Connected Speech and Intonation in English

Posted on Mar 6, 2025 in Photography and Digital Creation

Introduction

In connected speech, some phenomena occur at word boundaries, often changing the sounds substantially. For example, sounds may be:

  • Elided (e.g., give [h]im a letter)
  • Assimilated (e.g., give me > ‘gimme’)
  • Linked (e.g., the idea of > the idea[r]of)
  • Geminated (e.g., this summer > this[s:]ummer)

Elision

Elision is the deletion of a sound or group of sounds. It is normally optional, but widespread. It is mostly motivated by articulatory economy. It affects individual consonants (Cs), vowels (Vs), or consonant clusters.

Consonant Elision

A common elision is /t, d/ elision between consonants. For example: Christmas, kindly, behind this, last night

Conditions:

  • The preceding C must have the same voicing as /t, d/.
  • The following C can be any except /h/.

Another common elision is /h/ deletion (not /h/-dropping). In Received Pronunciation (RP), it occurs in unstressed function words only and phrase-internally (conditions are less strict in other accents). /h/-deletion is optional, but widespread.

Example: The men have gone, but we pronounce When HE comes back.

Other consonant elisions involve:

  • /l/: ‘dark l’ ([ɫ]) is often deleted after /ɔ:/, before a C. Example: always /ˈɔ:wɪz/
  • Various consonants in CCC or CCCC clusters. Example: jumped /ʤʌm(p)t/

Vowel Elision

Schwa (/ə/) is often elided between voiceless consonants. For example, potato /p(ə)ˈteɪtəʊ/.

In this example, schwa can be elided in unstressed syllables before /n/ and /l/, which then become syllabic.

Example: not alone /ˈnɒt ə ˈləʊn/ > /ˈnɒt l ˈəʊn/

This elision of schwa before word-final /m, n/, or /l/ is largely lexicalized, and the consonants are syllabic.

  • /ə/ + /l/: people /ˈpiːpəl/ > /ˈpiːp.l/
  • /ə/ + /m, n/: button /ˈbʌtən/ > /ˈbʌt.n/

Elision of schwa is particularly common when schwa is followed by an unstressed syllable with a weak V (/ə, i/).

  • /ə/ + /l, r/ + /ə/: camera /ˈkæmərə/ > /ˈkæmrə/
  • /ə/ + /l, r/ + /i/: secretary /ˈsekrətəri/ > /ˈsekrətri/

A syllabic consonant often leads to compression: the syllable containing the syllabic C disappears altogether, and the syllabic C becomes an onset of the following syllable.

Assimilation

Assimilation is a process whereby a segment takes on a property (e.g., voicing, place of articulation (POA), manner of articulation (MOA)) of an adjacent sound. As a connected speech phenomenon, it is mostly optional but widespread.

  • POA assimilation
  • Voicing and MOA assimilation
  • Coalescence

POA Assimilation

POA assimilation is particularly common in English. The most common form involves the movement of the POA of alveolar consonants to the POA of the following sound.

Three common examples are:

  1. Alveolar /n, d, t/ to bilabial /m, b, p/ when followed by any bilabial /p, b, m/ or labial-velar /w/: ten boys /tem bɔɪz/
  2. Alveolar /n, d, t/ to velar /ŋ, k, g/ when followed by velar /k, g/: ten cars /teŋ kɑ:z/
  3. Alveolar /s, z/ to post-alveolar /ʃ, ʒ/ when followed by postalveolar /ʧ, ʤ, ʃ, ʒ/ or palatal /j/: this shop /ðɪʃ ʃɒp/

Voicing and MOA Assimilation

Voicing assimilation is rare in English and rarely compulsory. However, it is lexicalized in some words.

Examples: it’s /ɪts/, used to /ˈju:s(t) tuː/

There is also compulsory voicing assimilation in regular suffixes –s and –ed in English, but it is also lexical.

Examples: dreams /dri:mz/, cats /kæts/

MOA assimilation is rare in English.

Example: /v/ > /m/ in ‘gi[m]me’ for ‘give me’

However, /ð/ tends to be weakly articulated and may be vulnerable to a preceding sound.

Example: is this OK? /ɪz zɪs ˈəʊ keɪ/

Coalescence

Assimilation typically takes either of two directions, depending on the direction of the change:

  1. L → R (left-to-right, progressive, or perseverative): features of segment 1 carried over to segment 2.
  2. R → L (right-to-left, regressive, or anticipatory): features of segment 2 anticipated in segment 1.

However, there is a third type of assimilation:

L ↔ R (mutual or coalescent): both segments take on features from each other.

In coalescent assimilation, both sounds influence each other (it involves progressive and regressive assimilation). A well-known case of coalescent assimilation is yod coalescence: /t, d/ + /j/ = /ʧ, ʤ/

Coalescent assimilation: This process has occurred before and continues to occur in English word-internally.

nature (Old English/Middle English) /naˈtju:r/ > (Modern English) /ˈneɪʧə/

tube (Modern English) /tju:b/ > (Modern English) /ʧu:b/

Yod coalescence is optional but widespread.

Linking /r/ and Intrusive /r/

In non-rhotic English, /r/ is never pronounced after a non-close vowel followed by a consonant or phrase-finally. Example: fork /fɔ:k/

But /r/ is generally pronounced post-vocalically across word boundaries when followed by a vowel > ‘linking’ /r/. Example: it’s near a shop /ɪts ˈnɪər‿ə ˈʃɒp/

Linking /r/ is, historically, a syllable-final C that was elided but is recovered before a morpheme-initial V.

If necessary, linking can be shown with the IPA diacritic undertie [‿] between /r/ and the following V. It can be characterized functionally as a euphonic strategy for avoiding hiatus.

Intrusive /r/ is a non-etymological, non-orthographic r-sound in the same contexts as linking /r/. It appears due to analogy with linking /r/. Linking /r/ and intrusive /r/ are jointly called /r/-liaison.

Two subcases of /r/-liaison, based on spelling criteria:

  • Linking /r/: r-sound for the letter <r(e)>. Example: the star of the show /ðə ˈstɑːr‿əv ðə ˈʃəʊ/.
  • Intrusive /r/: r-sound, but no letter <r(e)>. Example: put a comma(r) in /pʊt ə ˈkɒmər‿ɪn/.

Weak Forms of ‘the’, ‘do’, ‘to’

Sandhi can also be seen in the weak forms of a few function words (FWs). The weak forms of ‘the’, ‘do’, and ‘to’ finish in a close vowel /i, u/ if followed by a V and in /ə/ if followed by a C.

  1. the: (/ði/ + V: the other day /ði ˈʌðə deɪ/ … /ðə/ + C: to the shop /tə ðə ˈʃɒp/)
  2. do: (/du/ + V: why do all go? /ˈwaɪ du ˈɔ:l ˈgəʊ/ … /də/ + C: why do they go? /ˈwaɪ də ðeɪ ˈgəʊ/)
  3. to: (/tu/ + V: to our country /tu aʊə ˈkʌntri/ … /tə/ + C: up to the counter /ˈʌp tə ðə ˈkaʊntə/)

Gemination

Gemination, or consonant elongation, is a sandhi process whereby two consecutive identical Cs are pronounced as a single but longer C, called a geminate (or geminated C).

In English, all manners of articulation (MOAs) can geminate except the affricates.

  • (plosives): bad day /bæd deɪ/ > /bædːeɪ/
  • (nasals): ten nights /ten naɪts/ > /tenːaɪts/
  • (fricatives): last summer /lɑːs(t) sʌmə/ > /lɑːsːʌmə/
  • (approximants): all letters /ɔːl letəz/ > /ɔːlːetəz/

Lenition /t/-tapping

Lenition is a sound change that alters a C, making it be articulated with less energy (from Latin lenis = weak). A well-known lenition in English is t-tapping (or t-flapping), which turns alveolar /t, d/ into a voiced alveolar ‘tap’ [ɾ] when followed by an unstressed syllable.

what are you doing /ˈwɒɾə ju ˈdu:ɪŋ/

Introduction

Intonation is the use of pitch variations in language.

Pitch is the highness or lowness of your voice (in Spanish, ‘tono’). It is the way we perceive fundamental frequency (f0) – or the rate of vibration of the vocal folds.

Intonation helps create the ‘music’ of language. Like music, intonation consists of high notes, middle notes, and low notes. Our voice would be flat and strange without intonation. Intonation is important to make speech sound ‘human’ and to avoid misunderstandings.

In writing, changes in pitch are often shown by means of punctuation marks (? ! …). However, fundamental frequency (f0) changes can be seen very well in pitch contour graphs.

Intonation in Languages

Languages like English or Spanish are known as intonation languages (ILs). In intonation languages, utterances have varied pitch contours, but pitch is not used to distinguish words. Yet pitch may affect the meaning of a whole tone unit.

In ILs, the information conveyed by intonation can be:

  1. Grammatical information (e.g., sentence types): She got a dog (statement). She got a dog? (question)
  2. Attitudinal and emotional information (e.g., finality, irony, joy…): That’s it. Great!
  3. Pragmatic/discourse information (e.g., new vs. old information): My name is Bond. James Bond.

Other languages are tone languages (TLs). In TLs, pitch changes are also used to contrast the meanings of words. For example, in Mandarin Chinese, the sequence [ma] can be uttered with five different pitches, leading to five different meanings.

The Tone Unit

The basic unit in intonation is the tone unit (TU), also called intonation unit (IU).

Its length is variable, and its boundaries are marked with vertical strokes (single or double).

Each stressed syllable in a TU is a rhythmic stress (also called accent) (cf. lexical stress!).

In TUs, one rhythmic stress is more relevant than the others (if there is more than one). This accent is called the nuclear accent of the TU.

The nuclear accent carries the tone, or most important pitch movement.

| ˈPeter is ˈGOing || ˈPeter is ˈgoing to ˈLONdon || ˈPeter is ˈgoing to ˈLondon for a ˈCONference |

In a TU like this ↑ we can find:

  1. A nuclear accent (e.g., ‘CON)
  2. Other rhythmic stresses (e.g., ‘Pe-, ‘go-, ‘Lon-)
  3. Non-stressed syllables (e.g., –ter, is, -ing, to, -don, for, a, -fe, -rence)

Aspects of Intonation

Non-specialists usually associate intonation exclusively with the pitch movement or ‘melody’ of speech. However, intonation involves three different aspects:

  • Tonality: where to place TU boundaries.
  • Tonicity: where to place the nuclear accent.
  • Tone: what pitch pattern is used in the nucleus.

Tonality

Tonality is the division of speech into tone units. Tonality is indicated by different ‘cues’ or signals:

  1. Phonetic (e.g., pauses, changes in pitch level)
  2. Written (e.g., punctuation marks, but not always)
  3. Grammatical (e.g., phrase/clause/sentence boundaries)

Tonality can disambiguate and shift emphasis.

Tonicity

Tonicity is the range of choices in the position that the nuclear accent can have in a given TU.

Default position: the rhythmically stressed syllable in the last content word of the TU.

However, any syllable in a TU may become the nuclear accent, including function words. The following situations alter the default rule:

  • Contrastive focus
  • Use of emphasis
  • New/important information focus
  • Use of ‘empty’ words or vocatives
  • Use of reporting clauses
  • Use of pro-forms

Non-default tonicity:

  1. Contrastive focus: Default vs. contrastive tonicity

    | He ˈlikes red ˈCARS || He ˈdoesn’t like ˈRED cars | He ˈlikes ˈGREEN cars |

  2. Emphasis: Default vs. emphatic tonicity

    | The ˈmovie was ˈBOring | | The ˈmovie was ˈVERy boring |

  3. ‘New information’ focus

    | Are ˈlegumes a ˈsource of ˈVItamins? || They are a ˈGOOD ˈsource of ˈvitamins |

  4. Semantically ‘empty’ words & vocatives

    | I ˈvisited ˈDIfferent places | I ˈliked ˈLOTS of things || ˈNice to ˈSEE you, ˈJohn | ˈMORning, Mrs ˈSmith |

  5. Reporting clauses

    | “How are you ˈDOing?” he asked. || “Nice to ˈSEE you,” she ˈreplied |

  6. Pro-forms

    | Could I borrow some ˈSUgar? | I haven’t ˈGOT any. |

Tone

Tone is the choice of pitch pattern in the nuclear accent. The nuclear accent is therefore also called the tonic syllable. Tones can be simple or complex.

  • Fall: | His name’s \PEter | (the pitch descends)
  • Fall-Rise: | His name’s \/PEter | (the pitch descends and then rises again)

Five basic tones in English:

  1. Fall \ | His name’s \PEter |
  2. Rise / | His name’s /PEter |
  3. Fall-rise \/ | His name’s \/PEter |
  4. Rise-fall /\ | His name’s /\PEter |
  5. Level — | His name’s —PEter |
  • The falling tone (or the Fall):
    • Reply to yes/no questions and statements (finality)
    • Imperatives and exclamations
    • Wh-questions: falling tones
    • Question-tags seeking confirmation (express certainty)
  • The rising tone (or the Rise):
    • Yes/No questions
    • Question-tags requesting for more information (uncertainty)
    • Listing (all items, except the last on the list)
    • Statements that question
    • Statements showing inconclusiveness (something else is to follow)
    • Statements: encouraging someone.
  • The falling-rising tone (or the Fall-Rise):
    • It expresses limited agreement, response with reservation, uncertainty, or doubt.
    • Often followed in the following TU by a fall.
  • The rising-falling tone (or the Rise-Fall):
    • It conveys the idea that the speaker is impressed (or expects listeners to be impressed) and shows enthusiastic agreement. However, it can also express suspicion, indignation, incredulity, mockery, or disapproval.
  • The level tone (or the Level):
    • It expresses routine, boredom, or sarcasm.

English vs. Spanish Intonation

English and Spanish use intonation to convey emotional states and attitudes (Spanish is usually flatter). They also use intonation with grammatical and discourse purposes.

There are many differences between English and Spanish intonation. Three examples:

  • List: In English, listed items are often pronounced with rising intonation, except for the last item (with falling intonation). In Spanish, all listed items are typically pronounced with falling intonation.
  • Wh-questions: Wh-questions almost invariably have falling intonation in English, while in Spanish, they typically have rising intonation (occasionally falling, too).
  • Use of contrastive focus (Contrastive nuclear accent): In both languages, the last stressed σ in a sentence is usually the most important (→ nuclear ‘accent’). However, to express a focus on contrasts, the nuclear accent in English may be in some σ before the last, but rarely in Spanish.