Stops and Affricates
Chapter Objectives
By the end of this chapter, you will be able to:
- Describe the three phases of a stop: closure, hold, and release.
- Distinguish stop voicing from aspiration as two separate, sometimes independent, properties.
- Interpret the release-related diacritics introduced in this chapter, including aspiration ʰ, the tie bar ͡, and no audible release ̚.
- Distinguish a stop from an affricate using articulatory evidence rather than spelling.
- Compare voice onset time patterns across named language varieties without treating one pattern as the default for all languages.
- Read and produce controlled stop and affricate transcriptions at a broad and at a moderately narrow level of detail.
An Observable Problem
Hold a strip of paper or a single hair loosely in front of your lips. Say tip and watch what happens to the paper right after the initial consonant. Now say dip the same way. For most English speakers, the paper moves noticeably after the t in tip but moves far less, or not at all, after the d in dip. Chapter 5 already gave you a name for this kind of puff of air: aspiration.
Now say chip the same way, attending only to the first sound. The consonant begins like a stop — there is a brief moment where the airflow is fully blocked — but it does not end like one. Instead of a single clean release, you should be able to hear a brief period of friction immediately afterward, closer to the sh sound you will study more closely in Chapter 10. A consonant that behaves this way is neither a simple stop nor a simple fricative. This chapter gives you the tools to describe both kinds of complexity precisely: what stops actually do in time, and what happens when a stop's release is itself turbulent.
- Before reading further, write one sentence predicting whether the t sound in tip, stem, and cat is likely to be identical in all three words, or different depending on its position in the word.
The Anatomy of a Stop
Purpose: describe a stop as an event with internal structure rather than as a single instant.
A stop (also called a plosive) is produced with a complete obstruction of the airstream at some point in the vocal tract, followed by a release of that obstruction. Although a stop is often treated as a single symbol, it is more accurately described as a sequence of three phases:
- Closure (or approach): the active and passive articulators come together, sealing off the airflow at the place of articulation.
- Hold: the articulators remain in contact while air pressure builds up behind the closure. Voicing may or may not continue during this phase, depending on the language and the individual sound.
- Release: the articulators separate, and the built-up air is released, often producing a brief burst of noise.
Place of articulation, introduced in Chapter 7, determines where the closure occurs. This chapter adds a second dimension: what happens during the hold and at the moment of release, which is where voicing, aspiration, and release type become relevant.
A single IPA stop symbol such as t represents an entire closure-hold-release event, not an instantaneous sound. Two stops at the same place of articulation can differ considerably in what happens during the hold and release phases even though they share the same base symbol — which is exactly the source of the aspiration contrast you observed at the start of this chapter.
Voicing and Aspiration: Two Separate Properties
Purpose: prevent the common assumption that "voiced" and "unaspirated" are the same thing.
Chapter 6 introduced voicing as vocal-fold vibration. Chapter 5 introduced aspiration as a property of certain English stops in certain positions. It is tempting to treat voiced stops as one category and aspirated stops as the opposite category, but this conflates two independent properties of timing.
Aspiration is most precisely described using voice onset time, or VOT: the interval between the release of a stop closure and the start of voicing for the following sound. The relationship works like this:
- [pʰɪn]pin — long positive VOT: release happens well before voicing begins, heard as aspiration
- [spɪn]spin — short VOT: release and the onset of voicing happen close together, heard as unaspirated
- [bɪn]bin — voicing may begin before or very close to release, heard as a voiced stop in General American English
Because aspiration is a matter of timing along a continuous scale, languages can and do divide that scale differently. A two-way system can be organized around voicing, around aspiration, or around both together, and a three-way system is also well attested.
VOT is one of the clearest acoustic measurements in introductory phonetics because it can be read directly from a waveform or spectrogram: the release burst appears as a brief spike, and voicing appears as regular, repeating waveform pulses. Chapter 28 returns to VOT measurement in more technical detail. For now, it is enough to recognize VOT as a continuous dimension, not a binary switch, and to understand that "aspirated" describes one region of that continuum, not a separate articulatory mechanism from voicing.
In Hindi, a four-way laryngeal contrast is described for stops at several places of articulation: voiceless unaspirated, voiceless aspirated, voiced unaspirated, and breathy-voiced (sometimes called voiced aspirated). Because aspiration and voicing combine independently rather than always pairing predictably, Hindi distinguishes word pairs using aspiration alone, something English cannot do, since English aspiration is entirely predictable from position and never contrasts two words by itself.
Release Types
Purpose: show that the release phase itself can be modified or absent.
Not every stop ends with an audible burst. Several release patterns are common enough to deserve their own notation:
No audible release
A stop may be produced with the closure and hold phases intact but with little or no audible release, often because the speaker moves directly into another consonant or simply does not release the closure forcefully. This is marked with the diacritic ̚, as in a careful transcription of the final consonant in cat when spoken in isolation: [kʰæt̚].
Nasal and lateral release
When a stop is immediately followed by a nasal or lateral consonant at the same or a nearby place of articulation, the stop's closure may be released directly into that following consonant rather than into an oral vowel. Chapter 11 returns to nasal and lateral release in more detail once nasals and laterals have been introduced as their own sound classes; at this stage, it is enough to recognize that release is not always followed by an open vocal tract.
Error: Assuming that every stop must end with an audible burst of air, and treating unreleased stops as somehow incomplete or mispronounced.
Correction: An unreleased stop is a fully formed, systematic variant, common in many languages and positions, including word-final position in careful English speech. Unreleased stops should be transcribed using ̚ when the task calls for this level of detail, not silently omitted or treated as an error.
Affricates and the Tie Bar
Purpose: analyze affricates as a coordinated stop-to-fricative contour rather than a simple stop.
An affricate begins like a stop, with a complete closure, but releases into a fricative at the same or a nearby place of articulation rather than into an open vocal tract. The closure and the frication are produced as a single, tightly coordinated articulatory event, which is why affricates are treated as their own manner category rather than as a stop followed by a separate fricative.
English has two affricate phonemes, transcribed using a stop symbol and a fricative symbol joined by a tie bar ͡:
- [t͡ʃɪp]chip — voiceless postalveolar affricate
- [d͡ʒɪp]Gyp, jib-like onset — voiced postalveolar affricate, as in jeep [d͡ʒip]
The tie bar signals that the two symbols underneath it represent a single segment, not a sequence of two segments. Many published sources omit the tie bar when the affricate analysis is unambiguous within the relevant tradition, relying on context or a stated convention instead. This textbook includes the tie bar whenever affricate status is the point under discussion, and states explicitly when an established convention permits its omission.
Three different things can sound superficially similar: a single affricate phoneme, such as English /t͡ʃ/; a true sequence of a stop phoneme followed by a separate fricative phoneme, which can occur across a word or syllable boundary, as in nightshade [naɪt.ʃeɪd]; and a stop simply released with audible frication noise that is not treated as a distinct segment at all. Distinguishing these requires looking at syllable structure, phonemic patterning, and sometimes duration, not just the surface sound. Chapter 19 returns to syllable boundaries in detail.
Voice Onset Time Across Languages
Purpose: prevent learners from treating the English aspiration pattern as a universal default.
Because VOT is a continuous scale, languages organize their stop contrasts along it in different ways. The table below compares three patterns at the same general place of articulation, using verified descriptions from the cited language varieties.
| Language and variety | VOT categories distinguished | Example contrast |
|---|---|---|
| English, General American | Two surface categories: aspirated voiceless [tʰ] and plain voiced [d]; unaspirated [t] is a predictable allophone, not a separate phoneme. | /t/ vs. /d/: tin vs. din |
| French, Standard Metropolitan | Two phonemic categories: voiceless unaspirated [t] and voiced [d], with voicing typically beginning before release for the voiced member. | /t/ vs. /d/: ton vs. don |
| Hindi, Standard | Four phonemic categories at the dental place: voiceless unaspirated, voiceless aspirated, voiced unaspirated, and breathy-voiced. | तल /t̪əl/ "bottom" vs. थल /t̪ʰəl/ "land" |
Sources: descriptions checked against established phonetic reference grammars for each named variety. Realizations within each language may vary by speaker, dialect, and speech style; these entries describe a commonly cited reference pattern, not the only attested pattern.
Notice that English and French both distinguish /t/ from /d/, and a learner relying only on the IPA symbols might assume the contrast works the same way in both languages. It does not: English's contrast is primarily one of aspiration, with voicing onset close to release for both members in many positions, while French's contrast is primarily one of voicing, with the voiced member often showing voicing during the hold phase, before release. Sharing a symbol does not mean sharing a realization.
Guided Observation
Before working with the dataset below, complete this short self-observation task. No audio recording is required — you are listening to your own speech.
Say each word aloud at a normal, careful pace, holding a strip of paper or your hand near your lips. For each word, note whether you can feel or hear a release burst, whether you can hear aspiration, and whether the final consonant seems to be released at all.
- top (initial t)
- stop (medial t, after [s])
- pot (final t, spoken in isolation)
- chop (initial affricate)
Write one or two sentences comparing the final consonant in pot, spoken carefully in isolation, with the initial consonant in top. Predict which diacritic from this chapter is most likely to apply to each.
Dataset Analysis Lab
Purpose: apply phase analysis, VOT reasoning, and stop-versus-affricate classification to a small, self-contained dataset.
The forms below are verified phonetic transcriptions of English words, General American reference variety, spoken in careful citation style. Use only this dataset to answer the questions that follow — do not rely on outside knowledge of English spelling.
| Orthographic form | Phonetic transcription | Environment |
|---|---|---|
| ⟨pie⟩ | [pʰaɪ] | Stressed syllable onset |
| ⟨spy⟩ | [spaɪ] | After [s] |
| ⟨cap⟩ | [kʰæp̚] | Word-final, isolation |
| ⟨chap⟩ | [t͡ʃæp̚] | Initial affricate, word-final stop |
| ⟨badge⟩ | [bæd͡ʒ] | Final voiced affricate |
| ⟨bag⟩ | [bæg] | Final voiced stop |
Source: forms checked against a standard American English pronouncing reference for General American citation pronunciation. The transcription level is moderate phonetic detail, sufficient to show the contrasts under analysis.
Step 1 — Identify the three phases
For pie [pʰaɪ], briefly describe what happens during the closure, hold, and release phases of the initial consonant, and identify which phase the aspiration diacritic describes.
Step 2 — Compare release across positions
Compare the final consonants in cap [kʰæp̚] and bag [bæg]. Which one is shown as unreleased in this dataset, and what diacritic marks that?
Step 3 — Distinguish stop from affricate
Compare cap [kʰæp̚] and chap [t͡ʃæp̚]. Both begin with a complete closure. Using the tie-bar notation and your understanding of release type, explain why the initial consonant of chap is analyzed as a single affricate segment rather than as a stop followed by a separate fricative.
Step 4 — Write the analysis
In two to three sentences, summarize what this dataset shows about the relationship between aspiration, release type, and the stop-versus-affricate distinction. Support each claim with a specific form from the table.
Practice — Reading narrow transcriptions: For each transcription below, state in plain English what the diacritic or notation tells you about the release or laryngeal timing of the marked segment.
- [tʰaɪ]
- [stop̚]
- [d͡ʒʌst]
- [æt̚]
Common Mistakes
Error: Treating "voiced" and "unaspirated" as synonyms, or assuming that every language's voicing contrast works the way English's aspiration contrast does.
Correction: Voicing and aspiration are independent dimensions measured by voice onset time. English happens to use aspiration as the primary cue for its voiceless-voiced contrast in stressed syllable onsets, but other languages, such as French, rely primarily on whether voicing begins before release. Always state which dimension is doing the contrastive work for the specific language under analysis.
Error: Transcribing any sequence of a stop letter followed by a fricative letter in ordinary spelling, such as nightshade, as if it were automatically a single affricate.
Correction: A true affricate is a single, tightly coordinated segment, normally analyzed as one phoneme. A stop and a fricative that happen to occur next to each other across a syllable or word boundary, as in nightshade [naɪt.ʃeɪd], remain two separate segments and should not be joined with a tie bar.
Error: Omitting the no-audible-release diacritic ̚ whenever a word-final stop happens to lack a clear burst, even when the task explicitly calls for narrow detail.
Correction: If a transcription task specifies a level of detail that includes release type, mark unreleased stops consistently. If the task specifies a broader level of detail where release type is not contrastive or relevant, omit the diacritic consistently across the whole transcription instead of marking it only sometimes.
Chapter Review
Part A — Definitions
1. Name and briefly describe the three phases of a stop.
2. Define voice onset time in your own words, and explain why it is described as a continuum rather than a binary choice.
3. Define affricate and explain what the tie bar communicates that two separate symbols, written without a tie bar, would not.
Part B — Voicing and Aspiration
1. Explain why [pʰ] and [p] are both considered realizations of English /p/, while in Hindi an aspirated and an unaspirated voiceless stop at the same place can belong to separate phonemes.
2. A learner claims that French /t/ and English /t/ "sound the same because they use the same IPA symbol." Explain what is wrong with this claim, using the VOT comparison from this chapter.
Part C — Stops Versus Affricates
1. Using the dataset from this chapter's Dataset Analysis Lab, explain why chap is transcribed with a tie bar and cap is not.
2. Explain why a stop immediately followed by a fricative across a syllable boundary, as in nightshade, is not analyzed as an affricate.
Repaso acumulativo / Cumulative Review
1. Using Chapter 5's phone-phoneme-allophone framework, classify [pʰ] and [p] in English: are they separate phonemes, or allophones of a single phoneme? Justify your answer using the contrast test.
2. Using Chapter 7's place-of-articulation vocabulary, identify the place of articulation for the affricate in chap [t͡ʃæp̚].
Chapter Performance Task
Performance Task
Building a Stop and Affricate Transcription Set
Choose four English words that include at least one stop in a stressed syllable onset, one stop after [s], one word-final stop spoken in isolation, and one affricate. Working independently or with a partner:
- Write each word in ordinary spelling, then transcribe it phonetically in square brackets at a level of detail that shows aspiration, release type, and affricate status where relevant.
- For each transcription, state in one sentence which diacritic or notation choice you made and why.
- Identify one place in your set where an unreleased or aspirated variant could be mistaken for a different phoneme by a learner unfamiliar with English allophony, and explain why it is not.
Answer Key (click to expand)
Dataset Analysis Lab
Step 1: During closure, the back of the tongue (or lips, depending on place) seals off the airflow completely; during the hold, air pressure builds up behind the closure while the seal is maintained; during release, the articulators separate and air escapes, producing a burst. The aspiration diacritic ʰ describes the release phase specifically: it marks a delay between that release and the onset of voicing for the following vowel.
Step 2: Cap [kʰæp̚] is shown as unreleased, marked with ̚; bag [bæg] is shown with an ordinary release, with no diacritic added.
Step 3: The initial consonant of chap is written with a tie bar, [t͡ʃ], because the closure and the following frication are produced as one tightly coordinated articulatory event at a single place of articulation, functioning as a single segment of English. This differs from a true sequence such as the stop-plus-fricative boundary in nightshade, where the two sounds belong to separate syllables and are not coordinated as one gesture.
Step 4 (model answer): This dataset shows that aspiration marks the release of a stop in a stressed syllable onset, as in [pʰaɪ], but is absent after [s], as in [spaɪ]. It also shows that word-final stops in careful citation speech can be unreleased, as in [kʰæp̚], without this affecting their status as stops. Finally, it shows that affricates such as [t͡ʃæp̚] are distinguished from simple stops such as [kʰæp̚] by the tie-bar notation, which marks a coordinated stop-to-fricative release rather than a plain release into an open vocal tract. Other well-reasoned answers that correctly cite forms from the table are acceptable.
Practice — Reading Narrow Transcriptions
1. [tʰaɪ]: the initial stop is aspirated, showing a delay between release and the onset of voicing. 2. [stop̚]: the medial stop is unaspirated because it follows [s], and the final stop is unreleased. 3. [d͡ʒʌst]: the initial consonant is a single voiced affricate, shown by the tie bar joining the stop and fricative symbols. 4. [æt̚]: the final stop is unreleased.
Chapter Review — Part A (model answers)
1. Closure: the articulators form a complete seal; hold: pressure builds while the seal is maintained; release: the articulators separate and air escapes. 2. Voice onset time is the interval between the release of a stop and the onset of voicing in the following sound; it is a continuum because languages and individual sounds can fall anywhere along a range of delay, not only at two fixed points. 3. An affricate is a single segment that begins with a complete closure and releases into frication at the same or a nearby place of articulation; the tie bar communicates that the two symbols beneath it represent one coordinated segment rather than two separate, sequential segments.
Chapter Review — Part B
1. In English, the difference between [pʰ] and [p] is entirely predictable from position (stressed onset versus after [s]), and no pair of English words is distinguished only by this difference, so they are allophones of one phoneme. In Hindi, aspiration is not predictable from position in the same way and can be the only difference between two words, so the aspirated and unaspirated stops belong to separate phonemes. 2. Sharing an IPA symbol does not guarantee identical realization. As shown in the VOT table, English /t/ is primarily distinguished from /d/ by aspiration, with voicing onset near release, while French /t/ is distinguished from /d/ primarily by voicing, with the voiced member's voicing often beginning before release. The two languages' contrasts rely on different regions of the VOT continuum despite using the same symbols.
Chapter Review — Part C
1. Chap begins with a single closure that releases directly into frication at a shared place of articulation, forming one coordinated segment, so it is transcribed with a tie bar, [t͡ʃæp̚]; cap begins with a stop that releases into an open vocal tract with no following frication, so no tie bar is used, [kʰæp̚]. 2. In nightshade, the stop and fricative belong to separate syllables and are not produced as one coordinated gesture; they are a true sequence of two independent segments rather than a single affricate, so a tie bar is not used.
Cumulative Review
1. [pʰ] and [p] are allophones of a single phoneme, /p/, because their distribution is predictable from context (stressed onset versus after [s]) and no English minimal pair contrasts them in the same environment. 2. The affricate in chap is postalveolar, the same place of articulation used for English /ʃ/ and /ʒ/.
Looking Ahead
What You Can Now Do
You can now describe a stop as a three-phase event, separate voicing from aspiration using voice onset time, recognize and transcribe several release types, and distinguish a true affricate from a stop-fricative sequence. These distinctions will be active whenever this book analyzes obstruent consonants in later chapters.