Chapter 3: Phonology: Contrast and Patterns

- Phonology is the study of how languages organize speech sounds into a system of patterns and contrast.
- There are three levels of phonological analysis.
o A level that is concerned with the segments where we study the relationship between individual
speech sounds.
o A level that is higher than segments where we study how sounds are organized in syllables.
o A level that is lower than segments where we analyze the features that build up the segment.

3.1 Segments
There are phonetic differences between segments. Some differences are easy to notice like the one between [n]
and [ŋ] in win [wIn] and [wIŋ]. Other differences are less obvious as the [n] in one time [wʌn] and one thing
[wʌn̪]. Such a difference does not change meaning unlike the first example.

3.1.1 Phonemes and Allophones

Phonemes are sounds that are organized in contrastive units based on their phonetic properties and if these
properties can distinguish words. Since the difference between [wIn] and [wIŋ] makes a difference in meaning,
we say that [n] and [ŋ] are different phonemes. Phonemes are represented in slashes \ \.
Allophones are the different variations of the same speech sound. These variations do not change meaning and
they are only adjustments to the sound because of a neighboring sound. So, the difference between [n] and [n̪]
does not change the meaning, we say they are allophones of the same phoneme. [n̪] is dental (rather than alveolar)
because it came before the dental sound [θ] in ‘one thing’. Allophones are represented in square brackets [ ].
Another example in vowels: bit [bIt], beat [bit] → /i/ and /I/ are phonemes, but [i] in beat and [i:] bead [bi:d] are
allophones.

3.1.2 Minimal Pairs

A minimal pair consists of two words that differ only by one segment in the same position. This one difference
should change the meaning. So, the words win [wIn] and wing [wIŋ] form a minimal pair, and also the words
bit [bIt] and beat [bit] form a minimal pair. Minimal pairs show us that two sounds belong to different
phonemes.

3.1.3 Complementary Distribution

This is related to the distribution of allophones when two sounds occur in a non-overlapping and mutually
exclusive environment. As in the example above, [n] and [n̪] are allophones that are distributed differently where
[n̪] occurs only before dental consonants while [n] occurs elsewhere. We say that those two sounds [n] and [n̪]
are in complementary distribution. Also, [i] and [i:] in the other examples are in complementary distribution as
[i:] occurs only before voiced consonants while [i] occurs elsewhere. So, they are allophones of the same phoneme
/i/.

1
3.1.4 Differences in Phonemes across Languages
Sounds that contrast in one language may not contrast in another. For example, the difference between [æ] and
[ɛ] is phonemic in English but it is not phonemic in other languages like Turkish and Arabic (although the
difference may exist, but it does not change meaning).
Conversely, sounds that do not contrast in English like long and short vowels may contrast in other languages
again like Turkish and Arabic.

3.1.5 Things to watch for:

Sometimes, we cannot find minimal pairs for every sound in a language, so we may rely on near minimal pairs.
It is when there is more than one difference between two words, this difference should not be in the immediate
environment of the targeted sounds we want to contrast. In other words, the sound before and after the contrasting
sounds should be the same.
Examples: mission [mIʃǝn] and vision [vIƷǝn] are near minimal pairs, and we say that /ʃ/ and /Ʒ/ are contrasting
phonemes, and also /m/ and /v/. But, cut [kʌt] and put [pUt] are not near minimal pairs.
Sometimes, we cannot even find near minimal pairs for two sounds, so we should rely on the phonetic similarities
between them as well.

3.1.6 Differences in the Distribution of Allophones across Languages

The distribution between allophones can also vary from language to language. It is common to have a nasal
allophone for vowels when there is a nasal consonant nearby. However, vowels in English are nasal only before
nasal consonants as in ‘on’ and ‘bank’, not after them as in ‘no’ and ‘neat’. However, in other languages, like
Scots Gaelic, vowels are nasalized on both sides.

3.1.7 Phonetic and Phonemic Representation

As we saw before, a language can have contrastive phonemes (that are formed by minimal pairs) and then
allophones for each phoneme that are in complementary distribution. We can choose to represent the word
phonemically or phonetically.
Phonemic representation / / includes only the information that are essential to the meaning. It is like mental
categories that exist only in the mind. Phonetic representation [ ] includes information about phonetic details that
are produced in actual articulation like aspiration and nasal vowels.

3.2 Syllables
We can divide words into syllables which are another type of phonological units. Speakers of a language can
subconsciously divide words into syllable, but they do not know how important it is. We use a dot to mark the
boundaries of syllables.
Examples: exit /ɛk.sIt/ decline /dI.klajn/

2
3.2.1 Types of Syllable Patterns
Syllable Patterns are represented by C (consonants) and V (vowels). Some languages like Hawaiian allow only
(C)V1. Thus, the pattern of their syllables only have V or CV.
Example: aloha ‘greetings’ /a.lo.ha/ = V.CV.CV

Korean allows more complex syllable structure (C)V(C) which are V, CV, VC, CVC. French is more complex as
it allows two consonants at the beginning and the end, so there are additionally CCV and VCC.

English has even more complex syllable structure with three consonants at the beginning (as in stream /strim/)
and up to four consonants at the end as in (sixths /sIksθs/). However, only few words have such a complex
structure, and many constrains apply to such structures. To have three consonants at the beginning, the first must
be /s/, the second must be a voiceless stop and the third a glide or a liquid.

These constrains are called phonotactics which refers to the branch of phonology that is concerned with
permissible combination of phonemes. These phonotactics are language-specific.

3.2.2 Syllable Structure

Words are composed of syllables, which is in turn consist of an onset and a rhyme. The rhyme consists of the
nucleus of the syllable and a coda. Languages differ in the type and number of sounds they allow in the nucleus,
onset and coda. For example, English allows vowels and sonorous consonants in the nucleus while other
languages do not. For the onset, some languages require an onset, but all languages allow at least one consonant
in the onset, like Korean. French allows two while English allows three. For the coda, some languages ban coda.

Building Syllable Structure

There are three steps to build syllables of words. First, we start with the nucleus formation where we link each
vowel in the word to N, then link each N to an R then to the syllable (σ). Second, we start the onset formation
by linking as many consonants to the left as possible in the language to an O head, and we link it directly to the
syllable. Finally, we start coda formation by linking the remaining consonant at the right side to a C then to R
and then to the syllable. Syllables that have a coda are called closed syllables while those with no coda are open.

1 Putting parenthesis means that the elements inside are optional.

3
3.2.3 Basic Syllables
There are two basic requirements that we should fulfill when we build the structure of the syllable.
First, there is the Sonority Requirement which states that in basic syllable, sonority rises before the nucleus
and declines after the nucleus. Remember that vowels are the highest in sonority (that is why they are the
nucleus), followed by glides, then liquids then nasals and finally obstruents (oral stops, fricatives and affricates)
which has zero sonority.
Second, there is the Binarity Requirement which states that within the basic syllable, each constituent can be
at most binary (branching into two).

3.2.4 Syllables with a More Complex Structure

English allows more than two consonants in the onset (actually three) and more than two consonants in the coda
(up to four) which violates the Binarity Requirement. Also, in words like stream /strim/, the /st/ are not raising
in sonority but they are both zero, which violates the Sonority Requirement.
Such complex onsets and coda are rare and there are special rules for them. First, the extra consonants are
always at the edges (as in stream and ranks). Second, the extra consonant in the onset (in English) is always /s/,
and in the coda, the extra consonant is always voiceless and made with the tip of the tongue (as in ranks,
clamped, twelfth). Also, note that these extra consonants are added as suffixes. We link these extra consonants
directly to the syllable.

3.2.5 Stress and Syllables

Syllable structure affects stress placement which is (in English) sensitive to the syllable weight. The weight of
the syllable is determined by its rhyme. Heavy syllables consist of a nucleus and at least one consonant in the
coda while light syllables have only a nucleus.
For verbs with more than one syllables, we put the stress on the final syllable if it heavy, otherwise, we put it on
the second-to-the-last syllable. Examples are like arrive /ǝ.ˊrajv/ and hurry /ˊhʌ.ri/. For nouns with more than
one syllables, we put the stress on the second-to-the-last syllable.

3.2.6 Syllable-Based Phonology

Syllables play a major role in many of the phonological processes. Here we have two examples; aspiration and
vowel length.
Aspiration in English: English voiceless stops are aspirated syllable-initially. Examples, take [tʰek] and attack
[ǝ.ˊtʰæk].
Vowel Length in English: English vowels are lengthened when followed by a voiced obstruent in the coda
position of the same syllable. Example, obey [ow.bej] and lobe [lo:wb]).

4
3.3 Features
They are the basic building blocks of speech sounds.

3.3.1 Features as Phonetically Grounded Elements

As mentioned before, speech sounds are produced by different articulatory activities like voicing and tongue
position. For example [p] is voiceless bilabial stop and its production involves spreading vocal folds and
pressing the lips. So, we use the features [LABIAL] and [+ voice]as building blocks for the [p]. Using features
have many advantages. Here we mention three.

1. Features and natural Classes

Using features allow us to characterize natural classes in an economical way. Natural classes are groups of
sounds with similar properties. A natural class needs fewer features to define than to define any of its members.
Natural classes are important because members behave alike (e.g. only voiceless stops are aspirated).
Features of English obstruents
- sonorant
- voice + voice
ptk bdg - continuant
fsʃ vzƷ + continuant

2. Features and Contrast

Using features allows us capture phonemic contrasts. So, it is not enough to say that the phoneme [p] contrast
with [b]. We need to say that all voiceless obstruents contrast with voiced obstruents. Thus, the feature [voice]
is a distinctive feature. Also, we use the feature [continuant] to distinguish stops and fricatives. Other features
show other contrasts as well.
3. Features, Process and Allophonic Variation
Using features allows us to describe the environments of allophones better. So, for vowel nasalization, we say
that vowels become [+ nasal] in front of a [+ nasal] consonant in the same syllable.

3.3.2 Features Representations

Now we define the features we need to analyze sounds in languages.
Defining the Features of English
Most features have labels that reflect articulatory terms such as [voice] while fewer features have less familiar
terms such as [COROAL]. Most of the features are either ‘+’ or ‘-‘whereas three features (LABIAL,
CORONAL, DORSAL) are written in upper case and have only one value.
▪ Major class features
They are the features that represent the classes consonant, obstruent and sonorant (nasal, liquid, glide, vowel).
[+ consonantal] all consonants are + except for the glottals, glides are vowels are -
[+ syllabic] vowels and syllabic liquids and nasals are +
[+ sonorant] vowels, glides liquids and nasals are +, obstruents are -

5
 ▪ Manner Features
They are the features that represent manner or articulation. They distinguish stops and affricates from stops
[continuant], affricates from stops [DR] and nasals and laterals from other sounds.
[+ continuant] vowels, glides, liquids and fricatives are +
[+ delayed release] ([+ DR] only affricates are +
[+ nasal] only nasals and nasal vowels are +
[+ lateral] only [l] is +

▪ Laryngeal Features
They are the features that represent laryngeal activities. They include voicing, aspiration and glottal sounds.
[+ voice] all voiced are + and all voiceless are -
[+ spread glottis] ([+ SG]) all aspirated consonants and [h] are +
[+ constricted glottis] ([+ CG]) only [Ɂ] is + in English

▪ Place of articulation features

They are the features that represent supra-laryngeal activities (above the larynx).
[LABIAL] this feature represents the labial articulator (lips).
[+ round] rounded vowels and [w] are +
[CORONAL] produced by the tip or the blade of the tongue
[+ anterior] interdentals and alveolars are + but alveopalatals are -
[+ strident] the noisy fricatives and affricates are +
[DORSAL] produced by body or back of the tongue including vowels.
[+ high] vowels and velar and palatal consonants are +
[+ low] only low vowels are + (not needed for English consonants).
[+ back] palatal consonants (and behind) and back vowels are +
[+ tense] only tense vowels are +
[+ reduced] only [ǝ] is +

Note: diphthongs are treated as a vowel-glide sequence.

3.4 Derivation and Rules

We learned the difference between phonemic and phonetic representations. The rules (of devoicing, aspiration
and etc.) are applied to the phonemic representation (input) to derive the phonetic representation (output).
The Form of Rules
There is a formal way to express the rules by using symbols and features. For example, the rule of nasalization
is as follows.
V → [+ nasal] / ___ + cons
+ nasal