ʌ/ as spectrally overlapped and distinct from /ɔ/, which is realized as higher and backer in the vowel space. While each of these vowel classes exhibits significant spectral overlap, each is differentiated by vowel length for all age groups, suggesting that Hawaiʻi Creole (at least for speakers sampled in the 1970s) exhibits contrastive vowel length. By contrast, 2000s speakers realize /ɪ/ and /ʊ/ as distinct in spectral space from /i/ and /u/, respectively, and the low back vowels /a/ and /ʌ/ are less overlapping in spectral space for the youngest age group. 2000s speakers also realize /ɔ/ as fronter in comparison to older speakers. 2000s speakers also exhibit a number of other differences with respect to 1970s speakers, including lower and backer realizations of /æ/, fronter realizations of /e/ and /i/, fronter realizations of the high back vowels /u/ and /ʊ/, and higher realizations of the nucleus of /ai/. Despite the number of changes that manifest between 1970s speakers and 2000s speakers, few differences in vowel realizations arise across gender. Over time, only /a/ and the nucleus of /au/ raise for females but not males. Females also exhibit slightly lower variants of /ɪ/ and more similar realizations of /a/ and /ɔ/ than males. That relatively few differences arise across gender in Hawaiʻi Creole is noteworthy, especially since English (the main lexifier language for Hawaiʻi Creole and a language with which Hawaiʻi Creole is in heavy contact) exhibits many differences across gender in terms of vowel realizations. Taken together, these findings provide evidence that the vowel space of Hawaiʻi Creole speakers has changed substantially over time; many of these changes have caused Hawaiʻi Creole vowel spaces to approximate English vowel spaces. However, younger speakers of Hawaiʻi Creole who exhibit higher rates of Hawaiʻi Creole morpho-syntactic markers are more resistant to these changes. Together, findings from this study help characterize and describe the vowel system of Hawaiʻi Creole and how it has changed over time, as well as contributing to an understanding of how creoles interact at a structural level with their main lexifier language over time. The current research like the above mentioned research have certain similarities, this study will delve into and find out if position of vowels will differ as a result of generational difference.

Auszmann and Neuberger (2014) also conducted a research on age- and gender-related differences in formant structure during the stabilization process of vowels. The aim of the study was to investigate the effect of age and gender on vowel production in the speech of Hungarian children between the ages of 7 and 13. Eighty typically developing monolingual Hungarian-speaking children participated in this study. The analysis was cross-sectional and included four age groups: 7-, 9-, 11-, and 13-year-old children. The 7- and 9-year-olds were from the lower grades, while the 11- and 13-year-olds were from the upper grades of elementary schools in the capital city of Hungary. The reserchers compared their data with the data on preschoolers’and adults’ vowel structures from previous Hungarian literature (Gráczi and Horváth 2010; Gósy and Beke 2010; Bóna 2014). There were 20 children in each age group. None of them had any hearing disorders, and their intelligence fell within the normal range. All the participants were tested individually in a quiet room in their schools. The children talked about their family, school, and free time activities for various lengths of time. In order to compare the children’s data, the researchers analyzed a one-minute part from the middle of each recording. The data set contained approximately 15,000 tokens of nine manually measured Hungarian vowels. The recordings were annotated by one of the authors while the other one controlled and checked the annotations and the measurements were conducted using the Praat 5.3 software. All of the children had similar social and cultural backgrounds. The results confirmed that there were large individual differences in the vowels, irrespective of age and gender. However, there was evidence for maturation in F0 and vowel duration values across ages and gender. The vowels pronounced by 11-year old children are still different from those of adults (in terms of the parameters that were analyzed); however, they become more like those of adults by the age of 13.With age the duration of vowels becomes statistically shorter and vowel spaces get smaller. The most important gender related difference is that the F0 of boys is higher at the age of 7, becomes similar to girls’ at the age of 9 and 11, and then is lower by the age of 13. Though these researcher, researched on how age and gender affects vowels production, this current will be centered on only how age difference affects vowels production among Asante-Twi speakers of Akan.

Fox and Jacewicz (2016) have also conducted cross-dialectal and cross-generational changes in point-vowel locations. This study examined the changes in the location of four point vowels [i u ɑ æ] for three English dialects (North, Midland and Inland South) as produced by three generations of speakers (children, adults aged 35-50 and 65+ years). This research was an extension of Chung et al. (2012) who used a normalization approach that calculates distance and angular displacement from the speaker centroid. Eliminating the confounding effects of different vocal tract sizes on formant frequency values, this approach allowed for a comparison of the relative location of each of the three corner vowels in a speaker-specific vowel space by calculating distance and rotation of these vowels. These researchers posited that their study was an extension of the above mentioned methodology. They further argued that the speaker-specific centroid is determined on the basis of frequency values of all monophthongs and not only on the speaker’s corner vowels. This calculation provides a more stable characterization of the spectral center of the speaker’s vowel space. They used this approach to examine variation in the location of four corner vowels in three dialects and cross-generational variation related to sound change in progress. 10 American English monophthongs /i, ɪ, e, ε, æ, a, ɔ, o, u, ʊ/ were produced in h_d context in citation form using the prompts: heed, hid, heyd, head, had, hod, hawed, hoed, who’d, hood. 236 speakers participated, divided into 3 age groups: A4 (66+ yrs), A2 (51-65 yrs), and A0 (8-13 yrs). 81 speakers were born and raised in western North Carolina (NC): A4 (9 M, 9 F), A2 (16 M, 16 F), A0 (16 M, 15 F), 77 speakers were born and raised in central Ohio (OH): A4 (10 M, 11 F), A2 (9 M, 15 F), A0 (16 M, 16 F) ,78 speakers were born and raised in southeastern Wisconsin (WI): A4 (11 M, 9 F), A2 (14 M, 15 F), A0 (14 M, 15 F). Each speaker produced 3 repetitions of each token for a total of 7080 tokens. The first two formants of the four corner vowels /i, æ, a, u/ were measured at the vowel’s midpoint and checked using a series of Matlab programs. On the basis of these measurements, normalized and unnormalized plots were first created, followed by polar histograms for each age group and dialect. Finally, mean vectors from the centroid to each vowel were calculated and plotted on a polar coordinate system. The researchers found an anti-clockwise rotation of the corner vowels, particularly /æ/ and /a/, with each younger generation. Notably, this common trend occurs in all three different dialects studied here regardless of their dialect specific vowel dispersion. This rotation constitutes a new vowel development in North American English. This study will also examine how generational difference affects vowel production, however, the current research will focused on only one dialect (i.e. Asante-Twi) unlike the afore mentioned research who compare three dialects of English.


Excerpt out of 26 pages


Experimental phonetics. Changes in the location of vowels
University of Education
Catalog Number
ISBN (eBook)
ISBN (Book)
File size
1190 KB
experimental, changes
Quote paper
Job Anane (Author), 2017, Experimental phonetics. Changes in the location of vowels, Munich, GRIN Verlag, https://www.grin.com/document/431134


  • No comments yet.
Read the ebook
Title: Experimental phonetics. Changes in the location of vowels

Upload papers

Your term paper / thesis:

- Publication as eBook and book
- High royalties for the sales
- Completely free - with ISBN
- It only takes five minutes
- Every paper finds readers

Publish now - it's free