On the next page are two spectrograms + waveforms, labeled X and Y, both from recordings of me
saying “Toronto”. Here are two transcriptions of this name:
A) [tʰə.ˈɹɑn.ˌtʰoʊ]
B) [ˈtʃɹɑ.nə]
Examine the spectrograms/waveforms, and answer the following questions:
1) Which spectrogram/waveform matches which transcription? How can you tell? Explain with at
least two concrete examples how you decided which is which. (If you want, you can answer this
one AFTER answering the ones below, as their answers might help you choose/decide…)
2) In one of these pronunciations, where I produce a voiceless stop right after a nasal – my
‘voiceless’ stop nevertheless has a voicing bar! And I was unable to record this word in a natural-
sounding way without voicing it! … So, FIND THAT STOP with the surprising voicing bar in one of
the spectrograms and label it.
3) According to the transcription of these two place names I’ve given above: there should be a
total of three periods of voiceless fricative noise in these spectrograms. Using the transcriptions
as a guide: locate where the voiceless fricatives are, and label them with their correct IPA
symbol. Remember to use their voicing as a guide!
4) In each of the spectrograms, locate and label one period where there is voicing, and describe
which segment or segments it is associated with in the transcription (e.g. “this bit of the
spectrogram is a voiced lateral [l]”, which is clearly just an example because there are no laterals
in either of these pronunciations.)
5) In class we discussed several phonetic correlates of stress in English and two of them were pitch
and loudness. Looking at the blue pitch tracks on these recordings, and the waveforms, and the
transcriptions: do you think the changes in pitch and amplitude in these recordings reflect the
expected stress patterns of these two words? Explain with a sentence or two per word.
Spectrogram/Waveform X:
Spectrogram/Waveform Y
Part Two: Tones in Angas
The Angas languages are spoken in north-central Nigeria, and they have three level tones, High, Mid and
Low. In the transcriptions below (using the vowel ‘a’ as an example), I transcribe these as [á], [a] and [à]
respectively (that is: if you see NO tone mark on a vowel, that means it has mid tone.) Recall also that [ǎ]
represents rising tone, and [â] represents falling tone. (Read that sentence again and be sure you
understand it before continuing!)
There are three sets of Angas data: the bare nouns; the nouns as they appear when followed by a case
marker; and the nouns as they appear when followed by an adjective.
Bare Nouns Nouns when followed by case Nouns when followed by an adjective
[téŋ] ‘rope’ [téŋ] rope (+ case) [têŋ] rope (+ adj.)
[mús] ‘cat’ [mús] cat (+ case) [mûs] cat (+ adj)
[tʃén] ‘hoe’ [tʃén] hoe (+ case) [tʃên] hoe (+ adj)
[njí] ‘elephant’ [njí] elephant (+ case) [njî] elephant (+ adj)
[ʒwal] ‘boy’ [ʒwǎl] boy (+ case) [ʒwâl] boy (+adj)
[ʔas] ‘dog’ [ʔǎs] dog (+ case) [ʔâs] dog (+adj)
[ɟem] ‘child’ [ɟěm] child (+ case) [ɟêm] child (+adj)
[màs] ‘locust bean’ [mǎs] locust bean (+ case) [màs] locust bean (+adj)
[pùk] ‘soup’ [pǔk] soup (+ case) [pùk] soup (+ adj)
[ʔàs] ‘tooth’ [ʔǎs] tooth (+ case) [ʔàs] tooth (+ adj)
Q1) When you add the “+ case” meaning to a noun, sometimes its tone changes and sometimes
it doesn’t; same for adding “+ adjective”. Describe how these changes work in prose, as
succinctly as you can. In addition to using words, illustrate these changes as mappings from
bare nouns to affixed forms: e.g. (just as an example, not true of the data):
/báse/ + “+case” [bàse]
Q2) There are two affixes in these data: one that means “+ case” and one that means
“+adjective” … But these affixes contain no segments! … What are the underlying forms of
these two morphemes? Explain.
Q3) If you say a bare noun in this language of the form [dùlí]: what do you think its “noun +
case” and “noun + adjective” forms would be? Explain your answer. (There are multiple
possible right answers here! Just explain your reasoning.)
Part Three: Syllables in Child Spanish
Here are data from two children learning Spanish: it includes adult pronunciations of these words, taken
to be the ‘input’ to the children’s phonology, and their own output (as well as the gloss.)1 Both kids do
something systematic, but different, with these words, and syllable shape is relevant. Note that [tʃ] in
Spanish, as in English, should be considered a single segment, an affricate.
SD1’s Spanish at age 3;4
adult input child output gloss adult input child output gloss
/plato/ [plato] plate /kampanas/ [kapanas] bells
/bloke/ [bloke] block /fuente/ [fuete] water fountain
/fresa/ [fresa] strawberry /gantʃo/ [gatʃo] hook
/tren/ [tren] train /dulses/ [duses] sweets
/tʃikles/ [tʃikles] pieces of
gum (pl.)
/tʃaɲklas/ [tʃaklas] sandals
Q1) Assume (as I believe is correct) that both of these children, and Spanish more generally, syllabifies
all the segment sequences you see here the same that English would. With that in mind: give
syllabification trees (including Onset/Nucleus/Coda) for the adult, target pronunciation of the words
meaning ‘strawberry’, ‘piece of gum, singular’, ‘sweets’, ‘sombrero’ and ‘sandals’.
Q2) Given just the adult data you’ve seen in this table: describe using phonological features the natural
classes of consonants which can appear in Spanish Onset and Coda positions. If there can be more
than one segment in a position (i.e. two onset consonants in a row or two coda consonants…), describe
their individual properties: (e.g.: the first onset C has to be [+something]…)
Q3) With those syllabifications in mind: describe how each child changes the adult target Spanish
inputs. That is: compared to your answers in Q2, which adult Spanish syllable pattern does each child
appear to NOT allow, and how do they go about avoiding that structure? Hint: you will need to say
1 Note: some segmental changes have been suppressed here for clarity’s sake.
BL4’s Spanish at age 2;8
adult input child output gloss adult input child output gloss
/plato/ [pato] plate /manzana/ [manzana] apple
/bloke/ [boke] block /dulses/ [dulses] sweets
/fresa/ [fesa] strawberry /falda/ [falda] skirt
/tren/ [ten] train /arbol/ [arbol] tree
/tʃikle/ [tʃike] piece of
gum (sg.)
/kumpleaɲos/ [kumpeaɲos] birthday
/negro/ [nego] black /sombrero/ [sombero] hat
something both about the syllable shape and WHERE in the word it occurs. You can do all of this just
with words, and with reference to just Cs and Vs in different syllable positions.
Q4) Write phonological rules, as simple as you can make them, to capture each child’s pattern. If you
can get away with just Cs and Vs, do so. If you need to make reference to some phonological features,
do so. Remember that your rules canNOT reference syllable positions – e.g. we have no way of saying
/A/ [B] / ‘in an onset’.
Q5) We discussed in class that one way in which Spanish syllable structure differs from English is that it
does not allow s+stop onset clusters. With that in mind: how do you think both children above, BL4 and
SD1, would pronounce the Spanish word for ‘star’, whose adult input form would be /estreja/? Make
sure you refer to the adult word’s syllabification in your answer.
Part Four: Allomorphy and phonological rules in Ditidaht
Ditidaht is a Wakashan language spoken traditionally at the southern end of Vancouver Island. Here are
some words of the language, with their glosses. Every word has one root, and some of them also have
one affix, although the meaning of that affix may be more like an entire phrase or sentence in English.
Ditidaht gloss Ditidaht gloss
baʔas ‘house’ daqwuɫ ‘pub’
bakwuɫ ‘store’ tʃuʃʼas ‘tree’
biɫsaq ‘the salmon spear shaft’ baʔsa ‘it is a house’
tʃuʃʼsaq ‘the tree’ bakwɫaq ‘the store’
biɫis ‘salmon spear shaft’ daqwɫa ‘it is a pub’
ɫuʔaɫ ‘board’ ɫuʔɫa ‘it is a board’
baʔsaq ‘the house’ ɫuʔɫaq ‘the board’
In this data set: the affixes have only one allomorph, but the roots each have two allomorphs. Bearing
that in mind:
Q1) What are the two affixes’ phonological shapes, and what does each one of them mean? (Just give
the relevant gloss.)
Q2) Now with respect to the root: describe the difference between their two allomorphs, and in what
phonological context you see each of them (i.e. not what morphological context.) … Your answer
should be (basically) the same for all the roots – that is, describe the root allomorph shapes in a
sufficiently general way that it describes all of them (while still being accurate!)
Q3) Write a phonological rule that accounts for the alternation you see in the roots.
Q4) Now suppose that the syllable structure of Ditidaht is CV(C) (i.e. obligatory onset, optional coda.)
With this in mind: do you think that the phonological rule you wrote in Q3) could be driven by the
language’s syllable structure? If so, what aspect of syllabification is the rule trying to repair? If not, why
not? Use at least one concrete example from the data set to make your point.
