Phonology is where you put into practice all you’ve learned in phonetics. It is the study of how sounds are strung together (phonotactics), how they interact with each other, and the rules that account for these processes.
The focus of phonology at an introductory level course can be categorized into the following areas.
1. The Distribution of Phonemes
Phonemes are individual sounds described as a bundle of phonetic features that differ from each other in at least one feature. For example, we know that /p/ and /b/ are both bilabial plosives and yet they are not identical. In order to understand how they are distinguished from each other, we can create a simple feature chart, marking features with binary values as ‘+’ or ‘-‘. In doing so, we discover that the distinguishing feature is voicing. This is referred to as a minimal distinction or a distinctive feature.
Distinctive features enable us to classify phonemes into categories, distinguish classes of phonemes from one another, formulate predictions as to how classes of phonemes will behave, and display language-specific constraints. They also provide articulatory, perceptive (auditory), and acoustic information that help us distinguish between closely related phonemes and allophones, provide categories for showing contrast between sets of sounds, and are the basis for creating natural classes.
Non-distinctive features are those that are predictable for any sound. For example, in Standard English, voiceless stops are aspirated when they are the first sound of a word (word initial) or when they are in the onset position of a stressed syllable. The distribution of the feature [aspiration] is predictable or non-distinctive since there is a rule that determines where it will surface. Thus we can predict that /k/ in the word /ki/ ‘key’ will be aspirated, [khi]. Aspiration is not a distinctive feature since, when aspiration is added to /k/, it does not create a different phoneme as in the case of (1) with voicing.
Phonemes can be expressed in phonemic form or phonetic form. Phonemic form (also referred to as the underlying representation) is that which exists in the mind of the speaker, while phonetic form (also referred to as the surface form) is that which is actually articulated by the speaker.
Levels of Representation
Every language has its own rules that permit or prohibit various interactions. This knowledge is intuitive to native speakers; to such an extent that most people are not aware of what is really coming out of their mouths! A speaker of any language has an idea, an abstract concept of the sounds s/he articulates. However, more often than not, there is a disparity between the form of the sound that is in the mind of the speaker (underlying representation or UR), and the form that is actually articulated (surface form or SF).
In other words, phonemes are stored in the mind as a bundle of features.
This abstract notion of a sound, referred to as the underlying representation, is surrounded by back-slashes / /. The surface form that is realized as articulatory gestures, and articulated according to phonological environment, is surrounded by brackets [ ].
Allophones are variants of a phoneme that surface under specific phonological conditions. Here is an example from Standard English.
/t/ is a voiceless alveolar stop. We use this sound in words such as ‘top’, ‘train’, ‘stop’, ‘butter’, and ‘button’. But do we really? As discussed earlier, when /t/, as a voiceless stop, is the initial sound in a word, it bears the feature [+aspiration], as in the word ‘top’. Thus a speaker believes s/he is articulating the word as /tap/ when the surface form is actually [tʰap].
Consider the word ‘train’. This word initial /t/ is affected by the following /ɹ/. It thus surfaces as [tʃ] and will be transcribed as [tʃɹejn].
When a /t/ is found in the middle of a word (word medial) as in ‘butter’, it has yet another pronunciation. It will surface as an alveolar flap or [ɾ], as in [bʌɾəɹ], or as a glottal stop in words such as ‘button’ [bʌʔən].
Interestingly, the only phonological environment in which /t/ actually surfaces as [t] is in a consonant cluster such as in the word ‘stop’ [stap].
These various surface forms (SF) of /t/ are called allophones. Allophones are all (allo) the sounds (phones) that may represent one sound in differing phonological environments.
You will see these allophones represented like this:
(2) /t/ (the sound in the mind of the speaker)
/ | | | \ (surfaces as)[t] [tʰ] [ʔ] [ɾ] [tʃ] (allophones of [t])
In other languages, such as Hindi, [-aspirated] stops and [+aspirated] stops are not allophones of one sound but are contrastive, meaning that when [p] and [pʰ] surface in the same environment, the meaning of the word will actually change.
For example, the [p] in [pəl] ‘moment’ and the [pʰ ] in [pʰəl] ’fruit’ are two distinct phonemes.
You will see these phonemes represented like this:
(3) /p/ /pʰ/
| |[p] [pʰ]
So in this language, aspiration is distinctive.
Allophones can also be distinct phonemes in a language, surfacing in specific, rule-driven environments. For instance, we know that /s/ and /z/ are two separate, distinct phonemes in English. However, when marking a noun for plurality, there is a rule that determines whether the SF will be /s/ or /z/. (4) shows that the plural morpheme will be the voiced alveolar fricative /z/, when a word ends in a voiced sound, and the voiceless alveolar fricative /s/, following a voiceless sound.
(4)[bægz] ‘bags’ [bæks] ‘backs’ [sidz] ‘seeds’ [sits] ‘seats’ [flajz] ‘flies’ [flajts] ‘flights’
Since /s/ and /z/ are variants of a morpheme, they are called allomorphs.
Allophones are generally found in complementary distribution meaning that one form of a phoneme will never appear in the environment of another. In other words, the sounds alternate with each other, depending on where they are found in a word or phrase, just like [t] and [tʰ] in English.
2. Classes of Sounds
Phonemes can be categorized by shared features, creating classes that give insight as to how sounds interact in specific phonological environments.
Major classes are distinguished from one another based on degree of sonority. This corresponds with where sounds are located in a syllable.
syllabic [+/-syll] sounds are those that constitute syllable peaks and may fill the nucleus of a syllable. They are typically more prominent in perception than contiguous non-syllabic sounds.[+syll]: vowels and syllabic consonants [-syll]: stops, fricatives, affricates, approximants, and [-syll] liquids and nasals
consonantal [+/-cons] sounds are produced with a sustained vocal tract constriction at least equal to that required in the production of fricatives; non-consonantal sounds are produced without such constriction.[+cons]: obstruent sounds, nasals, liquids [-cons]: vowels, glides and approximants
sonorant: [+/- son] sounds are produced with a vocal tract configuration sufficiently open that the air pressure inside and outside the mouth is approximately equal. These include vowels, glides, liquids, approximants, and nasals. [-son] sounds are produced with vocal tract constriction sufficient to increase the air pressure inside the oral cavity significantly over that of the ambient air. These include stops, affricates and fricatives.[+ son]: liquids, nasals, glides, approximants, and vowels [-son]: stops, fricatives, affricates
obstruents [+/-obs] basically correspond inversely to [+/- son] such that [+obs] includes stops, affricates and fricatives and [-obs] vowels, glides, liquids, approximants, and nasals.
Manner classes are based on obstruction of airflow.
continuants [+/-cont] are formed with a vocal tract configuration allowing the airstream to flow through the mid-saggital region of the oral tract; plosives are produced with a sustained occlusion in this region.[+cont] consists of vowels, glides, approximants, and fricatives, whereas [-cont] includes plosives, affricates, and nasals.
strident [+/-stri] sounds are produced with complex constriction forcing the airstream to strike two surfaces, producing high-intensity fricative noise[+stri] encompasses fricatives and affricates; sibilants, labiodentals, and uvulars. In Standard English
/f, v, s, z, ʃ, tʃ, ʒ, dʒ/ are marked [+stri].
sibilant [+/-sib] sounds are fricatives and affricates that are [+stri, +cor] and are produced with the front of the tongue raised towards the teeth or the alveolar ridge. As strident, they generate a high amount of turbulence or white noise. English has 6 sibilant sounds: [s, z, ʃ, tʃ, ʒ, dʒ].
Place classes are based on the location of constriction of airflow. Those with which you should become familiar at the introductory level include:
These sounds involve a constriction of the lips, while non-labial sounds do not. Labials include all consonants that involve one or both lips as an articulator, as well as those that bear the secondary feature [w] such as [kw].
Coronal sounds are produced by raising the tongue blade toward the teeth or the hard palate and consist of dental, alveolar, palato-alveolar/post-alveolar, and retroflex consonants.
Anterior consonants are articulated with the front of the tongue anywhere between the alveolar ridge to the back of the upper teeth. They include post-alveolar, alveolar, dental, (interdental), labio-dental, and bilabial sounds.
Dorsal sounds are articulated with the back of the tongue (dorsum) raised towards or in contact with the velum or the uvula, encompassing palatal, velar, and uvular sounds.
*there is sometimes overlap with post-alveolars (palato-alveolars). This is language specific.
Radical sounds are produced with constriction between the tongue root and the pharyngeal wall and include pharyngeal and epiglottal consonants.
Laryngeal sounds are those in which the airflow is occluded or unobstructed by the vocal folds and include glottal consonants.
Natural classes are groups of sounds that share at least one or more features that distinguish them from all other sounds. The minimal number of features is used to describe these sets of phonemes to the exclusion of all others. Feature values are specified by + and – and only those that are relevant to the sounds are used. See our tutorial on Redundancy for more on this topic.
Members of a natural class will generally behave similarly in the same phonological environment. For example, the natural class [+syll, -cons, +high] (high vowels) in English tends to cause palatalization of preceding consonants.
(5) /t/ /tʃ/ /d/ /ʤ/
/stejtli/ /stætʃəɹ/ /gɹejd/ /græʤuəl/
‘stately’ ‘stature’ ‘grade’ ‘gradual’
/z/ /ʒ/ /s/ /ʃ/
/siz/ /siʒəɹ/ /fejs/ /fejʃiəl/
‘seize’ ‘seizure’ ‘face’ ‘facial’
Interestingly, natural classes seem to surface in identical environments across languages. For example, Japanese [+syll, -cons, +high] (high vowels) have the same effect.
/t/ /tʃ/ /s/ /ʃ/
/kat-itai/ [katʃitai] ‘win’ /kas-itai/ [kaʃitai] ‘lend
For more information on these classes, refer to our tutorial on Features and Classes.
3. Analyzing Data
In general terms, the goal of phonological analysis is to identify sound patterns of a language in order to identify distinct phonemes and account for allophonic variation. When conducting phonological analyses, make sure you follow the steps listed below. The more systematic your approach, the more likely you are to come up with a strong analysis.
3.1. Minimal Pairs
The first step is to look for minimal pairs. These are words that differ by a single phoneme. If you are looking at the distribution of /p, b/, find pairs of words whose meaning changes when these sounds are interchanged. An example would be:
These two words are identical except for the first sound of each word, yet they have entirely different meanings.
3.2. Near Minimal Pairs
When we suspect that two phonemes are distinct, but cannot find them in identical phonological environments, we may look for near minimal pairs: two words that share identical sounds in immediately surrounding environments.
For example, /mɪʃən/ ‘mission’ and /vɪʒən/ ‘vision’ show that /ʃ/ and /ʒ/ are distinct phonemes since, although the initial sound in each word is different, the phonemes of interest are found in immediately surrounding environments that are identical, i.e., preceded by /ɪ/ and followed by /ən/.
Another type of near minimal pair exists in which the addition of one sound changes meaning. For instance, in the comparison of /lip/ ‘leap’ and /slip/ ‘sleep’, when /s/ is added, word meaning is altered. In this example, the environment of /s/, word initial, is exchanged with /ø/ which is also in word initial position.
3.3 Free Variation
Some phonemes surface in identical phonological environments and yet meaning is not changed. Sometimes a speaker will use one allophone, and sometimes another, simply based on social context. For example, an English speaker may pronounce the word ‘speaking’ as [spikIn] in certain situations and as [spikIŋ] in others. [n] and [ŋ] are allophones but are freely alternated in word final position. It’s important to note that free variation refers to how one individual speaker may vary pronunciation based on social context, not to language/dialectal-specific rules.
If you find minimal pairs then go no further. They indicate that the phonemes of interest are distinct sounds.
3.4 Complementary Distribution
If you do not find minimal pairs, it is likely the phonemes of interest are allophones found in complementary distribution. Allophones are members of one underlying representation (UR), the abstract sound in the mind of the speaker, that are never found in overlapping phonological environments.
Distinct phonemes may be found in allophonic variation. Consider (7).
[t] [tʃ] SF[stap] [tʃɹejn]
‘stop’ vs ‘train’
In this example, the SF /t/ is found as the second consonant in the word initial cluster; however, when followed by /ɹ/, /t/ becomes /tʃ/. See (2).
Allophones can also be variants of an underlying phoneme, of which the speaker is often not aware. This is true of [th] in word initial position. Speakers of Standard English are not aware that they are adding the feature [aspiration] to this sound. What’s more, if a variant such as [th] is exchanged in the place of /t/, meaning will not be altered. The pronunciation may sound a bit strange, but meaning is clear. Another example is nasal vowels in English. All vowels that precede a nasal stop become nasal.
/flowt/ but /flõwn/
Speakers are not aware of this variation; however, this is the only form of the vowel that will ever surface in this phonological environment. Furthermore, non-nasal vowels will never appear before a nasal.
/flowt/ but not * /flõwt/ and not */flown/
This is a rule-driven distribution of nasal vs. non-nasal vowels.
4. Phonological Environments
In order to locate the distribution of sounds, you can create a chart that lists the phonological environments in which the phonemes of interest appear. This way you can visualize whether or not the sounds are found in complementary distribution or overlapping distribution. Consider (8).
(8)[læf] ‘laugh’ [liv] ‘leave’ [kuɫ] ‘cool’ [stIɫ] ‘still’
In this example there is no overlap in the distribution of /l/ and /ɫ/ since /l/ will only surface in word initial position and /ɫ/ word finally. When charting which sounds appear in which phonological environments, we see that they are in complementary distribution.
Note the symbols for various environments.
#___ (at the beginning of a word)
___# (at the end of a word)
V__V (between vowels)
C___ (after a consonant)
___C (before a consonant)
V___(after a vowel)
___V (before a vowel)
voicing[+/-vce]___ (after a voiced sound)
___[+/-vce] (before a voiced sound)
# = word boundary
$ = syllable boundary
+ = morpheme boundary
5. Writing Rules
Once you find the phonological environments in which the phonemes of interest surface, you can write a rule that accounts for their distribution.
In order to summarize the concept of writing rules we can say:
The phonemic form (UR)
Undergoes phonological rules
To obtain the phonetic form (SF)
Learn to write rules according to your professor’s expectations. Some simply require the phonemes between brackets while others will want a description based on natural classes and distinctive features. Whatever the case may be, here are the standard symbols used in writing rules:
/ / the UR
→ = becomes[ ] the SF
/ = in the environment of
___ = represents the phoneme which undergoes change
In other words, you first state the UR; then state how it changes; then show the circumstances under which it changes (the environment). Here’s a short tutorial on the underlying representation.
6. Types of Rules
There are several types of rules with which you should become familiar.
Rules of assimilation cause a phoneme to take on certain features of an adjacent sound. For instance, all vowels in English become [+nasal] when they precede a [+nasal] consonant. This assimilation of the [+nasal] feature facilitates production of the vowel sound. The rule looks something like this:
V → Ṽ/ ____ C̃
Dissimilation rules cause two adjacent phonemes to become different from each other to facilitate pronunciation. An example of a dissimilation rule would be the omission of [θ] in the word final position of words such as [sIksθ] ‘sixth’ or [fIfθ] ‘fifth’. The [-voice] fricatives [s] and [f] share similar features with [θ] which is also a [-voice] fricative. The fricative [θ] is usually changed to the stop [t], so that the words are actually pronounced [sIkst] and [fIft].
/θ/ → [t] / [fricative] ____
Insertion rules cause a phoneme to be added to a syllable or word. This is illustrated in English plurals. When [mℇs] becomes plural, a plural [z] or [s] cannot be added without first inserting [ə] since two [sibilants] cannot be adjacent to each other. Thus [mℇs] becomes [mℇsəz]. This type of rule will look like this:
∅ → [ə] / [+sibilant] ____ [+sibilant]
Rules of deletion cause a phoneme to be removed from a syllable or word. These rules are applied to facilitate pronunciation and are often seen in fast speech. Most English speakers will pronounce the word [mℇmɔri] ‘memory’ as [mℇmri] ‘memry’, omitting the [ɔ]. This type of rule will look like this:[ɔ] –>∅ /[+sonorant] ____ [+sonorant]
Metathesis rules change the order of sounds. The most common example of metathesis in English can be seen in the pronunciation of the word [æsk] ’ask’ as [æks] ‘aks’.
There is a lot of material to master here. Consistent, systematic study will get you the results you’re looking for!
For more practice and information, check out this site.
Aronow, Okum, Molnar 2015