4.3 Results

The results of the auditory and the acoustic method differed. Using the auditory method, the speakers’ consistency was on average 80 %. Hence, the speakers show some evidence of their chosen variety of RP or GA when pronouncing the given words. But on the other hand, the range of the speakers’ individual degrees of consistency was very wide – between 16.7 % and 100 %. Three out of eleven speakers had a consistency of 100 %, which means that they were inconsistent in just one of the given words.

Despite their preferred variety, 15 out of 16 speakers pronounced the word vibrate with the American stress pattern. There was no word that was pronounced with the correct stress pattern by every speaker. Nine out of 16 speakers made a mistake in stress placement on the word miscellany. But it has to be considered that these are difficult words that the average learner rarely comes across – at least in their spoken form – during his/her studies.

Using the acoustic method, the speakers had an average consistency of 37.5 %, which is slightly less than that of the auditory method. But with the acoustic analysis, the speakers’ consistencies ranged from 0 % to 83.3 %. These differences between the auditory and the acoustic method were unexpected. The following analysis of the strengths and weaknesses of both methods offers some explanations for these differences.

The auditory method proved very easy to carry out. Besides, it was not very time-consuming since it took around 5 minutes to analyse the stress placement of the words of one speaker. On the other hand, it was not always easy to detect the stress placement – this was especially dependent on the speaker: did he/she speak clearly or too artificially? The problem is that speakers who read out a list of words are in an artificial situation. Therefore, they put much more effort in pronouncing the words clearly than they would usually do in spontaneous speech. One solution for that problem would be to embed the words in a sentence context. Despite the fact that the pronunciation was in some cases unclear, I classified the stress placement of every word as either RP or GA.

With the help of the acoustic method, the syllables were analysed regarding vowel duration, maximum intensity and maximum pitch. This method was much more time-consuming since it took around 40 minutes to analyse the words of one speaker. It is advantageous if the examiner has some experience with Praat since it takes a lot of time to define the boundaries of a vowel clearly.

The main problem of the analysis with Praat is how to evaluate the results. As Gut proposes, the combination of the three factors vowel duration, intensity and pitch height were used as indicators for stress placement. It proved that in the case of a stressed syllable not all three factors were higher compared to the unstressed syllable. The factor of vowel duration, however, was in most cases higher in stressed syllables. That is why a longer vowel duration was taken as the main predictor of a stressed syllable. Besides, a syllable was only classified as the stressed one if it had not only a longer vowel duration, but also a higher maximum intensity and/or a higher maximum pitch. Nevertheless, it was only possible in 66 % of the words to classify their stress placement as typically RP or GA. In 20 % of the words, a classification was impossible since the three factors did not differ a lot between the two syllables in question (see results in appendix). In the remaining 14 % of the words, a classification was made, but a question mark in brackets behind the variant (RP/GA) indicates that this classification is uncertain, since the differences between the two syllables in question were not high. The problem is that there is no existing set of rules that states how much stressed and unstressed syllables should differ in vowel duration, intensity and pitch height. One solution might be to investigate syllable duration and contrast it with vowel duration of this syllable because in the case of a stressed syllable not only the vowel but the whole syllable is longer.

Especially the evaluation of speaker 7 proved to be difficult and this speaker serves as a good example of clear speech. In contrast to spontaneous speech, speakers in artificial situations speak very clearly and distinctly. In this case, more than one syllable of a word appears to be prominent. The phenomenon of vowel deletion that appears frequently in spontaneous speech was avoided by many speakers. Regarding the stress placement of speaker 7 it can be seen that 4 out of 6 words could not be classified at all.

All in all, using the acoustic method the speakers’ consistency is in more than 60 % of the cases lower than when using the auditory method. In some cases, the difference of a speaker’s consistency is significantly high – as in speaker 7, where the consistency was 66.7 % with the auditory method, but only 16.7 % with the acoustic method. But the reason for these differences is the problematic classification of stress placement with the help of Praat, as was explained above.

Another finding is that the results strongly depend on the speaker. In this context, not only the contrast of clear and spontaneous speech is important but also the fact that some speakers tend to mumble. Furthermore, words were oftentimes mispronounced, which exacerbated the analysis.

Everything considered, the auditory method is very easy to execute, while the acoustic method is both time-consuming and especially for beginners difficult to comprehend. For a profound study, however, the acoustic method has priority since it is more reliable and valid. To bring out the best of both methods, the acoustic method could be used in order to verify the findings of the auditory method randomly. Besides, the acoustic method might be used in cases where it was not easy to detect stress placement with the auditory method.

Created with the Personal Edition of HelpNDoc: Free CHM Help documentation generator