Instituto Superior “Josefina Contte”

Carrera: Profesorado en Inglés

Unidad Curricular: Fonética y Fonología III

Curso y División: 3° U / 2° Turno: Mañana / Tarde

Profesora: Lic. Mariana Sottile

Profesores Adscriptos: Melanie Gómez Garin – Flavio Frete

The following theoretical framework is built from extracts from:

Davenport, M. and S.J. Hannahs. (1998). Chapter 3. “Consonants”. Introducing Phonetics and
Phonology. London: Arnold.

Finch, D.F. and H. Ortiz Lira. (1982). Chapter 7. “The English Consonants”. A Course in English
Phonetics for Spanish Speakers. London: Heinemann Educational Books Ltd.

Kelly, G. (2000). Chapter 1. “The description of speech”. How to teach pronunciation. Essex:
Pearson Education Limited.

Underhill, A. (2005). Part 1: “Sounds in Isolation”. Sound Foundations. Learning and

Teaching Pronunciation. Oxford, UK: Macmillan Publishers.

Obstruent sounds – Voicing and Devoicing

As stated in previous notes, the class of consonants can be divided into a number of
subgroupings on the basis of their manner of articulation. The first division we will consider
here is obstruent vs. sonorant. For obstruents, the airflow is noticeably restricted, with the
articulators either in complete closure or close approximation. For sonorants, either there is no
such restriction in the oral tract, or the nasal tract is open; either way, the air has free passage
through the vocal tract. The class of obstruents can be further subdivided into stops, fricatives
and affricates, again on the basis of stricture type. The class of sonorant consonants can be
subdivided into nasals, liquids, and glides (vowels are also sonorants, but not sonorant
consonants).

A further important distinction between obstruents and sonorants is that, while the various
obstruent subtypes listed above may have both voiced and voiceless counterparts in most
languages, sonorant subtypes are typically only voiced. Thus English can distinguish “pad” from
“bad” due to the voicing contrast of the initial bilabial obstruents (stops) /p/ and /b/. With
sonorants no such pairs exist; for the nasals, for example, there is only one bilabial – the
voiced nasal found in “mad” – and no voiceless bilabial nasal.
As we’ve already seen, too, in connected speech some sounds undergo considerable variation.
Within sonorants, for instance, glides and liquids lose part of their voice as a consequence of
aspiration. That is, the vocal cords start vibrating a little later than expected when the voiceless
plosive preceding the liquid or glide in a consonant cluster is fully aspirated.

Now, we’ll have a close look at the way obstruents behave in connected speech. But before
doing so, let’s quickly recall how voice is produced.

A sound is said to be voiced if it requires the vocal cords to vibrate, and unvoiced (or voiceless)
if it does not. In English, the voiced/unvoiced distinction tends to coincide with gentle and
strong aspiration (also referred to as lenis and fortis). This means that voiced consonants may
be uttered with weaker breath force, while unvoiced consonants may be uttered with stronger
breath force (Underhill, 2005)

In the human larynx (or “voice box”, as it is commonly known), there are two flaps of elastic,
connective tissue known as vocal cords, which can open and close. During normal breathing,
and also in the production of unvoiced sounds, the cords are open. When the edges of the
vocal cords come close together, the air which passes between them makes them vibrate,
resulting in voicing (Kelly, 2000).

The larynx (Finch and Ortiz Lira, 1982).

The states of the glottis (Finch and Ortiz Lira, 1982).

As stated above, the terms fortis and lenis are used to describe the amount of breath force
required to articulate sounds. Fortis sounds require a forcefully expelled airstream, whereas
lenis sounds are uttered with relatively weak breath force. In spoken English, fortis happens to
equate with unvoiced sounds, and lenis with voiced sounds. This is partly because voiced
sounds take energy from the breath in order to drive the larynx, and partly because unvoiced
sounds need to compensate for their lack of voice with force and clarity in their articulation
(Underhill, 2005).

Voicing and Devoicing - Rules

While the nine English fortis consonants can be safely labelled voiceless, lenis consonants can
lose either part or all of their voice, according to the phonetic environment.

The following rules can be stated:

1- Lenis consonants are fully voiced between vowels and/or other lenis consonants. E.g.,
all the consonants in the utterance “on the living room door” are fully voiced.

[ɒn ðə ˈlɪvɪŋ ruːm dɔː]

Fully voiced consonants are left unmarked in allophonic transcription.

2- Lenis consonants are devoiced after and/or before pauses and fortis consonants. E.g.,
all the lenis consonants in the utterance “that good cause” are devoiced.

[ð̥æk‿ˈg̊ʊg̊‿ˈkʰɔːz̥]

Devoiced consonants are marked with the diacritic [ ̥] below the phoneme symbol, or [
̊] on top if the symbol has a tail.

3- Liquids and glides are devoiced as a consequence of aspiration. E.g., in the utterance
“certain pupils got quite cross”, devoicing of /j, w, r/ occurs.

[sɜːt‿m̩ ˈˈpj̊uːpɫ̩z gɒk‿kw̥aɪk‿ˈkɹ̥ɒs]

4- /h/, which has no voiced counterpart at the phonemic level, becomes voiced in
between vowel sounds. From the point of view of articulation, the vocal cords keep
vibrating when /h/ is produced in intervocalic position. Place your fingers on your
Adam’s apple and feel vibration of the vocal cords in the pronunciation of the word
“rehearse”:

[ɹiˈɦɜːs]

The glottal fricative gains voice in intervocalic position within the word or at word
boundary. Voicing of the glottal fricative is marked using the symbol [ɦ] in narrow
transcription.