The perception of natural speech sounds is based on phonetic features, the smallest units of distinction between any two phonemes. We use features such as voicing to differentiate between the words /kejm/ ‘came’ and /gejm/ ‘game’. The only distinguishing feature between these 2 sounds is that the vocal cords not do vibrate for during the articulation of /k/, and do for /g/. Place and manner of articulation, as well as voicing are the basis for contrast between consonants. For vowels, it is the height of sections of the tongue.
Speech sounds can be classified into broader categories, or superordinate classes that entail specific phonetic features. These classes provide descriptive information as to how sounds are distributed, how they interact, and how they can be combined or distinguished in any given language. Phonological features also show patterns across languages. For example, in many languages, the voiced stops /b, d, g/ undergo spirantization and are realized as fricatives when they occur inter-vocalically.
Spanish (Kaplan, 2010)
/nube/ –> [nuβe] ‘coud’
/lado/ –> [laðo] ‘side’
/lago/ –> [laɣo] ‘lake’
Proto-Bantu (Halle and Clements, 1983)
/abag/ –> [aβaɣ] ‘loincloth’
/sɔdɔj/ –> [sɔɾɔj] ‘remain’
/gigiul/ –> [gigiul] ‘spatula’
Classes are used to show the interactive patterns of features. We see that in categorizing phonemes into classes, various features interface, defining a sound as a group of like sounds. For example, bi-labials and labio-dentals are separate categories when understood as phonetic features of place. However, the merging of the two places of articulation yields a broader category of sounds, labials that actually behave similarly across languages since the lips are the key articulators.
Features can be used not only to describe speech sounds, but also to arrange them in groups or classes. Take, for example, /t/ and /d/ in English. These phonemes can be described using two features:
alveolar or (place of articulation)
stop/plosive (manner of articulation).
The only two phonemes in English that bear both features are /t/ and /d/. Since there are no other phonemes that can be included in this category, we can say they form a ‘class’.
Using the features [+voice] and stop/plosive without specifying the place of articulation, we form a class composed of [d, b, g] to which no other phonemes in English can be added.
Again, Standard English, when specified for back, height, and tenseness, the class [u, o] is formed, i.e., [+back, -low, tense]. There are no other vowels that are permitted into this class.
Description of Classes
All languages stipulate default feature values for phonemes. The descriptions below do not take into account any specific language.
Major Class Features
Major classes are formed based on degree of sonority and syllabic structure.
The broadest distinguishing category of speech sounds is [consonantal] [±cons]. In most languages phonemes are either/or. [+cons] sounds are produced with varying degrees of airstream obstruction due to the constriction of articulators the vocal tract. These include plosives (stops), fricatives, affricates, nasals, trills, approximants, ejectives and implosives, while [-cons] phonemes are produced with no constriction. In most languages this category is restricted to vowels.
[±syllabic] [±syll] is the inverse correlate for [consonantal]. [+syll] phonemes are considered the most sonorous segments of a language and are singularly permitted in the nucleus of a syllable, whereas [–syll] are not.
The second major class is [sonorant] [±son] that inversely pairs with [obstruent]. [+son] describes phonemes that are produced with relatively free airflow since there is little or no constriction of the articulators. These sound segments are said to be ‘sing-able.’ In other words, they can sustain resonance. Approximants, liquids, nasals, glides, and vowels are included in this class. [-son] is also sometimes referred to as [+obs]. These phonemes are articulated with noticeable air obstruction, i.e., stops, affricates, fricatives, ejectives, and implosives. Note the binary distribution of these classes.
Manner classes categorize sounds based on the type and degree of airflow through the vocal tract.
Continuants [±cont] are described in terms of sustained obstruction of airflow through the oral cavity. The class [+cont] includes fricatives, liquids, glides, as well as vowels. In the production of these phonemes, the vocal tract is always slightly open, allowing the passage of air with varying degrees obstruction. For example, in English, during the production of labiodentals [f,v] and interdentals [θ,ð] a continual flow of air passes through a narrow opening between the lips and the teeth. Although there is obstruction, it is partial and airflow is never completely cut off. In the production of glides and liquids, the air passes through a wider area of the oral cavity with no constriction of the articulators.
The second major class is [sonorant] [±son] that inversely pairs with [obstruent]. [+son] describes phonemes that are produced with relatively free airflow since there is little or no constriction of the articulators. These sound segments are said to be ‘sing-able.’ In other words, they can sustain resonance. Approximants, liquids, nasals, glides, and vowels are included in this class. [-son] is also sometimes referred to a
Non-continuants [-cont] include oral and nasal stops, and affricates, which are complex phonemes that have an initial stop closure with a fricative release. Although in the production of nasals the passage of air is not completely obstructed at any point, air does not pass through the oral cavity thus they are considered non-continuants.
*Note that in certain languages approximants are [-cont].
Here are a few more features that are used for making more detailed distinctions.
Distributed [+/-distr] specifies the extension of constriction along the mid-sagittal axis of the oral tract.
[+distr] sounds include bilabials, laminal alveolars, alveo-palatals, palatal, velar, uvular, and pharyngeal sounds.
[-distr] sounds include labio-dentals, apical alveolars, and retroflex consonants.
Coronal sounds, i.e., constriction created with the front of the tongue against the alveolar ridge and immediately surrounding areas (dental, alveolar, palato-alveolar, and retroflex) can be further specified as apical or laminal.
Apical consonants are articulated with the tip of the tongue, e.g., /t, d, n/ whereas laminals are articulated with the blade of the tongue, e.g., /tʃ, dʒ/.
The feature nasal [±nas] applies to either consonants or vowels and describes phonemes that are articulated with a lowered velum (soft palate) such that airflow is directed through he nasal cavity.
Lateral consonants [±lat] are characterized by airstream that passes along the side(s) of the tongue.
Delayed release [±del rel] is a feature used mainly when comparing stops and affricates. Phonemes bearing this feature are characterized by a slow, frictive release of air. Some phonologists use this feature for all phonemes that do not undergo a stop release.
Strident: [+/-stri(d)] sounds are produced with complex constriction forcing the airstream to strike two surfaces, producing high-intensity fricative noise.
[+stri] only applies to fricatives and affricates and includes labio-dental, dental, alveolar, and alveo-palatal consonants. All other phonemes are [-stri].
Sibilants [+/-sib] are fricative and affricate consonants that are [+stri, +cor].
They 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.
The feature spread [+/-spr] refers to phonemes produced with the vocal cords drawn apart, producing a non-periodic (noise) component in the acoustic signal.
[+spr] sounds include aspirated, breathy voiced or murmured consonants, voiceless vowels and glides. All others are [-spr]
Voicing [+/-vce] distinguishes between sounds are produced with a laryngeal configuration permitting periodic vibration of the vocal cords [+vce] and those that lack such periodic vibration [-vce].
Labial [±lab] refers to phonemes produced with the lips as an active articulator and includes bilabials and labiodentals. [p, b, m, f, v, w].
Round [±rnd] is a feature attributed to phonemes (consonants and vowels) that are produced with rounded lips. The distribution of rounding is very language specific.
Anterior phonemes [±ant] are produced with the active articulator positioned toward the alveolar ridge and forward toward the lips. This includes alveolars, alveolar-dentals, interdentals, labiodentals, and bilabials. [-ant] sounds are sometimes labeled as [+posterior].
Coronal [±cor] phonemes are articulated with the tongue against anywhere between the alveolar ridge and the hard palate. This includes alveolar, post-alveolar, and palatal consonants.
Dorsal [±dor] sounds are articulated with the back of the tongue (dorsum) against the area between posterior section of the hard palate and velum (soft palate) and includes palatals, velars, and uvulars.
Subcategories which specify more detailed information about the position of the dorsum are:
high: [+hi] are sounds produced with the dorsum raised such as palatals, velars, and high vowels.
[-hi] are sounds produced with the dorsum lowered such as mid and low vowels and all [-high] consonants.
low: [+low] are sounds produced with the body of the tongue lowered such as low vowels [æ, a, ɒ, ɶ] and pharyngeal. [-low] are all sounds which are not [+low].
back: [+bck] sounds are produced with the body of the tongue behind the hard palate and includes velars, uvulars, and back vowels. [-bck] are all sounds which are not [+bck], i.e., front vowels and coronal and anterior phonemes.
pharyngeal [±phar] sounds are made with the root of the tongue against or towards the pharyngeal wall.
Gaining a clear understanding of ‘how’ and ‘why’ of classifying phonemes according to shared features is the first step in conducting sound phonological analyses.
These feature classes allow us to recognize both basic and complex phonological rules by revealing distributional and interactive patterns.
When you feel you have mastered this material, your next read is: