Lin 205 Basic Phonology by Kola Adeniyi
Lin 205 Basic Phonology by Kola Adeniyi
Lin 205 Basic Phonology by Kola Adeniyi
BASIC PHONOLOGY
By Kolawole ADENIYI
ABOUT THE COURSE
This course is the beginning of our building on the foundation laid by phonetics. Remember that
assertion by Pike 1947 that “phonetics gathers the raw material, phonology cooks it”. It is here that
we will begin the processing. You will be introduced to the concepts germane to phonological
analysis, their application and relationships between them. And of course, you will be guided
through the analyses of some data. But there is no cause to worry: phonology is as simple as saying
that “if I can see 2 people in front of me right now, then the 2 of them are here”. This is however
based on the condition that you know your phonetics.
COURSE OUTLINE
1) Definition of Phonology
2) Steps in Practical Phonological Analysis (premises)
3) Concepts: Contrast, Distinction, and Redundancy
4) The Phoneme
5) Phonemic Features
6) Suspicious Sounds (Pairs/Sets)
7) Allophones
8) Distribution (definitions and differences): Complementary Distribution (CD), Free Variation,
Parallel Distribution (PD), Neutralisation
9) Discovery Procedure
10) Phonetic and Phonemic Transcriptions
11) Phonological Processes 1
12) Phonological Processes 2: assimilation
13) Environments for Phonological Processes
14) Phonological Rules
15) Naturalness
LECTURE 1: Definition of Phonology
Introduction
In this lecture, we’ll define phonology, and see why structure is so important in achieving meaningful
communication.
Objectives
At the end of this lecture, you should be able to:
Define phonetics
Define phonology
Understand the link between phonetics and phonology
PRE-TEST
What is phonetics?
Phonology is the study of how speech sounds are used in natural language. Hyman 1975 views it as
the study of sound system of language. But whatever form the definition takes, the point remains
that phonology deals with the organisation of how speech sounds are organised in the process of
human communication, and what function each sound performs in the organisation. You will agree
with me that if a speaker of English brings [b], [i:], and [d]together, he will have *bi:d+ “bead”. But if
the same speaker mistakenly put *k+ where *d+ should be, you will have *bi:k+ “beak”, which is
another word entirely.
The point here, therefore, is that there is something about the choice of [d] when you want
to refer to the form of necklace called “bead”, and the choice of *k+ when you want to refer to a part
of the body of a fowl called “beak”. Only the two sounds vary in the two words; so what is it about
them that results in the changes in the meaning of the words? That is the sort of question phonology
answers.
So we will be looking at the sounds that each language uses in its inventory, the status of
each of the sounds, and the processes that each sound undergoes (and under what conditions they
undergo the changes).
Let me make a point clearer at this point: in phonetics, we studied the sounds in isolation;
we described them by means of their phonetic features. We even grouped the sounds into natural
classes (remember natural classes?). But in phonology, we will be looking at what happens to the
sounds when they are in the company of one another. Take a look at the Yaqui data below and tell
me whether there is any link at all between *t+ and *t∫+
teput “flea”
teput∫-im “fleas”
Again, take a look at the Yoruba data below and see whether there is a relationship between [n] and
[ɱ]
nl “is going”
ɱf “is breaking”
To give you a clue, in Yaqui, *t+ changes to *t∫+ because there is now *-im] after it (it is no
longer at the end of the word!). In Yoruba, you will observe that the syllabic nasals share the same
place of articulation with the sounds following them. Meaning that a syllabic nasal sound in Yoruba
“borrows” the place of articulation of the sound directly following it! Phonology simply looks at such
effects as these, their patterns (and those patterns change from language to language), among
others.
Summary
Phonology takes up the study of speech sounds where phonetics stops. It goes on to throw light on
the structure of sounds, particularly when they are used in context. And because sounds are always
used in context, phonology becomes central to our study of language.
Recommended texts:
Hyman, L.M. 1975. Phonology Theory and Analysis. New-York: Holt, Rinehart and Winston
Collinge, N. E. Ed. 1999. An Encyclopedia of Language. London and New-York: Rutledge.
Pike, K.L. 1947. Phonemics: A technique for Reducing Languages to Writing. University of Michigan
Fudge, Eric. 2005. Language as Organised Sounds: Phonology. In: N. E. Collinge, An Encyclopaedia of
Language. London & New York: Rutledge, 17-37
Clark, J., C. Yallop and J. Fletcher. 2007. An Introduction to Phonetics and Phonology. Oxford:
Blackwell
LECTURE 2: Steps in Practical Phonological Analysis (premises)
Introduction
Phonology is not done “just like that”. One thing must logically precede another for the expected
result to materialise. In this lecture, you’ll learn the stages involved in analysing data.
Objectives
In this lecture, you’ll learn:
The evergreen premises of phonological analysis given by pike (1947)
Why you still need the premises today
PRE-TEST:
Why must you wear cardigans during winter in Canada?
Give a phonetic explanation for the slight variations in the production of the same sounds by the
same speakers at different times
Now that we have defined phonology in the last unit, let us take a look at some of the factors that
ensure that the Phonologist achieves his aim.
There are many factors conditioning the behaviour of speech sounds in context. But for ease of
remembering, Pike 1947 summarises the factors under 4 categories, which he calls premises of
phonological analysis. In short, a premise means a proposition guiding the behaviour of speech
sounds in context. Let me put it differently, a premise, in our context, is a condition guiding the
behaviour of sounds in words (or larger units). These premises are crucial to your survival in
phonology because they derive from language universals, in the sense that they tend to be the same
in all languages. Here are the 4 premises:
I. Sounds tend to be modified by their environments (they tend to slur into one another, or
slur into silence). By way of illustration, this is like saying that if a leaf stays long on the
soap, it eventually becomes soap itself. Yet another illustration is that of a Nigerian who
after two Months in America begins to pronounce the word /betә/ as *beɾә] etc. The
point is that just as it happens to humans, it happens to speech sounds, such that when
/t/ occurs before /u:/ in /tu:/ English ‘two’, the /t/ itself being an unrounded sound,
acquires a level of roundedness from /u:/, and you actually have *tʷu:+. Now try it out
phonetically: you’ll see that the lip rounding begins as soon as you configure your
articulators for the production of /t/. Great! So, this is how sounds behave; only that it
depends on the environmental influence.
Now, remember that we have done assimilation in LIN 241 and 252. Here is the principle
that explains that concept right! We defined assimilation, then, as the process by which
sounds take on features from their environments. Take another look at the data on
Yoruba syllabic nasal we analysed in unit 1 of this course. Let me give you an example
that you should find funny: in certain languages, you will never find a voiced sound at
the end of a word! Even if the sound is voiced, it must become voiceless the moment is
occurs word-finally! Can you think of the reason for this? The next activity to a word-
final position is silence, and voicelessness is only a component of silence; so in such
languages, voicelessness is that feature that is in a word final position that transfers to
every word-final sound to ensure they are all voiceless/devoiced.
II. Sound systems have a tendency toward phonetic symmetry. Here is an illustration: there is a
gathering strictly for married men. If you find ten men and nine women, you should
expect to find one more woman somewhere around, if you search well. Why?
Something tells you that they should all be in pairs! That’s it! Phonetically speaking,
sounds that are similar behave in similar ways. For instance, if a particular language has
/p t k/ and then /b d/, you should expect to also find /g/ because this symmetry predicts
that plosives come in voiced/voiceless pairs in the language. This symmetry applies to
different categories of sounds.
III. Sounds tend to fluctuate. Come to think of it, if you produce the same sound two
consecutive times, do you think you can say exactly the same thing the two times?
Phonetically, you can’t, and this is the idea of sound fluctuation. Sometimes these
fluctuations are so pronounced that they become perceptible to the ear of foreigners
(native speakers are often unaware of it). At that point, the phonetics of that language
will record it, and it becomes a duty for the phonology to tell whether the fluctuation is
significant or not.
IV. Characteristic sequences of sounds exact structural pressure on the phonemic interpretation
of suspicious segments or suspicious sequence of segments. Structural influence on the
interpretation of sounds in line with the predominant syllable structure or the word or
morpheme structures in any particular language. This is more serious with sounds of
unstable qualities (or which are phonetically complex) e.g. semivowels, affricates, long
vowels etc.
Consider when a language has a syllable structure that forbids clusters of any kind, and
yet in your transcription, you feel you heard something like *[pai]. The prevailing syllable
structure tells you that the last sound should not be a vowel, because it breaks the rule
that forbids cluster by resulting in a vowel cluster. [i] is therefore a sound of
indeterminate quality. But the syllable structure suggests that it is actually a consonant
[y] which is a semi-vowel.
This rule many times result in glide formation in languages. Take a look at the Emai data
below:
Gloss underlying form surface form
Toe /chiɛ/ [ehjɛ]
Elder /ᴐdiᴐ/ [ᴐdjᴐ]
You will notice that where there are contiguous vowels underlyingly, the first of those
vowels comes out as a glide in the surface form (which is the form spoken). This is because the
language does not allow vowel clusters. You should be able to deduce by this that the language
allows consonant clusters.
Lastly under this premise is the determination of sounds like long vowels and affricates.
Whether to recognise a sequence of long vowel as just one or two, or whether to recognise such
sounds as [t∫+ as just one sound segment or two. These depend on the prevailing structure in the
language. By this I mean that if you have heard clearly other sequences of consonants that are seen
as one, and you have no doubt about them, and you now have a problem determining just a few,
you can invoke the prevailing structure principle to resolve your dilemma. Also, if you have already
convincingly found sequences of vowels such as [aa], [ee], and [oo], why should you not consider
interpreting [u:] as [uu]? After all, it is the prevailing pattern. And remember the premise of pattern
congruity.
Now, these are only the major premises; there are more, so I recommend that you read Pike
1947 yourself. Begin from 57 (it’s not a crime if you read to the end of the book, after all, its all
about phonology).
Summary
You’ve learnt in this lecture that there are certain reasonings behind phonological analysis, and that
once you understand them, having a direction for approaching any data becomes easy for you.
POST-TEST:
Profer a phonetic explanation to the principle of pattern congruity
What can make a sound become labialised?
Recommended texts
Pike, K.L. 1947. Phonemics: A technique for Reducing Languages to Writing. University of Michigan
Press
Schane S. 1973. Generative Phonology. New Jersey: Prentice-Hall Inc.
Gleason’s Workbook, Series 3
Hyman, L.M. 1975. Phonology Theory and Analysis. New-York: Holt, Rinehart and Winston
Welmers, W.E. 1973. African Language Structures. Berkeley: University of California Press.
Clark, J., C. Yallop and J. Fletcher. 2007. An Introduction to Phonetics and Phonology. Oxford:
Blackwell
LECTURE 3: Concepts: Contrast, Distinction, and Redundancy
Introduction
In this lecture, you’ll learn what it takes for sounds to be different from one another and the effect
of the differences on the languages.
Objectives
At the end of this course, you’ll learn
The role of features in the meaning of words
Why some features are not as important as others in certain contexts
PRE-TEST
What does it mean for sounds to be different from one another?
Contrast
Baudoin de Courtenay (1895) was said to be the first to use the word ‘contrast’. The basic idea
expressed by the term is that of a sound being different from another. Usually, such differences
referred to by the term are of features. The difference however may be phonetic in the sense that it
does not affect meaning, and it can affect meaning. Let me illustrate this for you:
*tʷu:+ ‘two’
[teɪbl] ‘table’
The initial sounds in the two English words above are not exactly the same in the sense that
one is labialised*tʷ+ and the other is not *t+. When you look well, you will find that this difference
(labialisation) does not affect the meaning of words in English. Another way of saying this is that
removing the feature in question does not change the meaning of the word; it only indicates
inaccurate pronunciation. You will also find that there is something in the environment that
warrants the presence of the feature: it comes immediately before [u:] a rounded sound! So it is the
phonetic environment that makes it surface there. So we say contrasts like this are only phonetic, or
allophonic.
*tʷu:+ ‘two’
[dʷu:] ‘do’
The only way we can explain the difference in the meaning of the two words in this data is
the presence of *tʷ+ in the first and that of *dʷ+ in the second. Recall that this is only a matter of
voicing feature. Because this ‘little’ feature difference has affected the lexical meaning of the words
concerned, we say that it is distinctive (makes a sound different), or phonemic. Note that there is
nothing in the phonetic environment, in the case of distinctive contrast, that can explain it.
Now, to define the terms distinction and redundancy. Distinction is when contrast affects the
lexical meaning of a word, and redundancy is when it does not.
Summary
When we say that a sound contrasts with another, we mean it is different such that the difference
can affect meaning in a language. Sometimes, the raw data collected on the field may reflect
differences that really don’t affect meaning; in which case we say such features are redundant.
Post test
What is contrast?
Give another word for redundancy
Recommended Texts
Collinge, N. E. Ed. 1999. An Encyclopedia of Language. London and New-York: Rutledge.
Odden 2004. Phonemes, contrast, and phonetic detail. http://www.ling.ohio-
state.edu/~odden/IntroducingPhonology/Chapter%203%20theory%20extension.pdf
LECTURE 4: The Phoneme
Introduction
What is our aim when we are confronted with a mass of data to analyse? It’s to determine which of
the sounds are core and which are not; in other words which ones are phonemes, and which are not.
Thus, in this lecture, you’ll learn what it means for a sound to be a phoneme.
Objectives
At the end of this lecture, you’d have learnt:
What makes a sound a phoneme
How to determine the factors and prove they valid
Pre-test:
With your understanding of Unit 3, define the term phonemic feature
It is defined in Odden 2004 as ‘a separate type of symbolic object which stands for a
particular class of phones encountered in a language, following certain analytic rules’
Hyman 1975: ‘minimal units of sound capable of distinguishing words of different meaning’
Pike 1947: ‘one of the significant units of sound arrived at for a particular language by the
analytical procedures developed from the basic premises’
Jones 1931: ‘a family of sounds in a given language, consisting of an important sound of the
language, together with other related sounds, which take its place in particular sound
sequences’
Gleason 1955: ‘a class of sounds which are phonetically similar and show certain
characteristic patterns of distribution in the language or dialect under consideration’
You will find different definition in the literature; but one thing that is common is that for a
sound to be regarded as a phoneme, it has to appear in an environment phonetically identical
(analogous) to a similar sound, and still be the cause of change in meaning. Let’s take a look at a
Yoruba data. Tone is not marked.
[ka] ‘read/confess’
[ga] ‘stand’
[ta] ‘sell’
[da] ‘pour’
[ba] ‘perch’
[sa] ‘pick’
As the first sound in each of the 6 sounds above changes, meanings change. And note that
each of the changing sounds occurs word-initially, and before [a] (identical environment). For the
fact that the alternation of the consonant sounds in identical environment result in change in
meaning, we say that each of the sounds is a phoneme.
At this point, we want to examine certain definitions of the term phoneme; our focus is
going to be the significance of each of those definitions to the understanding of the term.
Of course we have already established that a phoneme can effect change in the meaning of
words; that is Hyman 1975. Pike 1947 also makes this point by the phrase ‘significant unit’. Another
point from Hyman’s definition is that the unit to be called a phoneme must be the minimal unit, that
is, it can no longer be broken down (if you break it down, it’ll no longer stand as a sound). This is also
a point made my Pike’s 1947 definition.
Odden 2004 says a phoneme is ‘symbolic’ and it ‘stands for a particular class of phones’;
Jones 1931 says ‘a family of sounds in a given language’; Gleason 1955 says ‘a class of sounds which
are phonetically similar’. The implication of these is that a phoneme is not necessarily a single sound;
it refers to a collection of sounds that within themselves don’t affect meaning.
This is what I mean: English has the different variants of /p/, aspirated [pʰ], unreleased [p°],
to mention only 2. Now, take a look at the following data.
[әplaɪ] „apply‟
[pʰɪn] „pin‟
p°] „cup‟
But supposing you put /b/ in each of the slots above, I can assure you that the 1 st and 3rd
words will become jargons, while the second will become another word entirely: „bin‟. Thus, /b/
must be another phoneme (member of another group of sounds)
Let me also point this out to you: Gleason said ‘in the language or dialect under
consideration’. This means each language has its own different sets of phonemes. You must
interpret each language independently. Consider the hypothetical language below
Language K
[әkɪp°] ‘Mother’
[әkɪp] ‘Aunt’
The only difference between the two words is [°] the unreleased superscript; but the
meaning has changed. This means [p] and [p°] are two different phonemes (or belong to two
different phonemes). Now, compare this result with the English data I gave you above! What do you
find?
A fall out of this idea that a group of sounds THAT ARE PHONETICALLY SIMILAR can form a
phoneme is that the distribution of each of the members of the group is predictable. That is, in the
English data we just treated (and of course in the totality of English), you always find [pʰ] word-
initially, [p°] word-finally, while [p] is wider in its occurrence. A statement of where each member of
the group occurs is called the distribution. Distribution is possible because the occurrence of each
member of the group of sounds forming the phoneme is predictable. We’ll say more on this when
we get to complementary distribution in unit 7.
Summary
A phoneme is a cover term for a group of sounds that are contextual variants of each other. This fact
shows that the phoneme is usually discovered through analysis; it is analysis that will bring to the
fore the different variants to be together called a phoneme, and the relationship between them.
Post-test
What is a phoneme?
Recommended Texts
Clark, J., C. Yallop and J. Fletcher. 2007. An Introduction to Phonetics and Phonology. Oxford:
Blackwell
Collinge, N. E. Ed. 1999. An Encyclopedia of Language. London and New-York: Rutledge.
LECTURE 5: Phonemic Features
Introduction
You’ll learn here that some features can singlehandedly change or alter meaning. Those features are
our focus
Objectives
Pre-test
If aspiration does not affect meaning in English, but it affects it in language K, what do you think is
happening? As a Linguist, how can you explain this?
Features
Features are properties of linguistic units, e.g. voicing, labial, velar, nasal or nasalised, low (tone),
high (tone) etc., (remember your phonetics?). Each sound is composed of features, for instance:
[t] has the following features, among others: voiceless, alveolar, etc.
Do you still remember that in labelling a vowel sound, you are required to talk of stuffs like
horizontal position, vertical position and lip rounding. These are all features.
Phonemic Features
Now, let us return to the questions I gave you in the pre-test, /e/, /ø/
/e/ is front close-mid unrounded vowel
/ø/ is front close-mid rounded vowel
Let’s put this sounds in context, using our hypothetical language, language K.
bebe ‘book’
bebø „teacher‟
bøbe „librarian‟
bøbø „library‟
You will observe that the difference between these two sounds is lip-rounding. That is, other
than lip-rounding, they both have the same features. Now, look at the data. The only difference
between items 1 and 2 is restricted to the two sounds we just labelled; and that is what results in
the change in meaning. Do you really get the gist?
The question here is whether nasalisation is phonemic in this Yoruba data. Remember, we
have made two points about phonemic features: it independently affects meaning; you won’t be
able to predict where it occurs. At this level in phonology, keep those two points in mind.
Now looking at predictability, can you predict where nasalised vowels occur in this data?
Yes. It always occurs after nasal consonants. Always! Since it can be predicted, it is not a
phonemic feature here in this data. It is only phonetic because of that premise we mentioned:
sounds tend to be more like their environments! Meaning the nasal feature in the vowels is only an
environmental influence. That is why it can be predicted. Once a nasal sound occurs before it
(environment), the automatically becomes nasalised.
To recapitulate, a phonemic feature cannot be predicted by reading the environment; it
doesn’t surface as a result of the environment.
What are the functions of phonemic features? Schane (1973:28), under the caption
‘distinctive features’ outlines three of the functions. (1) they are capable of describing systematic
phonetics, that is, separating which sounds are core and which are not to any particular language;
(2) they serve to differentiate lexical items, for instance the voicing feature solely distinguishes
between /pɪn/ ‘pin’ and /bɪn/ in English; (3) they define natural class. We’ll talk more about this by
and by.
Summary
Any feature that either alters or changes the meaning of a word when tampered with is phonemic
(distinctive). Such features are not predictable.
Post-test
What are the other words that have the same meaning as phonemic.
Texts: Zwicky, A. M. 1982. Phonemes and Features. Innovations in linguistics Education 2.2, 55-76
Schane S. 1973. Generative Phonology. New Jersey: Prentice-Hall Inc.
Clark, J., C. Yallop and J. Fletcher. 2007. An Introduction to Phonetics and Phonology. Oxford:
Blackwell
Collinge, N. E. Ed. 1999. An Encyclopedia of Language. London and New-York: Rutledge.
LECTURE 6: Suspicious Sounds (Pairs/Sets)
Introduction
Phonological analysis is not just done arbitrarily. Before you even begin to investigate any pair or set
of sounds, you will have to establish a certain level of similarity between them. That level of
similarity is the focus of this lecture.
Objectives
Basically, you’re going to learn:
the criteria for setting up suspicious pair/set
minimal pairs/sets and their importance
importance of such groupings
Pre-test
What are the signs that will make you conclude that certain sounds are suspicious?
Sounds are said to be suspicious when they have strong phonetic similarities. I mean the phonetic
similarities between them will be so strong that you begin to think that they may actually be
allophones of the same phoneme. (Don’t worry; I will explain what I mean by allophone shortly).
When you have only two sounds that are similar in this manner, you call it a pair, but when there are
more than two sounds, you say that you have a set.
In setting up suspicious pairs or sets, Pike 1947: 69 advises that you ensure that the sounds
are similar in all but one feature. On the basis of this, I will now give you examples of suspicious
sounds usually encountered in the course of language documentation. But note suspicion can come
in any form: voicing difference, place of articulation, features relating to secondary articulation, etc.
[P pʰ]
[g gh]
[f v]
[k g]
[ ᴐ]
[P pʲ]
[l ɫ]
[p b]
[∫ ʒ]
[m ɱ]
I expect a particular question to have formed in your mind: what happens if a particular
sound is not suspicious with any other sound in the phonetic inventory?
I simply present to you the answer given by Pike 1947: 71...THEY ARE SEPARATE
PHONEMES!
Now I don‟t expect you to be at a loss here; if Shehu‟s native language is Hausa, and
Kemi‟s is Yoruba; Shehu‟s linage has always been Muslims, and Kemi‟s has always been
Christians, you don‟t need to waste your time investigating whether they are members of the same
family! And now, it‟s up to you to relate that to Linguistics.
Minimal Pairs and Minimal Sets
A minimal pair is regarded as a Linguist’s delight, because once you find it, you need not investigate
any further. Elugbe (class notes) says ‘often, two sounds occur in identical environments, and there
are pairs of words which would have been identical but for the difference in the two sounds, we call
the two words minimal pair’. Let me illustrate this with this Yoruba words
[ija] ‘mother’
[ij ] ‘feather’
In the Yoruba data, the only thing we can identify as the source of the di erent meanings of
the two sounds is the [ ]-[ ] alternation. In the Ebira data, the only observable cause of the different
meanings is the differences in the tone on each vowel. Observe that the sounds (or features) causing
the difference in meaning in each of the two data occur in exactly the same environment, that is,
analogous environment. This fact of similarity of the environment of occurrence of the deferring
sounds is vital to our determination of what is a minimal pair or set.
Why is a minimal pair (or set) a Linguist‟s delight?
When a Linguistic investigator has identified a suspicious pair, he scans his data to see
whether the two sounds can be found occurring in analogous environment. Finding the suspicious
sounds in analogous environments is the easiest way to determine whether sounds are independent
phonemes. For instance, if the suspicious sounds are found in a minimal pair (analogous
environment), then the Linguistic investigator easily concludes that they are different phonemes.
What if you really don‟t find the suspicious sounds in exactly the same environments; but
they are in almost the same environments?
That‟s what is called NEAR MINIMAL PAIR (or near minimal set). This is when you have
more than one difference in the words concerned, but the allowance is still there to compare the
suspicious sounds. The implication for analysis is still the same as of minimal pair, but with less
force. This is exactly the point made in the following data from Lewis (class notes)
[senɪt] „senate‟
[zenɪɵ] „Zenith‟
You will observe that the first and last sounds of the two words are different. But if we want
to compare the first sounds [s] and [z], on the ground that they are both at the beginning of the word
and are both followed by [e], we can, because the other area of difference (word-final position) is
too distant to possibly affect the suspicious sounds under investigation. This is called near minimal
pair.
Summary
Sounds are usually regarded as suspicious when they are similar in all but one phonetic features.
Phonological analysis is largely about investigating whether suspicious pairs/sets are the same or
different. Finally, suspicious sounds found in exactly the same environments belong to different
phonemes.
Post test
Why does a Linguist need to search for minimal pairs in his data?
Recommended Texts:
Pike, K.L. 1947. Phonemics: A technique for Reducing Languages to Writing. University of Michigan
Hyman, L.M. 1975. Phonology Theory and Analysis. New-York: Holt, Rinehart and Winston
Clark, J., C. Yallop and J. Fletcher. 2007. An Introduction to Phonetics and Phonology. Oxford:
Blackwell
Collinge, N. E. Ed. 1999. An Encyclopedia of Language. London and New-York: Rutledge.
LECTURE 7 Allophones
Introduction
Phonemes sometimes have contextual variants. We‟ll learn why and how to find the variants in this
lecture.
Objectives
Here, you‟ll learn:
what allophones are
why they‟re easy to find (they‟re always predictable)
the relationship between allophones, and between them and phonemes
Pre-test
If Pike (1947) says that sounds tend to be more like their environments, do you think that on the
basis of this, a sound occurring in three different environments can have three different
appearances?
„When the difference between two similar sounds is completely predictable...from the phonetic
context, we say that they are allophones of the same phoneme‟ Fudge (2005). Lewis (class notes)
defines allophones as different representations of the phoneme at the phonetic level. By this, we
mean allophones occur in phonetically predictable environments, and they are usually restricted to
such environments. See how Hyman (1975: 63) puts it: „An allophone is...a phonetic realisation of a
phoneme in a particular environment‟. Let me give you an example from English
[ pʰɪt] „pit‟
[ɪm pʷᴐ:t] „import‟
[ә plaɪ] „apply‟
[sɪp] „sip‟
In this data:
[pʰ] is aspirated because it occurs in a word-initial stressed syllable;
[pʷ] is labialised because it occurs before a rounded vowel;
[p] is not modified by its environment in the last two occurrences.
Because the first two occurrences are predictable we say they are only environmentally-
conditioned variants of /p/; they are allophones of /p/. Let me sound this warning however, it is the
three variants of /p/ attested in the data that allophones of the same phoneme (a phoneme the
appearance of which we do not know), so we choose [p] to represent that phoneme for reasons we
will talk about later.
Summary
Allophones are phonetic (predictable) variants of the same phoneme.
Texts
Fudge, Eric. 2005. Language as Organised Sounds: Phonology. In: N. E. Collinge, An Encyclopaedia of
Language. London & New York: Rutledge, 17-37
Hyman, L.M. 1975. Phonology Theory and Analysis. New-York: Holt, Rinehart and Winston
Clark, J., C. Yallop and J. Fletcher. 2007. An Introduction to Phonetics and Phonology. Oxford:
Blackwell
Collinge, N. E. Ed. 1999. An Encyclopedia of Language. London and New-York: Rutledge.
LECTURE 8: Sound Distributions
Introduction
After learning about allophones and their different contexts, you‟ll now learn in this lecture how to
state those contexts in “phonological terms”.
Objectives
At the end of this class, you should be able to state the alternations between allophones:
in prose; and
in phonetic notation
you should also be familiar with and use effectively terms like CD, PD, free variation, etc.
In lecture 4, I gave you Gleason‟s (1955: 261) definition of the phoneme as a „class of sounds which
are phonetically similar and show certain characteristic patterns of distribution in the language or
dialect under consideration‟; what this means is that if we have different sounds in a particular
group to be called a single phoneme, then we must understand where each of the sounds occurs.
Remember Unit 6 where we said when 2 sounds occur in the same environment and yield difference
in meaning, we are dealing with different phonemes. This should tell you that you that, under
normal circumstances, sounds that are members of the same phoneme will not occur in the same
environment.
Recall also unit 2 where we said sounds are modified by their environments. So each
member of the class of sounds known as phoneme occurs in a customised environment, where that
environment dictates its appearance. When it is another environment dictating the appearance of
another member of that class, it will have to conform to that environment, yielding another slightly
different appearance. Thus we can predict hoe a member of a phoneme will appear by looking at
what is in its environment. This is the idea I want to teach you under different sub-headings in this
unit.
[b β]
[b] [β]
#- v-v
[d ð]
[d] [ð]
#- v-v
„v‟ stands for „vowel‟, „#‟ stands for word boundary, while „-„ stands for the location of the sound
we are analysing.
Free Variation
When sounds confirmed to be in CD occur in the same position (of course with no change in
meaning), we talk of free variation. This is better said with data, and I now give you a very easy one
from Language K
Language K
gu ‘pestle’
atak ‘fight’
din ‘cook’
set ‘cutlass’
baluu ‘bird’
orᴐp ‘head’
g lәk ‘fish’
k lәk ‘fish’
You can see that except in the last two items in this data, the voiceless member of each pair
([g k], [d t], [b p]) always occurs at the end of words, while the voiced occurs at the beginning, which
is clearly a case of CD. Remember [g] and [k] are already established as allophones of the same
phoneme in this data; now they both occur in the same position, which seems to break our CD rule!
But there is something strange yet: the two sounds occur in the same position and still don’t affect
meaning, which means they’re not different phonemes; they are still allophones. This is what is
called allophones in free variation.
Neutralisation
Now recall what we mean by contrast. When the contrast between two or more phonemes is
suspended in a particular environment, we talk of neutralisation. Elugbe (personal communication).
In such situations, it is going to appear as if the phonemes concerned share a common allophone. In
other words, the data suggests that a single phone (allophone) represents two or more phonemes in
a particular environment.
Polish
Singular plural gloss
1 klup klubi club
2 trup trupi corpse
3 snop snopi sheat
4 zwup zwobi crib
5 kot koti cab
6 lut lodi ice
7 ko∫ ko∫e basket
8 nu∫ noʒe knife
Our suspicious pairs are [p b], in 1-4; [t d], in 5-6; and *∫ ʒ], in 7-8. When you analyse this
data in this light, you will see that the sounds contrast between vowels (in the plural column). Can
you see that?
Now zero in 1 and 2; there is this allophone [p] which is the variant of /b/ in word-final
position. And remember that the same sound has already been said to contrast with /b/. This means
two different phonemes share a common allophone:
/p/ has allophone [p]
/b/ has allophone [p] and [b]
The sharing of allophone [p] is what we call neutralisation. Now check this out on other
items in the data.
Summary
When suspicious pairs/sets do not occur in the same environment, they are said to belong to the
same phoneme; that is they are in CD. If they are found in the same environment, they are said to be
in PD, that is, they belong to different phonemes. Lastly, only allophones can be found in free
variation, n which case they are found in the same environment without any effect on meaning.
Post Test
What is the relationship between each of these pairs of sounds *t d+, and *∫ ʒ] in the Polish data
above?
In your own words, explain the relationship between CD and free variation
Recommended Texts
Fudge, Eric. 2005. Language as Organised Sounds: Phonology. In: N. E. Collinge, An Encyclopaedia of
Language. London & New York: Rutledge, 17-37
Hyman, L.M. 1975. Phonology Theory and Analysis. New-York: Holt, Rinehart and Winston
Ladefoged, Peter. (1967). Linguistic Phonetics 6. Working Papers in Phonetics, Department of
Linguistics UCLA.
Collinge, N. E. Ed. 1999. An Encyclopedia of Language. London and New-York: Rutledge.
Clark, J., C. Yallop and J. Fletcher. 2007. An Introduction to Phonetics and Phonology. Oxford:
Blackwell
LECTURE 9: Discovery Procedure
Introduction
In this class I want to further break down the steps to phonological analysis for you
Objectives
At the end of this class, you’d have learnt the step-by-step approach to identifying phonemes, and
be adequately armed to practice with any available data
Pre-test
I have already told you many of the items we will be discussing here in this unit: how to determine
phonemes and allophones, as well as distributions of sounds. Before you begin to read, make a list
of the steps you feel you should take in analysing sounds.
The procedure
The idea here is of the steps that a linguistic investigator needs to take in his field work to ensure
accuracy in his work. Here, you will notice yet again that knowledge of phonetics cannot be
discounted. This is because it all begins from hearing the sounds as said by your informants
(including those features that even the native speakers are not aware of, remember narrow
transcription?), and documenting them. If your ears fail to pick everything about each sound, your
analysis will be flawed. This also requires that you have all your phonetic symbols and diacritics at
your finger tips. So, let’s see the steps
1. Data collection. I’ve already explained this, but I need to add that you will do yourself a lot of
good if you use more than one informant. This will help you to eliminate idiolectal variances.
Also confirm every sound you feel you don’t hear well right there on the field. It does no
harm when your informants repeat certain items a number of times.
2. Identify every phone in your data. It is always better to list them on a plain sheet of paper
first. Here you will list [p], [pʰ] and [pʲ] separately, for instance. Every minute phonetic
variance leads to a different item to be listed.
3. Put the sounds you have identified in a phonetic chat. Lewis (class notes) advises that here,
you would need extra rows for such forms as secondary articulation, etc. And, of course, you
know that consonants will be in phonetic consonant chat while vowels will be in phonetic
vowel chat. Let me show you what I mean:
Phonetic consonant chat for language K
pb kg
pʰ kʰ
pʷ kʷ
sz
4. List all pairs of segments that suspicious because they are phonetically similar
5. List all pairs of segments that are not suspicious with any sound in your data
6. Among the suspicious pairs (or sets), look for minimal pairs (sets). If you get, and it is the
presence of the suspicious segments that result in difference in meaning, then the suspicious
sounds involved belong to different phonemes. But if there is no difference in meaning, the
sounds are likely to be allophones in free variation.
7. For suspicious pairs you do not find in minimal pair, describe the environments in which
each occurs.
8. Look for complementary distribution. If you get, then the sounds are allophones of the same
phoneme.
9. After determining that a set of suspicious sounds are allophones of the same phoneme (CD),
see the one that has the wider distribution. It is the symbol of the one with the wider
distribution that you should choose to represent the phoneme.
10. Try to understand the phonological processes that made the phoneme change to its
allophone(s) in their respective environments.
11. Write out rules for the phonological processes. Note that your rules are only formal
representations of distributional statements, for instance, if you distributional statement
says that a vowel becomes nasalised if it occurs after a nasal consonant, your rule (informal)
will look like: V V/N- .Don‟t worry, we‟ll talk more on this later.
12. Look for parallel distribution (PD) of suspicious sounds for which neither minimal pairs nor
CD has been found. If you get PD, then the sounds belong to separate phonemes, since they
occur in analogous environments.
13. Transcribe your data phonemically by replacing all the allophones with their respective
phonemes.
14. Present a phonemic chart of the sounds of the language under study. A phonemic chat will
contain ONLY the phonemes you have found by your analysis. These are the sounds you
establish for the language, which you will eventually use to design an orthography for that
language.
Let me point out to you that the procedures here may not be exactly what you find in other
materials, e.g. Lewis (class note) has 12, many other materials have even less, and you may be lucky
to find a few materials that have more. The whole idea is to give you a summary of the steps to take
when confronted with a fresh data (when your lecturer is not in sight). I believe that the more steps
we’re able to brake the process to, the more explanatory. You need not cram, only practice!
Summary
There are easy-to-follow steps to phonological analysis. Though these steps are not iron clad, they
provide you with a reliable guide, especially at this level of your exposure to phonology.
Post test
Get a data, and try this procedure out for yourself. If you’re stuck at any point, you may bring the
data for discussion.
Recommended Text
Pike, K.L. 1947. Phonemics: A technique for Reducing Languages to Writing. University of Michigan
Hyman, L.M. 1975. Phonology Theory and Analysis. New-York: Holt, Rinehart and Winston
Collinge, N. E. Ed. 1999. An Encyclopedia of Language. London and New-York: Rutledge.
LECTURE 10: Phonetic and Phonemic Transcriptions
Introduction
You learnt transcription in LIN 241. If you still know that (I expect you do), then you know phonetic
transcription already. But there is a transcription you still do after separating phonemes from
allophones. You’ll learn ‘why?’ in this class, as well as the difference between the two transcriptions.
Objectives
At the end of this class, you should be able to:
Do raw (phonetic) transcription
Separate phonemes from allophones
Do a kind of transcription having only the phonemes
Pre test
What is the difference between phonetics and phonemics?
Your phonetic transcription simply refers to your raw data. If in the course of data elicitation, you
notice that certain plosives tend to be produced extra puff of air than normal (aspiration), you have
to represent that in your transcription. If you notice that every time a vowel is adjacent to a nasal
consonant, the vowel tends to be nasalised (nasalisation), you will also reflect that observation in
your transcription. In other words, the raw data you take on the field captures every detail of the
speech of your informant.
You will then have such sounds as [pʰ], [a], and [tʷ] in your data.
Note that the diacritics, superscripts and other attached features make this kind of transcription
appear complex. Also note that phonetic transcriptions are put in square brackets [ ].
So the English transcriptions [pʰɪt] ‘pit’ and [tʷu:] ‘two’ are phonetic for instance because they
capture lip rounding and aspiration.
In essence, a phonetic transcription is also known as allophonic transcription, in the sense that
most times, features represented are mere allophones.
Exercise: provide ten transcriptions in your language that reflect such secondary features as lip
rounding etc. If you succeed, then you have done a phonetic transcription of you language in your
little way!
Phonemic transcription: for one, phonemic transcription is put in slants / / and it rarely has such
complexities as phonetic transcription. In other words, phonemic transcription is simpler than
phonetic transcription.
There are many reasons for this: one, it is not the raw data. Two, the raw data you take is passed
through phonological analysis such that features that are not significant would have been
eliminated. Remember that non-significant features are those features that do not affect meaning in
a language.
So, let us to the two transcriptions done above: [pʰɪt] ‘pit’ and *tʷu:+ ‘two’.
As for the transcription for ‘pit’, the presence of *ʰ] does not make any difference in terms of
meaning. That is, whether you say [pʰɪt] or [pɪt], you will still understand that there is no difference
in meaning. So, in phonemic transcription, we can do without it. We therefore have /pɪt/. Note that
it is now in slants. And it only contains features that affect meaning; those features not affecting
meaning have been dropped by during phonological analysis.
Note also that a phonemic transcription has fewer symbols than phonetic. Confirm this by
comparing the transcriptions we have given so far in this lesson, and then, get more examples from
your mother tongue as practice.
Let us conclude this chapter by relating our transcription to charts. Your phonetic chart of a
language has all the sounds recorded in that language whether the sounds are significant or not;
that is, phonemes and allophones. On the other hands, a phonemic transcription will not include
allophones.
Summary
Phonetic transcription is what you do when you record raw data on the field and is expected to
reflect every nuance of speech, which you’ll then go on to sift in the form of phonological analysis.
Phonemic transcription, on the other hand, is that which you do for only the significant sounds of
the language after your “sifting”.
Post test:
We have said previously in this class that features are sometimes predictable. Where will you put a
predictable feature; phonetic or phonemic chart?
Recommended Texts
Collinge, N. E. 1990. An Encyclopaedia of Language. London and New York: Routlege
Clark, J., C. Yallop and J. Fletcher. 2007. An Introduction to Phonetics and Phonology. Oxford:
Blackwell
Pike, K.L. 1947. Phonemics: A technique for Reducing Languages to Writing. University of Michigan
LECTURE 11. Phonological Processes 1
Introduction
I’m introducing you to the wide subject of phonological processes in this class. We’re going so
continue over the next few lectures. So, fasten your seat belts.
Objectives
One major point you must have learnt by the end of this class is why there are phonological
processes in language
Pre-test
What is the first premise of phonological analysis as given by Pike?
What are secondary articulations and why do they occur in language?
Phonological process is a term used to capture sound changes that result when sounds are used in
context. When transcribed, the English word electric has [k] as the final sound, but in electricity, a
word that is clearly derived from electric, the [k] changes to [s].
Why? You may ask? And the only answer I will give you in this class is that what changes [k] to [s]
is exactly what we mean by phonological processes. let me give you another illustration by means of
a data from Ghotuo.
Ghotuo
[ti] ‘push’
[tja] ‘push them’
Explanation: the *i+ sound in ‘push’ is a basic component of the word. But when another vowel is
added at its back, it changes to a glide, because somehow, the language does not want two vowels
to stand side-by-side in that environment. Now, the change from [i] to [j] is what we capture by
phonological processes.
Why do we actually have phonological processes in language?
Most phonological processes arise due to the need for ease of articulation. That is particularly
critical because the same limited set of articulators is used to produce all the sounds used in speech.
And in rapid speech, you produce so many sounds within seconds. So the articulators sort of blend
the the sounds in such a way that it is easier to keep up the required speed for speech. Schane
(1973: 61). This is what leads to the loss or change of some features. Let’s take the
electric/electricity example I gave you above.
In producing electricity, which has more sounds than electric, you’ll notice that *k+ changes to
[s], right? Knowledge of phonetics tells us that this is actually a stop becoming a non-stop. You still
remember that phonetic point?
Okay, this is why the articulators played that game on you: by adding the suffix ‘-ity’ to electric,
you have put [k] between two vowels. So the articulators just say “this guy want us to produce a
non-stop followed by a stop, and then followed by another non-stop. It’s not easy for us; we’ll blend
them and make the three sounds non-stops for ease.” So they keep your voicing feature for
instance, but change the stop feature (later I will tell you that they change [-cont] to [+ cont]) to non-
stop. And they say to themselves “at least this has not changed the meaning of the word; so he
should get used to it!”
How do you respond? You get used to it! Unfortunately, you never knew you’ve fallen for that
trick, except you’re a linguist like me!
So, this is how we come by phonological processes, though reasons may vary. We’ll talk about
reasons by and by.
Before we leave this lesson, let me introduce you to the categorisation of phonological
processes. There are four of them, according to Schane (1973): (1) assimilation; (2) syllable
structure; (3) weakening and strengthening; and (4) neutralisation. But I can tell you that more than
70 % of the processes you’ll find are assimilatory. So we’ll take it in the next lesson.
Summary
Most phonological processes arise in language because of the articulators push to ease articulation,
and this is all because sounds have to be used in context to communicate. This is however only the
introduction to this subject of discussion. We shall continue in the next lecture.
Post test:
Why do we have phonological processes in language?
Recommended Texts
Collinge, N. E. 1990. An Encyclopaedia of Language. London and New York: Routlege
Clark, J., C. Yallop and J. Fletcher. 2007. An Introduction to Phonetics and Phonology. Oxford:
Blackwell
Introduction
In lecture 11, I taught you why there’s phonological processes in language,, and I hope you all
understood. This lecture now focuses on those phonological processes that are context-driven; that
is, those that result as an effect of influence from the environment.
Objectives
At the end of this class, you should have learnt that:
Certain contexts spur specific phonological processes
And that this is language-specific
Pre-test:
If indeed sounds tend to be like their environments and all we can see in the environment of a
particular sound x are sounds; does it mean that that sound x is likely to become like one of the
sounds around?
Assimilation
Assimilation is a term used by some phoneticians to refer to ‘‘the contextual variability of
speech sounds, by which one or more of their phonetic properties are modified and become
similar to those of the adjacent segments’’ Farnetani (1997: 376), quoted in Laniran and
Clements (2003). This is just a deeper way of saying what Pike (1947) said in his first premise
of phonological analysis, “sounds tend to be more like their environments”. Let me explain
this with what I expect you did in O’ Level Government or History. The French colonial policy
was that of assimilation, in which citizens of African French colonies are absorbed into
France by the offering of France’ citizenship. They were also made to speak French, which
makes them appear as if they were real French men and women. My God! I hope you still
remember. So by that relationship with France which also manifests in their language, we
say that they were assimilated to France, and have the citizenship of France. Assimilation in
phonetics is similar, only that we are not talking about human relationship; it is sound
relationship. So we say there is a kind of relationship that will happen between sounds and
one will become like the other. So the example of electricity I gave you in the last lesson in
which [k] becomes a non-stop because the sounds surrounding it are non-stops (vowel) is
actually an example of assimilation!
In Yoruba, every time a vowel comes after a nasal consonant, the vowel becomes
nasalised. This means the vowel is becoming like the nasal consonant in terms of nasality
(the vowel is assimilating to the nasal feature!). Check this out in the following data.
Yoruba
[ iʃan ] „Matches‟
[ n ] „Yesterday‟
[ m ] „Knowledge‟
[i ] „Pounded yam‟
Types of Assimilation
Assimilation can be of 2 types: partial, and complete. In the Yoruba data we just gave you,
vowels only assimilate 1 feature (nasality) from the nasal consonants preceding them. I
expect you to agree with me that only 1 out of the many features of nasal consonants is
partial. Assimilation can however be complete if a sound changes entirely to become exactly
like another in its environment. Let’s look at a hypothetical language.
Language K
/ilouo/ [ilouu]
You will see from this that rather than /o/ merely becoming more like /u/, it just
become exactly like it. This is complete assimilation!
Direction of Assimilation
Assimilation can be progressive (perseverative) or regressive (anticipatory) in nature.
Progressive assimilation is when the feature is being transferred forward, while regressive is
when the feature is being transferred backwards. Hey! Don’t get confused. I will explain.
Take another look at the Yoruba data above; you will see that the nasal consonants are
transferring their nasality feature to sounds on their right, that is, sounds coming after
them. That is progressive assimilation. On the other hand, you will see that in the following
English data, the feature is being transferred the other way: backward, that is, to preceding
sounds.
English
[win+ ‘win’
[piŋk+ ‘pink’
[h :n+ ‘horn’
This English data therefore illustrates regressive nasal assimilation.
Finally, there are many types of assimilatory processes. They include: palatalisation,
velarisation, labialisation, glottalisation, nasalisation, voicing etc. We will discuss these as
part of the next lesson.
Summary
If the leaf stays too long on the soap, it becomes soap itself. Or, when a sheep keeps
company with a dog, it will end up eating faeces. In the same vein, when sounds stay
together in certain contexts, they rub off on one another. This is what we mean by
assimilation. Depending of the nature of the effects sounds have on one another, there are
different names for assimilatory processes.
Post test:
Find a word to describe the process by which a sound that in not nasal becomes nasalised.
Recommended Texts:
Laniran, Y. O. and G.N. Clements (2003). Downstep and high raising: interacting factors in
Yoruba tone production. Journal of Phonetics 31 (2003) 203–250
Pike 1947. Phonemics: A technique for Reducing Languages to Writing. University of Michigan Press
Clark, J., C. Yallop and J. Fletcher. 2007. An Introduction to Phonetics and Phonology. Oxford:
Blackwell
LECTURE 13 Environments for Phonological Processes
Introduction
In this lecture, I’ll go deeper in the idea I started in the last class, that there are specific
environments for specific phonological processes.
Objectives
At the end of this class, you are expected to have learnt:
How to relate phonological processes to phonetic environments supporting them.
Types of phonological processes other than assimilation
Pre-test:
What do you still remember about secondary articulation, which you did in LIN 241?
I mentioned in lesson 11 that there are other types of phonological processes, namely syllable
structure, weakening and strengthening, and neutralisation. I also did say that all these other
phonological processes are like 30% of phonological processes attested in language. I will introduce
you to them in this lesson and supply data to demonstrate them. Thus this lesson will be in 4 parts:
assimilatory processes; syllable structure processes; weakening and strengthening processes; and
neutralisation.
Here, we will discuss some of those phonological processes that result as a result of sounds
becoming like each other, or simply put, being influenced by their environments.
i. Nasalisation: this results when a sound takes on a nasal feature from another sound
adjacent to it. Check the Yoruba data that I gave you in lesson 12 above; it’s a perfect
example of nasalisation. The diacritic for nasalisation is a tilde [ ] placed on the nasalised
sound, like this [a], which is a nasalised [a]. Environment: a sound will get nasalised
either before or after a nasal sound.
ii. Labialisation: Ckark, Yallop and Fletcher (2007: 63) define labialisation as “the addition of lip
rounding or lip protrusion to any sound which is normally articulated with the lips in a
neutral or spread position”. Have you ever observed that when you produce the English
[b] your lips are neither rounded nor protruded? Now try it. Good! But now, also
produce [bʷu:k+ “book”. You will see that while producing the very first sound, your lips
are already protruding and assuming rounding position. That’s labialisation! As you can
see, [b] becomes labialised because it is occurring in the environment of [u:] which is a
vowel produced with rounded lips. The diacritic for labialisation is a raised [ʷ] in front of
the labialised sound.
iii. Palatalisation: Lewis (n.d.) defines palatalisation as “the superimposition of a raised front of
the tongue to another place of articulation”. Now, don’t forget that the front of the
tongue is the place or articulation for the palatal sound. So, that’s the logic. And it
happens when a sound (not palatal - palatal sounds cannot be palatalised), is in the
environment of a high front vowels [i], [ɪ], [y], [ʏ], and in some languages, [e]. The
diacritic for palatalisation is a raised [ʲ] in front of the palatalised sound. Lewis (n.d.)
has an interesting data from Bwatye Language. Let‟s look at it together.
[ɓʲelto] “boundary”
[ɗʲekto] “hen”
mʲent ] “life”
You can see that a consonant becomes palatalised before [e] in this language.
iv. Velarisation: this happens when the back of the tongue is raised at the time of production of
another sound (which is not itself a velar sound – velar sounds cannot be velarised).
This is because the configuration for the velar sound interferes with the production of
that sound, and it goes on to influence the quality of the sound so produced. The
diacritic for velarisation is a raised symbol of a voiced velar central fricat
[~] across the symbol of the velarised sound, like this [t ] or [t]. The environment is the
environment of high back vowels.
well, you‟ll see that the back of the tongue is raised while producing the [l].
v. Voicing: this is when a voiceless sound becomes voiced in the environment of a voiced
sound. This is sometimes called Intervocalic Voicing because very often, it occurs when
a voiceless sound comes between two vowels. Let‟s look at this English example:
/haus/ “house”. When you pluralise this word, what you get is [hauziz] “houses”. This
shows that the word-final /s/ in house, has become [z] in “houses” because it is now
between /au/ and /i/.
These are phonological processes that alter the structure of the syllable of languages. Schane
(1973) says they affect “the relative distribution of consonants and vowels within the word”.
Schane goes on to say that “for example, a cluster of two consonants could be simplified in one
of three ways: one of the consonants could be deleted, a vowel could be inserted between the
two consonants, or the two consonants could coalesce into a single segment”. Let me give you
an example from a major Nigerian Language, Yoruba. In Yoruba, the syllable structure forbids
consonant clusters. The structure also forbids a consonant ending a syllable. So when a typical
Yoruba speaker wants to pronounce the English word [brᴈ:d]“bread” (which is only 1 syllable
having CC C structure), he will say [b r d ]. Don‟t be afraid – CCVC means consonant,
consonant, vowel, consonant, which is what you have in the transcribed word!
Now, can you see what I‟m seeing? Our Yoruba speaker has inserted [u] between [b] and [r]
which has altered the syllable structure of English, but now conforms to that of Yoruba. Our
speaker also inserted [i] at the end of the word. By the insertion of these two vowels, our
speaker has turned just one syllable to three b – r - d ! What a serious alteration, you would
say!
Let me now illustrate another type of syllable structure process with a popular biblical name
– Moses (having 2 syllables, mo - ses. A typical Yoruba speaker calls this Mose, still preserving
the 2 syllables, but deleting the consonant at the end of the second syllable. This is vowel
deletion.
One major type of syllable structure processes is glide formation. Let‟s use our hypothetical
language to explain this.
Language K
fru „milk‟
ilre „cow‟
Here is the explanation of the data: the word for milk has a word-final vowel, while the
word for cow has a word-initial vowel. So to form cow milk, putting the word for milk first,
followed by the word for cow, we find ourselves with fru + ilre. A sequence not permitted by
the structure of the language. What do we do? Since the language does not forbid consonant
clusters, we convert the first vowel to a consonant so that we avoid the erring vowel cluster. The
consonant is however, the closest to vowels (a glide) among all the consonants.
The process of converting a vowel to a glide ([w] is a glide) is called glide formation, and it
is done to preserve the syllable structure of the language. So, it‟s a syllable structure process.
Let me emphasise that the motivation for syllable structure processes is to preserve the
syllable structure of the language; because this is necessary to ensure correctness (you still
remember deep structure of your syntax?). This is done by breaking up unwanted clusters of
consonants and vowels one way or the other.
Part 4: Neutralisation
Neutralisation occurs when the contrast between two or more sounds gets lost. Try and follow this
exposition: English has phonemes /s/ and /z/, right? Okay! But the plural marker in English is usually
/s/, except in certain specific circumstances where you have [z] or nothing at all (there are other
variants which are basically morphemic). Now the problem is with this variant [z] which sometimes
stands in the place of /s/ as the plural marker in words like boys [b ɪz]; it is still /s/ (when you
transcribe it phonemically!).
My whole story is that /s/ sometimes appears as [z], but this appearance can be confused
with the other phoneme /z/ which contrasts with /s/ in the minimal pair /su:/ ‘sue’, /zu:/ ‘zoo’. Every
time a sound appears such that it can ordinarily confused with another phoneme, we say there is
neutralisation – the absence of that feature that makes them physically different.
We will spend a lot more time on this issue in LIN 331. So, for now, let’s keep it in the cooler.
Summary
One good thing about phonological processes is that each of them is spurred in specific
environments, and by our knowledge of phonetics, we easily know why. That means if you can
master the environments, phonological analysis becomes child’s play. For syllable structure
processes; once you know the syllable structure of a language, the sequences allowed, and those
disallowed, it becomes easy to know why syllable structure processes happen: to force erring
sequences to conform.
Post-test
Collect data on borrowed words in your mother tongue and determine how erring syllables are
forced to conform to the structure of the language
Recommended Texts:
Lewis. Nd. Lecture notes on LIN 241 (production of Speech) and LIN 232 (Basic Phonology).
University of Ibadan
Hyman, L.M. 1975. Phonology Theory and Analysis. New-York: Holt, Rinehart and Winston
Clark, J., C. Yallop and J. Fletcher. 2007. An Introduction to Phonetics and Phonology. Oxford:
Blackwell
Clark, J., C. Yallop and J. Fletcher. 2007. An Introduction to Phonetics and Phonology. Oxford:
Blackwell
LECTURE 14: Phonological Rules
Introduction
In this class, I will explain to you how to reduce the prose form of analysis to phonetic notations. You
can be sure that you will do more works in this class (or after) than in others. This is because I’m to
just lay the foundation for you, and from now on, you’ll be expected to write rules for every analysis
you do. The rules we’ll do are however informal rules.
Objectives
By the time we finish this class, you’ll be expected to be able to write phonological rules for any
phonological process you come across.
Pre-test
Try and use only symbols to state the change of /k/ to [s] in the English pair electric/electricity.
We’ve dwelled so long on the process of phonemic discovery and the processes that condition the
alternation between a phoneme and its allophones. But it has all been talk, talk and
talk...explanation all the way. There are, however, very short, notational, ways of expressing these
processes. These involve the use of notations (or, if you like, say symbols). However, as short as the
symbols are, Clark, Yallop and Fletcher (2007) say “a phonological rule states that a certain class of
segments undergoes a change in some particular environment”.
/ (morpheme boundary)
{} braces (used to abbreviate sequences of partially similar rules)
Now, before I begin to guide you in the use of these symbols in constructing rules, let me
state that, basically, we‟re going to concern ourselves with only informal rules, at this stage.
So, we begin to demonstrate the use of the symbols: if you want to say, as we have in the
electric/electricity example, that [k] changes to [s], you will write:
1. [k] [s]
If you want to add the environment in which this takes place, you add „/‟ after [s], which means
“in the environment of”. After the „/‟, you will now add V- V, which means between two vowels
(remember that „-‟ indicates the exact environment of the changed sound).
Thus, we have:
2. [k] [s]/ V – V
This means [k] becomes [s] when it‟s between two vowels (remember that when [k] occurs
word-finally in electric it remains [k]. It is when you add another vowel at the back that it changes).
The English word /b :l/ “ball”. If a typical Yoruba man (I mean a village man) wants to say
this, he will say [b lu]. Many things have happened here, but I will only choose the insertion of a
vowel at the end of the word for the next rule.
In the English word, there is no vowel at the end; so in your rule, you say ø „null‟ (that is
“nothing”) becomes a vowel at the end of the word. Here is it:
3. ø V/ - #
4. ø [u]/ - #
Let‟s take the word Moses (I leave you to transcribe it). Yoruba deletes the word-final [s].
5. [s] ø/ - #
Do you still remember that Yoruba data in which oral vowels become nasalised after nasals? It‟s
in unit 5. Let‟s write a rule for it:
6. /N –
Braces are used when we want to conflate two or more rules that are structurally related. Let us
attempt to conflate rules 4 and 5 in 7 below:
7.
You can now see that phonological rules can be used to explain every phonological process;
they can explain sound change, insertion, deletion, coalescence, etc.
Summary
Phonological rules are convenient notational ways of presenting phonological analysis, which can
then be used to make the description of language easier by just stating the rules, rather than
repeating the process every now and then.
Post-test
Now revisit all the data we‟ve treated in this course and write rules for all the processes. a cheap
test, actually!
Recommended Texts
Schane S. 1973. Generative Phonology. New Jersey: Prentice-Hall Inc.
Hyman, L.M. 1975. Phonology Theory and Analysis. New-York: Holt, Rinehart and Winston
Clark, J., C. Yallop and J. Fletcher. 2007. An Introduction to Phonetics and Phonology. Oxford:
Blackwell
Introduction
In this final class, I want to emphasise the fact that some phonological processes are phonetically-
motivated, and so easily explained and understood. Also, sounds may be grouped based on features
they share and processes they undergo in similar environments, a situation having simplifying effect
on analysis. We will look into this in this lecture.
Objectives
After you’ve gone through this class, you’ll be expected to:
Know what it means for a phonological process to be natural
Know the motivation for, and the good things about natural classes of sounds
Pre-test:
Which of these 2 statements makes sense to you?
- [a] becomes nasalised after a nasal sound
- [a] becomes nasalised after a voiceless bilabial fricative
Try to explain whichever you choose.
Naturalness in phonology is a topic that has many parts. One deals with natural classes of sounds,
another links the natural classes with features uniting those sounds in the classes. Yet another part
deals with the phonetic naturalness of phonological processes, while an extension of this deals with
the naturalness of rules. But at the end of the day, we’re still talking about the same thing:
naturalness.
So, we take it one after the other. Sounds can be grouped into a natural class if they are more
closely related than each of them is to any other sound in the language. One good advantage of this
is that the more natural a class of sounds, the fewer the features needed to characterise them. This
means the easier our phonological analysis on the long run. Hyman (1975: 139-140) lists some
criteria for grouping sound into natural classes. Here is a paraphrase of those criteria:
a. If sounds undergo the same phonological rules;
b. If they both function in similar environments of phonological rules;
c. If one segment is converted into the other segment by a phonological rule;
d. If one segment is derived in the environment of the other segment.
He says sounds can be grouped into natural classes if one or more of these criteria are met. A
partial summary of this is that sounds belonging to a natural class undergo similar processes in
similar environments. Thus, in German, all voiced obstruents undergo devoicing in word-final
position. That is, [b], [d], and [g] for instance do not occur in word final position in the language.
Rather they change to [p], [t], and [k] respectively. So, in that language, instead of analysing these
voiced obstruents one after the other, you can just say you’re analysing the class, voiced obstruent,
in which case, you’ll end up doing only one analysis for all of them, rather than many analysis
repeating the same thing.
Now, look at [b], [d], and [g] with your phonetic eye. They all all plosives and voiced, and the
sounds they change to are also all plosives, but voiceless. This means that the sounds changing are
united in their manner of articulation and that unity is unique to them alone (they don’t have that
relationship with [m] or [l] for instance) in the inventory of German sounds. Now see that the sounds
being changed to are also united in their plosive feature (they also share this feature with their
voiced counterparts from which the change is coming). In addition, they are united by voicelessness.
And this unity of features reflect in the unique process they go through. Sounds not belonging to
their class do not go through the process. Now, the class may be expanded or reduced, depending of
phonetic considerations like features they share with another group of sounds on another level.
We’ll do more of this when we begin to talk of features in detail in LIN 331.
In another sense, naturalness is about phonological analysis making phonetic sense, being
logical. And this actually goes beyond phonology. Collinge (1990: 457) says “naturalness is
considered to be a property of human language in general”. When I tell you that a vowel becomes
nasalised in the environment of a nasal sound, you understand that there is something in the
environment to warrant that. That is therefore expected, going by the phonetic reality of the
environment. But when I say that a vowel gets nasalised in the environment of an [l], you should ask
what is it about [l] that should warrant this kind of change. Nothing. This means that this latter
process is not natural.
One more example before we leave this discussion: when I say that a consonant becomes
palatalised before [ɪ], you should understand that the high and front features of [ɪ] are capable of
causing that (the area of production of [ɪ] is around the hard palate, so palatalisation occurs when it
transfers features from that area to neighbouring sounds). But this argument will not do if I say
palatalisation occurs in the environment of [a] ([a] is a low vowel and is not produced anywhere
near the palata area of the oral cavity).
Lastly natural rules are rules written for natural phonological processes, while a phonological
process that is not phonetically plausible will not have a natural rule. But bear in mind that not all
rules are natural, actually. We‟ll return to that by and by.
So much for naturalness for this level; I assure you you‟ll return to it again and again until you
graduate.
Summary
Sounds belonging to natural classes also undergo similar processes. So establishing natural classes
makes analysis considerably easier. Naturalness also mean that phonetically plausible phonological
processes will also have phonetically plausible rules.
Post-test
Which of these 2 statements makes sense to you?
- [a] becomes nasalised after a nasal sound
- [a] becomes nasalised after a voiceless bilabial fricative
Try to explain whichever you choose.
Recommended Text:
Collinge, N. E. Ed. 1999. An Encyclopedia of Language. London and New-York: Rutledge.
Ohala, J. J. 1990. Naturalness in Phonology.
www.linguistics.ucla.edu/people/.../03_naturalness_ ordering.pdf
Hyman, L.M. 1975. Phonology Theory and Analysis. New-York: Holt, Rinehart and Winston
Clark, J., C. Yallop and J. Fletcher. 2007. An Introduction to Phonetics and Phonology. Oxford:
Blackwell