Skip to Content

Department of Linguistics


Oral-Nasal Airflow: Analysis

Click here to return to the Oral-Nasal Airflow Workshop home page

Click here to return to the results page

Analysis of Results

Your submission for this workshop:-

  • will probably be about 3-4 pages in length. Longer is not necessarily better and shorter is not necessarily worse. Concise well focused answers are preferred. You will be marked on your demonstrated understanding of the issues rather than the size of your submission.
  • should not include the tables and figures provided in these pages (simply refer to them by number if you feel the need). It is also unnecessary to include images from other sources in this report. That is, only submit text written by you.
  • does not need to be written up as a normal experimental report. In other words, you don't need to reiterate the details of the experimental procedure nor do you need to copy the details of the results into your submission.
  • should be structured in the following way. The sections 1-6 below are only included to provide context for the questions that you need to answer (and these section titles are not required in your assignment). You should answer the questions below (each preceded by a letter [A] to [N] ). Please keep each answer separate (this greatly improves ease of marking) and precede each answer with the question's letter.
  • DO NOT write an essay.
  • DO NOT merge the answers to two or more questions into a single answer.
  • YOU MUST answer each question separately. The answer to each question must start with the question letter.

1) Introduction

This assignment does not require an introduction (and any introduction will be ignored during marking). In marking this assignment I'm only interested in your answers to the questions [A] to [N] found scattered in the sections, below.

2) Pervasive nasality and Rate of Utterance.

The rainbow passage is (approximately) a phonetically balanced passage. This means that the frequencies of occurrence of each phoneme and of each phonetic class reflect their actual frequencies in normal continuous connected speech. This suggests that the rainbow passage can be used to examine the average behaviour of each subject in continuous speech. In this experiment it is assumed that this passage gives a reasonable measure of a subject's habitual or pervasive nasal airflow.

Table 1 on the results page summarises results from the rainbow passage and the "Naomi" passage. The following data are provided for each passage:-

  1. Rate of articulation
  2. Nasal airflow as a percentage of total airflow. It is well established that the best indirect method for determining velar opening relies on measurements of nasal airflow as a percentage of total airflow.

[A] Examine (for both the rainbow and the naomi passages) the relationship between nasal airflow and rate of articulation. Briefly summarise these results.

[B] What is the effect of using a passage that is phonetically unbalanced (in favour of nasal syllables) on our ability to accurately determine each person's pervasive nasality? Why does this difference occur? Which passage is better able to provide a reliable measure of pervasive nasality.

3) Pervasive (Airflow) Nasality and Rate of Utterance: Sex differences.

Refer to tables 1 and 4 on the results page.

[C] Examine your data for any significant sex differences. What conclusions can you draw from this data? Can you suggest any explanations for the results?

[D] How (if at all) are these results likely to relate to perceived nasality? Remember (from other units) that, typically, we perceive the unexpected more clearly than the expected.

4) Vowel Intrinsic Nasality.

Refer to tables 2 and 3 on the results page.

When a set of vowels are compared for their intrinsic "nasality" they must be contrasted in identical environments. These environments must not be affected by the presence of nearby nasal consonants which would cause (through coarticulation) contextual "nasality". These 4 sentences (i.e "ye bees ...", etc.) meet these requirements.

[E] Discuss both the muscular and the aerodynamic effects which would influence absolute nasal airflow and also nasal airflow as a percentage of total airflow. For example, is it possible for absolute nasal airflow to increase and yet for nasal airflow as a percentage of total airflow to decrease? Explain.

[F] With respect to "nasality" which is the most "natural" vowel phonetic group? You will need to examine especially the t-test results to answer this. You should note that the t-test results are not reported explicitly in these pages, but the relevant p values derived from these t-tests are reported at the bottom of the relevant tables on the results page. For the purposes of this assignment, the most "natural" groups are those groupings of vowels that significantly contrast with respect to some measure of nasality.

[G] The vowel /i:/ can belong to either the "high" group or the "non-high-back" group. With respect to "nasality" to which group does it most naturally belong?

[H] Why is /i:/ closer to /u:/ in terms of nasal airflow than it is to /o:/? You will need to know the acoustic and articulatory nature of Australian English high vowels to answer this question. It will do you no good to look up these vowels in American or British texts. See, in particular, this Australian English data and compare it with British monophthong and US monophthong patterns.

[I] Do any subjects display behaviour which would place /i:/ in a different phonetic class to that deduced from the results of the sample as a whole? This answer can't be based on statistical results as we're talking about individual behavior. Simply carry out an impressionistic analysis. Look at the individual results and see if you can find any individual differences from the group behavior. Which subjects seem to have different patterns? For example, do some subjects have a vowel that is more "nasal" than is the case for the other subjects? Do some subjects have no particularly "nasal" vowels? Do some subjects have two "nasal" vowels? Do some subjects have one or three "nasal" vowels?

[J] Do any subjects display behaviour which fails to make any separation of the vowels into two (or more) phonetic classes? Again, you are asked to carry out an impressionistic, rather than statistical, analysis. For example, does anyone have all vowels with the same degree of "nasality" (either high or low)?

[K] Different subjects utilise different articulatory strategies. What might these different strategies be? Consider the coarticulatory effects of tongue position on the (various) relevant muscle(s). Also consider how other muscles might be used antagonistically to prevent velum opening even when tongue position might otherwise enhance velum opening. In other words, do some people have more nasal high vowels, more nasal low vowels, more nasal high-back vowels, etc.? If so, what differing patterns of muscle activity would account for this? To answer this you need to think in terms of (a) the main muscle(s) involved in the gesture (e.g. tongue raising, backing, lowering), (b) what other articulators might the contraction of each of these muscles affect, (c) what other muscles might be involved as synergist or antagonist muscles for a particular articulatory pattern?

5) Vowel contextual nasality

Examine table 2 on the results page to answer the following questions. Also look at figures 1-4 on the results page (each of the tokens in these figures is about 0.4 ms in duration).

[L] What are the effects that you can see in the statistical tables relating to the presence or absence of an adjacent nasal consonant on a vowel's nasal airflow?

[M] Using the positive movement of the nasal airflow traces (figures 1 to 4) during speech as an indication of velum opening discuss when the velum opens relative to the start of a nasal consonant, when it closes relative to the end of a nasal consonant, and how quickly it opens and closes. That is, what is happening to the velum during a vowel in the context of no nasal consonant, a preceding nasal consonant, a following nasal consonant, etc.?

[N] Is there evidence for more than one allophone of /a:/ in this data? (How many allophones?)

6) Conclusion

A conclusion is not required (and will be ignored during marking). Your mark will be based entirely on how you answer questions [A]-[N], above.