LLT JOURNAL VOL. 15, NO. 2 ISSN 1410-7201 1 Voice Onset Time (VOT) and Vowel Duration YohanaVeniranda English Language Education Study Program Sanata Dharma University veniranda02@yahoo.com Abstract This study investigates the effects of voicing of a preceding and following plosive on the Voice Onset Time (VOT) and vowel duration. The data consist of words with CVC segments, with the four groups of samples that represent both voiced and voiceless plosives, both for the ones following and preceding the vowels. The preliminary hypothesis is that voicing of following plosives affects the length of the vowels, indicated by the vowel duration, and VOT is affected by the voicing of the preceding plosives, not the following plosives. To compare the effects of voicing on the lengths of the vowel duration and VOT, the sound files are analyzed using the Praat program. The results of this study show that the voicing of the following plosives affects the vowel duration, while the voicing of the preceding plosives does not affect the vowel duration. The results also show that it is the voicing of the C 1 plosives that affect the VOT. The word-ending plosives do not affect the VOT of the preceding plosives.This study is on English vowel duration and VOT affected by the following and preceding stops. A further study can be done to compare the results of the effects with another language. Keywords: VOT, vowel duration, the Praat program. INTRODUCTION This study investigates the effects of voicing of a preceding and following plosive on the Voice Onset Time (VOT) and vowel duration. The data consist of words with CVC segments, with the four groups of samples that represent both voiced and voiceless plosives, both for the ones following and preceding the vowels. The preliminary hypothesis is that voicing of following plosives affects the length of the vowels, indicated by the vowel duration, and VOT is affected by the voicing of the preceding plosives, not the following plosives. To compare the effects of voicing on the lengths of the vowel duration and VOT, the sound files are analyzed using the Praat program. A. METHODS a. The Data The data consist of the following sound files: 1) Four sound files, each file consists of 8 test words. The test words are in the carrier sentence “The word is….” There is one token of each word. The words, which have the segments of CVC, have been grouped in the following categories. Here are the words in this study: a) C[-voice] V C[-voice]: pat, pot, cot, putt, talk, puck, tuck, pick. b) C[-voice] V C[+voice]: pad, pod, cod, pud, tog, pug, tug, pig. c) C[+voice] V C [-voice]: bat, bought, got, but, dock, buck, duck, bic. d) C[+voice] V C [+voice]: bad, bod, god, bud, bug, dug, big. 2 2) Four sound files, each file consists of one test word. There are 10 tokens of each word. The test word is in the carrier sentence “The word is….” In these four sound files, there are four test words that represent each of the 4 CVC categories, and the vowel segment is the same. In this study, the words aretuck, tug, duck, dug. The recording is done using the sound recorder in the Praat program, recorded as mono sound, and the sampling frequency is set at 8,000 Hz. The sound files are saved as WAV files. b. The Analysis The calculation of the means and standard deviation of the vowel duration and VOT is done by grouping the consonants that follow the vowels into voiced and voiceless and that precede the vowels into voiced and voiceless. In the analysis, the abbreviation of C1 refers to the first plosive in the CVC words. Also, C2 refers to the second plosives in the words. T-tests, with the tails 2 and type 2, are performed to see the significance of the difference, using the standard p 0.05.The VOT and vowel duration obtained from the Praat program are converted to milliseconds. B. RESULTS AND DISCUSSION a. The results of the first 4 sound files, 8 test words in each file, one token. 1) The means and Standard Deviations The following tables show the means and the standard deviationsof the Vowel Duration and VOTof each set in milliseconds. Table 1. Mean and Standard Deviation of the Vowel Duration Followed by Voiceless and Voiced Plosives C V C [-voice] C V C [+voice] Mean 129.05 193.85 Standard Deviation 37.21 34.42 Table 2. Mean and Standard Deviation of the Vowel Duration Preceded by Voiceless and Voiced Plosives C[-voice] V C C[+voice] V C Mean 153.22 169.67 Standard Deviation 55.53 39.76 The results show that the voicing of the following plosives affects the vowel duration, while the voicing of the preceding plosives does not affect the vowel duration. This is confirmed by the t-test that show the difference between the means of the following plosives that are different in voicing is 1.69 E-5 (which means 1.69 X 10-5), so it is < 0.05. While the t-test of the vowel duration between the groups of different voicing of the preceding plosives shows that it is 0.34, which is not <0.05, so it is insignificant. The following tables show the results of the means and standard deviation of the VOT, of different voicing of the following and preceding plosives. Voice Onset Time (VOT) and Vowel Duration LLT JOURNAL VOL. 15, NO. 2 ISSN 1410-7201 3 Table 3. Mean and Standard Deviation of the VOT of C2 (the Word Ending) by Voiceless and Voiced Plosives C V C [-voice] C V C [+voice] Mean 53.76 47.51 Standard Deviation 43.56 40.80 Table 4 Mean and Standard Deviation of the C1 VOT by Voiceless and Voiced Plosives C[-voice] V C C[+voice] V C Mean 89.56 11.71 Standard Deviation 17.75 5.91 The results show that it is the voicing of the C1 plosives that affect the VOT. The following plosives do not affect the VOT of the preceding plosives. This result is also confirmed by the test of significance of the means. The t-test of the means of the different voicing of the preceding plosives (or the C1) is 1.066E-16 (or 1.066 X 10- 16), which is much smaller than 0.05, and it means the difference is significant. While the calculation of the t-test of the means of the VOT of the different voicing of the following plosives is 0.68, which is not smaller than 0.05, and it means it is not significant. b. The results of the first 4 sound files, 1 test word in each file, ten tokens. The following two tables show the results of the data on the vowel durations for plosives of different voicing, both following and preceding them. Table 5. Mean and Standard Deviation of the Vowel Duration Followed by Voiceless and Voiced Plosives C V C [-voice] C V C [+voice] Mean 96.08 211.40 Standard Deviation 16.22 34.10 Table 6. Mean and Standard Deviation of the Vowel Duration Preceded by Voiceless and Voiced Plosives C[-voice] V C C[+voice] V C Mean 151.8 152.75 Standard Deviation 61.17 68.48 The results in these sets of sound files confirm the previous results. The voicing the following plosives affect significantly the vowel duration, while the voicing of the preceding plosives does not affect it. The result of the t-test shows that the different voicing of the following plosives is 5.74E- 16 (or 5.74 X 10-16), which is lower than the 0.05 p value. While the t-test of the preceding different voicing of plosives is 0.96, which is higher than 0.05 p value, and so it means it is not significant. 4 The results of the VOT measurement show that the voicing of the preceding plosives significantly affect the values. While the following voicing of the plosives do not show significant effect on the VOT values. Table 7 Mean and Standard Deviation of the VOT of C2 (the word ending) by Voiceless and Voiced Plosives C V C [-voice] C V C [+voice] Mean 34.80 37.45 Standard Deviation 24.62 18.03 Table 8 Mean and Standard Deviation of the C1 VOT by Voiceless and Voiced Plosives C[-voice] V C C[+voice] V C Mean 52.34 19.91 Standard Deviation 13.54 14.069 The t-test of the different groups of preceding voicing is 6.58E-09 (or 6.58 X 10 -9), which is smaller than the p value 0.05, and so it is significant. The t-test of the VOT values for different voicing of following plosives is 0.70, which is higher than 0.05, and it means it is insignificant. All the results above show that the preliminary hypothesis, i.e. that voicing of following plosives affects the length of the vowels, indicated by the vowel duration, and VOT is affected by the voicing of the preceding plosives, not the following plosives, is proven to be correct. Ohala (1997) in his paper compared the contemporary view of the relation between phonetics and phonology with earlier attitudes on the matter. Phonetics and phonology did not exist as separate disciplines in earlier centuries. Ohala looks at the relation between phonetics and phonology as the relation between the domains of the study, and in his opinion, phonology has to be seen as the discipline that tries to answer questions about spoken language by employing the methods, data, and theories of phonetics, as well as psychology, social sciences, history, ethology, etc.He also points out that a phonetic account of how natural sound patterns will make a convincing explanatory scenario. He mentions the common practice within phonetics of making a given measurement, such as vowel duration on multiple tokens, like the one in this lab, is evidence of the integration of phonetics and phonology and that phonology can benefit from phonetic studies. C. CONCLUSIONS The results of this study show that the voicing of the following plosives affects the vowel duration, while the voicing of the preceding plosives does not affect the vowel duration. The results also show that it is the voicing of the C1 plosives that affect the VOT. The word-ending plosives do not affect the VOT of the preceding plosives. The current study shows the standard deviations but does not discuss further the implication of the deviations. Further study can provide analysis on it.More sample words with not only voiced and voiceless plosives, but also voiced and voiceless fricatives and affricates can be taken to provide further evidence if voiced and voiceless obstruents, not only stops or plosives, have the same pattern as the results of this study. This study is on English vowel duration and VOT affected by the following and preceding stops. A further study can be done to compare the results of the effects with another language. Voice Onset Time (VOT) and Vowel Duration LLT JOURNAL VOL. 15, NO. 2 ISSN 1410-7201 5 REFERENCES Ohala, John J. 1997. “The Relation Between Phonetics and Phonology”. The Handbook of Phonetic Sciences. William J Hardcastle; John Laver (ed). Oxford, UK ; Malden, Mass. : Blackwell Publishers. Source of sound files: Class website: http://udel.edu/~heinz/ classes/2012/4-633/log.html 6 Appendix 1: The calculations of the second set of sound files words VOT in ms Vowel Length in ms tuck1 50.3 103.3 Mean tuck2 52.3 122.5 VOT -voice C1 52.338 tuck3 54.93 118.04 VOT +voice C1 19.914 tuck4 56.1 93.49 tuck5 51.7 100.9 Mean tuck6 52.91 68.29 Vowel Length tuck7 54.09 80.83 VL C2 –voice 96.081 tuck8 86.49 83.34 VL C2 +voice 211.4042105 tuck9 70.14 95.98 tuck10 41.22 103.98 Mean VOT -voice C2 34.8025 VOT +voice C2 37.4495 tug1 68.28 181.2 tug2 48.58 187.77 Standard Dev tug3 46.41 171.61 VOT -voice C1 13.54345155 tug4 45.3 171.62 VOT +voice C1 14.06898358 tug5 73.46 206.49 tug6 32.75 195.57 tug7 38.71 211.45 Standard Dev tug8 33.75 255.13 Vowel Length tug9 41.69 252.21 VL C2 -voice 16.22142829 tug10 47.65 232.3 VL C2 +voice 34.09629425 Mean duck1 24.55 110.7 VL C1-voice 151.8 duck2 8.57 81.46 VL C1 +voice 152.7526316 duck3 11.3 99 duck4 10.53 75.52 Stand Dev duck5 14.42 102.13 VOT C2 -voice 24.61633409 duck6 7.79 108.75 VOT C2 +voice 18.02869598 Voice Onset Time (VOT) and Vowel Duration LLT JOURNAL VOL. 15, NO. 2 ISSN 1410-7201 7 duck7 14.03 117.72 duck8 15.98 98.61 Stand Dev duck9 9.35 91.99 VL C1 -voice 61.16815004 duck10 9.35 65.09 VL C1 +voice 68.48023031 T Test VOT (Voicing) 6.58548E-09 significant dug1 9.7 139.11 preceded by +-voice dug2 54.9 194.11 T Test Vowel length 5.73803E-16 significant dug3 49.33 207.-5 followed by +-voice dug4 37.2 208.67 T Test VOT (Voicing) dug5 29.11 253.96 followed by +-voice 0.700205483 insignificant dug6 8.89 200.58 T Test Vowel Length dug7 21.02 234.55 preceded by +-voice 0.963661394 insignificant dug8 34.77 214.33 dug9 12.94 236.69 dug10 14.55 269.33 8 Appendix 2: Samples of the measurement of the VOT and Vowel Duration tuck tug <---------Vowel Duration---------> duck dug <-------Vowel Duration------> Time (s) 12.68 13.32 -0.1074 0.1325 0 Time (s) 12.68 13.32 -0.1074 0.1325 0 Time (s) 1.322 1.892 -0.1852 0.143 0 Time (s) 1.322 1.892 -0.1852 0.143 0 Time (s) 1.322 1.892 -0.1852 0.143 0 Time (s) 1.322 1.892 -0.1852 0.143 0 Time (s) 1.286 2.022 -0.3008 0.1321 0 Time (s) 1.325 2.056 -0.1863 0.1504 0 Time (s) 1.325 2.056 -0.1863 0.1504 0 Voice Onset Time (VOT) and Vowel Duration Cover Vol 15 No2_okt 12_rep Isi LLT_Vol_15_No 2_Oct 2012_save as