D A Y 1 5
C E S A R K O I R A L A
Ling 403/603
Introduction to Phonology
Morphophonemic Analysis
The purpose of Morphophonemic Analysis is to discover a set of underlying forms and ordered rules that is consistent with the data.
Morphophonemic Analysis
The purpose of Morphophonemic Analysis is to discover a set of underlying forms and ordered rules that is consistent with the data.
Procedure of Morphophonemic Analysis:
Choosing the Underlying Representation
This is crucial (and not always easy!)
This , often, involves considering more than one hypothesis, with the final choice defended by its leading to a working analysis.
For example: Chimwiini:
Hence, we have two hypothesis:
1. Underlyingly, the vowels are long. Later, shortening rules apply.
2. Underlyingly, the vowels are short. Later, lengethening rules apply.
Long vowels alternate
with short
Choosing the Underlying Representation
For the segments that alternate, follow the hypothesis you made about the underlying forms, implementing it consistently throughout the data.
Under hypothesis 1, the underlying representation is /so:m/ for the root ‘read’ and /-o:w/ for the passive suffix.
The segments that don’t alternate can be assumed to be phonemically identical in their underlying representation to their surface representation.
A clue for Choosing the Underlying Representation
Contextually limited contrast.
For example, though vowel length is contrastive in Chimwiini, only short vowels are allowed when more than 3 syllables from the end of the phrase.
This means the contrast is neutralized in that particular environment.
Hence, there must be a rule that wipes out the contrast in that environment.
Don’t analyze in a direction opposite to that of a neutralization.
A clue for Choosing the Underlying Representation
If we analyze Chimwiini with ‘Shortening’, we analyze in the direction of neutralization.
If we analyze Chimwiini with ‘lengthening’, we analyze in the direction opposite to that of a neutralization.
Lardil
Lardil
What are the suffixes for accusative non-future and accusative future?
Lardil
What are the suffixes for accusative non-future and accusative future?
What should be our two hypothesis about the underlying forms?
More data:
Hypothesis 1 - Deletion
Hypothesis 2 - Insertion
Q. What determines the choice of vowel we insert?
No principled basis. Hence, we choose hypothesis 1 over 2.
Moreover, since Lardil makes no contrast of vowel vs. consonant in the position after a vowel (only consonants are allowed), deletion analysis, which is based on this pattern, will work.
Lardil
Some more data: Alternation of vowel quality
Lardil
Some more data: Alternation of vowel quality
Lardil
Again, we need to consider both hypotheses:
1. [a] changes to [u] in accusative forms.
2. [u] changes to [a] in uninflected forms.
Some more data: Alternation of vowel quality
Hypothesis 1: [a] changes to [u]
Q: why /a/ [u] when a suffix is present?
Q: Is this true for all the data we have seen so far?
Some more data: Alternation of vowel quality
Hypothesis 1: [a] changes to [u]
Q: why /a/ [u] when a suffix is present?
Q: Is this true for all the data we have seen so far? NO
Some more data: Alternation of vowel quality
Hypothesis 1: [a] changes to [u]
Some more data: Alternation of vowel quality
Hypothesis 1: [a] changes to [u]
Some more data: Alternation of vowel quality
Hypothesis 1: [a] changes to [u]
In fact:
Hypothesis 1 gives us wrong surface forms.
Hypothesis 1: [a] changes to [u]
In fact:
Hypothesis 1 gives us wrong surface forms.
[a] doesn’t change to [u]
Hypothesis 2: /u/ changes to [a]
Q: Is this analysis consistent with all the data we have seen so far?
Some more data: Alternation of vowel quality
Hypothesis 2: /u/ changes to [a]
Hypothesis 2: /u/ changes to [a]
Hypothesis 2: /u/ changes to [a]
#
Lardil
Some more data: Alternation of vowel quality
Lardil
We can see alternations between [æ] and [i]
[w] appears in Accusative future suffix.
Some more data: Alternation of vowel quality
Lardil
(Revised)
Lardil
Lardil
We can see alternations between [æ] and [i]
[w] appears in Accusative future suffix.
Some more data: Alternation of vowel quality
Lardil
We can see alternations between [æ] and [i]
[w] appears in Accusative future suffix.
Some more data: Alternation of vowel quality
[w] appears in Accusative future suffix.
Lardil
[w] appears in Accusative future suffix.
Lardil
This can be seen as another rule that does not allow to vowels together (Like vowel deletion)
It is common across languages that two rules can apply for taking care of the same constraint.
[w] appears in Accusative future suffix.
Lardil
This can be seen as another rule that does not allow to vowels together (Like vowel deletion)
It is common across languages that two rules can apply for taking care of the same constraint.
What should be the ordering of /w/ Epenthesis with respect to Vowel Deletion?
Lardil
Note: Vowel Epenthesis cannot apply when /w/ Epenthesis applies.
What is this kind of rule ordering called?
Lardil
Some more data
Lardil
We can see [k] in accusative future forms.
Some more data
Lardil
We can see [k] in accusative future forms.
Some more data
Lardil
We can see [k] in accusative future forms.
Some more data
Lardil
We can see [k] in accusative future forms.
Does the ordering of this rule matter? Say, with respect to vowel deletion or /w/ epenthesis?
Some more data
Lardil
New data:
Lardil
Morphemes for the roots are different in Uninflected and Accusative forms.
They end in different vowels.
New data:
Lardil
Two hypotheses:
1. Insertion: Vowel inserted in accusative forms.
2. Deletion: vowels are deleted in uninflected forms.
New data:
Hypothesis 1: Insertion
Problem: How do we determine which vowel to insert.
No Principled way to answer it.
New data:
Hypothesis 1: Insertion
Problem: How do we determine which vowel to insert.
No Principled way to answer it.
New data:
Hypothesis 2: Deletion
The rule will simply deletes all the vowels (no matter which) at word final position. It is a common phonological process seen in other languages too.
New data:
Hypothesis 2: Deletion
The rule will simply deletes all the vowels (no matter which) at word final position. It is a common phonological process seen in other languages too.
Looks promising…but is it consistent with our previous analysis?
Basically, are there other words that end in vowels in this language?
New data:
Hypothesis 2: Deletion
New data:
Hypothesis 2: Deletion
New data:
Is there any difference in
these two forms?
Hypothesis 2: Deletion
Hypothesis 2: Deletion
Lardil
Consider some more new data:
Lardil
Consider some more new data:
There is vowel ~ zero alternation like previous data.
There is also consonant ~ zero alternation.
Lardil
Consider some more new data:
There is vowel ~ zero alternation like previous data.
There is also consonant ~ zero alternation.
Lardil
Consider some more new data:
Like before:
1. Hypothesis 1: Insertion
2. Hypothesis 2: Deletion
Lardil
Consider some more new data:
Like before:
1. Hypothesis 1: Insertion
2. Hypothesis 2: Deletion
Can you think of a problem with
hypothesis 1?
Lardil
Consider some more new data:
Like before:
1. Hypothesis 1: Insertion
2. Hypothesis 2: Deletion
Can you think of a problem with
hypothesis 1?
Why are different consonants
inserted?
Hypothesis 2: Deletion
First, lets see what the analysis so far result in:
Hypothesis 2: Deletion
Notice Urs…consistent
with previous analysis
Hypothesis 2: Deletion (Cluster reduction)
1. Apocope applies word finally and creates consonant clusters.
2. Lardil seems to dislike word final consonant clusters. This is common across languages.
Notice Urs…consistent
with previous analysis
What should be the ordering of this rule with respect to Apocope?
Rule Orderings (Hasse Diagram)
What did we learn from today’s class?
What did we learn from today’s class?
How to do morphophonemic analysis?
1. It involves coming up with the set of underlying forms and ordered rules that is consistent with the data.
What did we learn from today’s class?
How to do morphophonemic analysis?
1. It involves coming up with the set of underlying forms and ordered rules that is consistent with the data.
2. Whenever you have more than two hypothesis for the underlying forms, consider both…and then choose the one that is: consistent with the data + grounded in phonological principles that you have seen in other languages.