Linguistics 001      Fall 2016     Homework 4      Due Mo 10/17

1. Phonetic duration measurements

The point of this first section is to get you started on using a computer program for acoustic analysis of speech. Your task is to embarrass Chuck Todd, the host of NBC's Meet The Press.

Todd's first interview as host was with President Barack Obama, broadcast on 9/7/2014. About eight and a half minutes into the 35-minute interview, Todd asserted, in a very serious voice, "You've not said the word 'Syria' so far in our conversation." Unfortunately for Todd, Obama had in fact said "Syria" four times at that point, most recently towards the end of the previous turn. (See here for some further details, if you're interested.)

Choose a convenient folder to work in, and download this 30-second-long selection from the interview's audio, containing Todd's statement along with part of Obama's previous turn. Depending on your computer and browser, you can do this via right-click>>Save link as... (typical for Windows) or control-click>>Save link as... (typical for Mac OS X).

Now download the free program Praat. In your recitation section, you'll learn a bit about how to use Praat to analyze audio recordings. Instructions for getting you started can be found here: "Creating an Aligned Transcript in Praat".

If you want to learn more, you can delve into online tutorials like this one.

Your task is

(1) to create an aligned transcript (as a Praat "TextGrid") for the 30-second selection;
(2) to measure the duration of Obama's rendition of the word "Syria" (in milliseconds);
(3) to measure the duration of Todd's rendition of the word "Syria" (in milliseconds);
(4) to measure the duration of the stretch between the end of Obama's "Syria" and the start of Todd's "Syria" (in seconds, with three digits after the decimal point).

(A part of) Your transcript will look like something like this this in Praat -- I've selected the word "compatible", indicating that it's 659 milliseconds long:

If you like, you can save time by transcribing in chunks between silent pauses, rather than finding the boundaries of every word -- except as required by the instructions above. So part of your transcript might look like this:

But you'll find that trying to segment (at least part of) the audio into words will be an educational experience.

You can take measurements directly from the interactive editing window in Praat -- but you should also learn to save your TextGrid as a text file for future reference, and so you should hand in your saved TextGrid along with the three requested measurements.

2. Luiseño morphology
(based on an exercise in Language Files):

The table below gives a set of 16 words in Luiseño, a Uto-Aztecan language of Southern California, with English glosses. Note that "(pl.)" means "plural".

English translation
1 nokaamay my son
2 ʔoki your house
3 potaana his blanket
4 ʔohuukapi your pipe
5 ʔotaana your blanket
6 noki my house
7 ʔomkim your (pl.) houses
8 nokaamayum my sons
9 popeew his wife
10 ʔopeew your wife
11 ʔomtaana your (pl.) blanket
12 čamhuukapi our pipe
13 pokaamay his son
14 poki his house
15 notaana my blanket
16 pohuukapi his pipe
17 nohuukapi my pipe
18 ʔokaamay your son
19 pompeewum their wives
20 pomki their house
21 čampeewum our wives
22 čamhuukapim our pipes
23 ʔomtaanam your (pl.) blankets
24 pomkaamay their son

A. List the Luiseño morphemes that correspond to each of the following English translations. Note that the plural marker has two allomorphs; list them both.

  Luiseño morpheme English gloss
1   son
2   house
3   blanket
4   wife
5   my
6   his
7   your (sg.)
8   your (pl.)
9   their
10   (plural marker)
11   pipe
12   our

B. Are the allomorphs of the plural marker phonologically conditioned?

C. If so, what are the conditioning environments?

    [course home page]    [lecture schedule]     [homework]