Assignment 1

Assignment  1:  Consonantal Place Assimilation

In the paper “The phonetics and phonology of aspects of assimilation” (one of the readings for this course),  it is noted that assimilation of place of articulation in medial –CC- clusters is very common,.  A few examples from Italian and other languages are given:

                Late Latin                             Italian

                noktu                                    notte     “night”

                oktu                                       otto       “eight”

and

                primu tempus                   Fr.  printemps  “Spring”

Traditionally these were explained as due to the speaker, being lazy, simplifying the complex heterorganic clusters by making both C‘s at the same place.

Part 1.

 I want you to make your own evaluation of this hypothesis by using Praat to make some heterorganic medial clusters and hearing what they sound like to you – and what they sound like to at least ONE other listener who is not in this course and who has not read the cited paper.  You can use your own speech samples if you wish or you can download a .wav file with the VCV utterances [apa ata aka ipi iti iki upu utu ukju] (here’s the link: )

https://docs.google.com/leaf?id=0B09lW9XhoFydOGNhMWRiYWUtZGI3ZS00ZjNhLWI0OTEtYTU0ODY1NTJmYTk2&hl=en_US

(The recording is not very good; it is noisy and has annoying clicks at beginning and start – these are a consequence of my having to increase the amplitude above what my ‘cheapo’; microphone provided.  Sorry.)

What I want you to do is relatively simple:

Make at least 4 different heterorganic medial clusters by combining the VC- from one of the VCV utterances with the –CV of another (where the place of articulation of the two C’s are different).  For the ‘default’  -VC1C2V-,  the duration of the medial silence should be the average of the two original medial stops.  If you splice the VC- and –CV exactly in the middle of the stop silence this is what you will have.    Ultimately you will be presenting your Praat-created heterorganic medial clusters to some listener in addition to yourself:  this listener should be naive (as to what you’ve done to make the samples).  But before you do that I’d like you to do at least ONE innovative variation on this experiment:  your choice, but here are some suggestions:

1.       Does it make any difference (as to what you and the other listener report hearing)  if the duration of the medial silence is varied?  (But the medial silence should not be shorter than 60 msec and not be longer than the combined duration of the original C durations from which the samples were taken).

2.       If you decide to do this experiment with your own speech samples:  does it make any difference (as to how they are perceived) if voiced stops are used instead of voiceless stops?

3.       As we all know, the place of articulation of a prevocalic stop is cued by the spectral properties of the burst AND the transition.  Does it make any difference to the percept if one cuts off the burst and leaves the transition?   (Since the speech sample I provide is Am. English, the stops’ releases have a burst and a period of aspiration following; if you opt to do this elaboration, you’ll have to be judicious in separating the burst from the aspiration since part of the transition is manifested during the aspiration.)

4.       Something else that I haven’t thought of.

I also leave up to your best judgment whether you present your samples to the other listener in some order or randomly. 

Part 2.

Here is a link to a .wav file with the utterances [ampa   anta  aŋka]. 

https://docs.google.com/leaf?id=0B09lW9XhoFydOGNhMWRiYWUtZGI3ZS00ZjNhLWI0OTEtYTU0ODY1NTJmYTk2&hl=en_US

In this case please use this speech sample for your manipulations, not your own.  From these you can create a total of 6 heterorganic –NC- clusters, i.e., labial  + alveolar, labial + velar, alveolar + labial, alveolar + velar, velar + labial, velar + alveolar.   In this case keep the duration of the voiceless stop to 50 msec.  How do you and your other listener(s) identify these creations?  The crucial question, of course, is whether they hear homorganic or heterorganic –NC- clusters.

There will not be enough data for a statistical analysis but I want to you report what you did and what the results were (what you heard and what your minimum one subject heard) and to interpret them with respect to the ‘lazy speaker’ hypothesis.  Due date:  25 July.