Voicing is the vibration of the vocal folds during the production of speech sounds. In phonetics, it describes a physical property of speech: a sound is voiced when its production includes vocal-fold vibration and voiceless when it does not. In phonology, voicing also refers to a property that distinguishes categories of sounds, even when their pronunciation does not consistently involve vibration throughout. The distinction is especially important for consonants, as in English /s/ and /z/, but can also apply to vowels and other speech sounds.
Physical production
The vocal folds are paired structures in the larynx. In ordinary voiced speech, they are brought sufficiently close together for airflow from the lungs to sustain repeated oscillation. Their vibration results from the interaction of air pressure, tissue elasticity, and aerodynamic forces, rather than from a separate muscular contraction for each cycle.
These oscillations periodically interrupt the airflow, providing a sound source that is shaped by the cavities above the larynx. The rate of vibration determines the fundamental frequency, a major acoustic correlate of perceived pitch. Whether vibration occurs is therefore distinct from how high or low its frequency is: both high-pitched and low-pitched sounds can be voiced.
Voiceless sounds are produced without this periodic vocal-fold vibration. For many voiceless consonants, the vocal folds are sufficiently separated to allow airflow without oscillation. Their sound may instead come primarily from turbulent airflow, as in [s], or from the release of an oral closure, as in [t].
Maintaining voicing is particularly demanding during a complete oral closure. As pressure builds behind the closure, the pressure difference across the vocal folds decreases, making vibration harder to sustain. Speakers may compensate by expanding the cavity above the larynx or adjusting airflow and timing. This helps explain why voiced stops can lose vibration during part of their closure.
Voicing and phonation
Voicing is closely related to phonation, the broader study and classification of laryngeal sound production. Ordinary, or modal, voice is only one possible configuration. Breathy voice combines vocal-fold vibration with relatively substantial airflow, while creaky voice involves a different pattern of vibration, often irregular and at low frequency.
These categories should not be treated as a simple scale from “voiced” to “voiceless.” They involve several interacting properties, including vocal-fold tension, glottal opening, contact patterns, and airflow. Languages can use such differences to distinguish words.
A vowel is normally voiced in ordinary speech, but vowels may also be voiceless or devoiced in particular environments. In Japanese, for example, high vowels frequently undergo devoicing between voiceless consonants, with the details varying by dialect and speaking conditions. Nasals, approximants, and other consonants normally produced with voicing can likewise have voiceless counterparts.
Transcription and sound classification
The International Phonetic Alphabet provides separate symbols for many voiced–voiceless consonant pairs:
| Voiceless | Voiced | Shared articulation |
|---|---|---|
| [p] | [b] | Bilabial stop |
| [t] | [d] | Alveolar stop |
| [k] | [ɡ] | Velar stop |
| [f] | [v] | Labiodental fricative |
| [s] | [z] | Alveolar fricative |
| [ʃ] | [ʒ] | Postalveolar fricative |
Members of each pair share a broad place of articulation and manner of articulation, while differing in their conventional voicing classification. Their actual pronunciation may differ in additional respects, such as timing and intensity.
For sounds without a dedicated contrasting symbol, a diacritic can indicate voicelessness or voicing. Examples include [n̥], a voiceless nasal, and [s̬], an [s]-type articulation marked as voiced.
Square brackets indicate phonetic transcription, describing pronunciation. Slashes indicate phonological categories: /d/, for example, denotes a phoneme, not a guarantee that every instance contains uninterrupted vocal-fold vibration.
Phonetic voicing and phonological contrast
A phonetic measurement asks whether, when, and how strongly the vocal folds vibrate. A phonological analysis asks how sounds function within a language’s system of contrasts. These questions are related but not identical.
In English, sip /sɪp/ and zip /zɪp/ form a minimal pair illustrating a contrast conventionally described as voiceless versus voiced. Yet English /b d ɡ/ may be partly or wholly voiceless in some positions, particularly at the beginning of an utterance. Conversely, sounds assigned to a voiceless category may acquire some vibration through the influence of neighboring voiced sounds.
Listeners identify the categories from multiple acoustic cues rather than from vocal-fold vibration alone. Depending on the sound and context, these can include:
- The timing of voicing relative to a consonant release.
- The duration of the preceding vowel.
- The duration of a closure or fricative.
- The presence and duration of aspiration.
- Fundamental-frequency patterns near the consonant.
For example, in many varieties of English, the vowel in bid is longer than the vowel in bit. This duration difference helps distinguish the words even when the final /d/ is substantially devoiced.
Phonological theories often represent voicing through a distinctive feature, conventionally written [±voice]. Other analyses use different laryngeal features or distinguish categories such as fortis and lenis. Fortis–lenis terminology is not an exact synonym for voiceless–voiced: its interpretation depends on the language and analysis.
Voice onset time and aspiration
Voice onset time (VOT) measures the interval between the release of a stop closure and the onset of vocal-fold vibration. It is especially useful for comparing stop consonants across languages.
Three broad patterns are commonly distinguished:
- Voicing lead: vibration begins before the stop is released.
- Short-lag VOT: vibration begins shortly after release.
- Long-lag VOT: vibration begins after a longer delay, often with aspiration.
English /p t k/ commonly have long-lag VOT at the beginning of stressed syllables, as in pin, tin, and kin. After /s/, as in spin, sting, and skin, they are generally unaspirated. English /b d ɡ/ often have short-lag VOT in utterance-initial position, although prevoicing also occurs.
By contrast, the major stop distinction in Mandarin Chinese is between unaspirated and aspirated voiceless stops. The letters b and p in pinyin, for example, represent this aspiration contrast rather than a straightforward voiced–voiceless pair.
VOT is not a complete measure of voicing. It describes timing around a stop release, but does not by itself capture every aspect of vibration during a closure, the voicing of fricatives, or differences in voice quality.
Cross-linguistic patterns
Languages organize laryngeal contrasts in different ways. Some have no phonemic voiced–voiceless opposition among stops. Others distinguish two or more categories involving voicing, aspiration, or phonation.
Sanskrit traditionally distinguishes voiceless unaspirated, voiceless aspirated, voiced unaspirated, and voiced aspirated stops. The last category involves a breathy-voiced release and is not simply the combination of ordinary voicing with the same aspiration found in voiceless aspirated stops.
Korean has a three-way stop contrast conventionally described as lenis, aspirated, and tense. A simple voiced–voiceless classification does not adequately describe this system. The relative importance of VOT, following-vowel pitch, and other cues varies across varieties and generations of speakers.
Such patterns show why the label “voiced consonant” cannot always be transferred directly from one language to another. Phonetic realization and the organization of contrasts must both be considered.
Contextual processes and historical change
Voicing frequently changes according to a sound’s surroundings.
Voicing assimilation occurs when neighboring sounds become more alike in laryngeal properties. As a form of assimilation, it can involve a voiced consonant becoming voiceless before a voiceless consonant, or a voiceless consonant becoming voiced beside a voiced sound. Languages differ in the direction, domain, and completeness of these processes.
Final devoicing removes or reduces voicing contrasts at the ends of words or syllables. German and Dutch, for example, conventionally neutralize voiced–voiceless obstruent contrasts word-finally. Some acoustic studies find small residual differences in particular speakers or conditions, so phonological neutralization need not imply perfect phonetic identity.
Intervocalic voicing is the voicing of consonants between vowels. It can occur as a contextual pronunciation pattern or become established through sound change. Historically, it often accompanies other changes grouped under lenition, such as weakening a stop to a fricative.
Changes in consonantal voicing can also affect neighboring pitch. If those pitch differences become contrastive after the original consonantal distinction disappears, they may contribute to tonogenesis, the historical development of tone.
Observation and measurement
In acoustic phonetics, voicing can be investigated through waveforms, spectrograms, and measures of periodicity. Regular vibration often produces repeated waveform patterns and harmonic structure. During a voiced stop closure, low-frequency energy may appear as a voice bar on a spectrogram.
Electroglottography provides complementary information by measuring changes in electrical conductance associated with vocal-fold contact. Laryngeal imaging can reveal aspects of vocal-fold movement more directly.
No single measurement is sufficient for every case. Creaky voice may be irregular rather than strictly periodic, breathy voice combines periodic and noisy components, and consonants may be voiced during only part of their duration. Descriptions therefore often specify the timing or proportion of voicing rather than assigning every sound token an unqualified binary label.