Intonation contour

Encyclopedia of Terminology for CA and IL: Intonation contour
Author(s):	Marina N. Cantarutti (University of York, UK) (https://orcid.org/0000-0002-1688-0896)
To cite:	Cantarutti, Marina N. (2023). Intonation contour. In Alexandra Gubina & Chase Wesley Raymond (Eds.), Encyclopedia of Terminology for Conversation Analysis and Interactional Linguistics. International Society for Conversation Analysis (ISCA). DOI: []

An intonation contour, in informal terms known as the “melody of speech”, normally refers to the perceivable falling, rising or steady (level) movements of pitch over a segment of time within a unit. In CA-IL research, intonation contours have been found to bear different kinds of relationship with TCU and action boundaries as well as with the design of different social actions, including, e.g., second assessments, other-initiations of repair, and lists.

Technically, an intonation contour is a phonologically contrastive, systematic, and language-specific configuration of pitch movement whose domain goes beyond the individual sound and extends into words and phrases. Intonation contours are made up of sequences of changes in f0 (an acoustic measure that measures the frequency in the vibration of molecules in the air, in our case, the effect of vocal fold vibration as cycles of opening and closing) which is perceived as differences in pitch. Major intonational contours tend to be initiated on the focal accent of a unit (e.g., for English, the most prominent and typically the last accent in the unit), and they are realized normally over a number of syllables towards the end of the unit.

CA-IL have incorporated studies of intonation contours as resources (Schegloff, 1988) in direct and less direct ways from the very start of the discipline (e.g., Jefferson, 1972; Sacks, Schegloff & Jefferson, 1974; Sacks, 1992). However, there are different aspects of what makes up an intonation contour that have been addressed in different ways in CA-IL and which need to be differentiated.

The whole intonational configuration of the unit as a gestalt includes not only the major pitch movements normally starting at the focal accent (falling, rising, level, and variants; see below), but also any prior pitch jumps or movements (falls, rises, step ups or step downs in pitch level often in early secondary accents), as well as terminal pitch levels reached at the end of the unit.

First, upsteps and downsteps, that is, noticeable pitch jumps to a point higher or lower in pitch than surrounding syllables (marked in CA-IL transcription systems with upward ↑ and downward ↓ arrows) are relevant for interactional ends. For instance, a high onset (i.e., first accented syllable in the unit) have been found to be interactionally relevant to mark shifts into the “reason for the call” (Couper-Kuhlen, 2001). High peaks on or preceding focal accents were also found in contexts where speakers pre-empt resistance (Nielsen & Morris, 2019). Pitch level at the start and end of discontinued units has a bearing in marking turn continuation or re-starts (Local, 1992). In the following example, a speaker marks topic transition with “anyway”, followed by a high onset on “how”:

[CallFriend _5926_3-10]

01  MOM:   I can't ima:gine. >anybody wanting to< listen. 
02         to this pho:ne call, hh. (0.2)
03  JUL:   hhh.
04  MOM:   or read a transcript. either. hhh. [But-]
05  JUL:                                      [no, ]
06  MOM:   >anyway.< ↑SO anyway. hh. So:: uhm:: (.) ↑↑how many people
07         are you having. o:f: yo:ur friends.

Figure 1: Acoustic visualization of "↑SO anyway. hh. So:: uhm:: (.) ↑↑how many people..."

Then, there is pitch movement initiated at the focal accent that is generally carried over to the end of the unit, that in IL work in English and German at least (following traditions in configurational approaches to intonation, e.g. Cruttenden, 1997) is normally taken to be falling, rising or level (including rising-falling, or falling-rising movements and their variants). Early CA research has used labels from Jeffersonian conventions to refer to terminal pitch movement: comma (low rise and other variants), period (falling), or question (rise-to-high) intonation (see relevant entries for a critical discussion of these terms).

Figures (2)-(6) below show acoustic visualizations of f0 contours (presented in Hz log to show contours in a way closer to human perception) in intonation units matching full TCUs in a variety of American English:

Figure 2: Fall

Jeffersonian 
I ga:ined about ten pounds.                                 

GAT-2 Fine
i `gA:ined about tEn `POUNDS;                               

Figure 3: Rise

Jeffersonian 
he:’s ↓lea:rni:ng?                               

GAT-2 Fine
he:s ↓´LEA:RNi:ng?

Figure 4: Level

Jeffersonian 
I’m ↑walking a↓rou:::nd                              

GAT-2 Fine
im ↑wAlking a↓¯ROU:::ND-

Figure 5: Fall-Rise

Jeffersonian 
So this time I’m taking adva:ntage of it,                              

GAT-2 Fine
so `thIs time im taking adˇVA:Ntage of it,

Figure 6: Rise-Fall

Jeffersonian 
I ↑ha↓:d one of those.                              

GAT-2 Fine
i ↑ˆHA:D one of those.

CA-IL work has found some regular co-occurrences of particular social actions with intonational configurations (see Couper-Kuhlen & Selting, 2017 and Ogden, 2021 for a thorough overview). The production of a complete intonation contour has been taken by some CA-IL work to be a cue for TCU completion (Ford & Thompson, 1996; but see Szczepek Reed, 2010). Sequences of matching intonation contours are taken as prosodic cues of the constitution of lists (Selting, 2007). Particular intonation contours have been associated to specific social actions, such as high rise-falls for repair on repeats indexing problems of acceptability (Benjamin & Walker, 2013) in English, and for displays of surprise (Reber, 2012). For French, Persson (2020) offers a complete overview of the role of repeats for repair (low rise to request completion) as well as forms of registering (rise + fall contour) or topicalising (rise). The role of intonation contours in action differentiation for different particles has been described at length. For example, Betz and Depperman (2021) describe the role of rising intonation in “okay” to index counter-to-expectation informings and not-yet-complete undertandings. Sørensen (2021) found that Danish response tokens ja and nej doing (dis)confirmation display affiliation with rising tones, and disaffiliation with level tones.

Work on the phonetics of talk-in-interaction, however, warns that form and function correspondences may not be as easy to come by in interaction (Walker, 2014) and that the intonation-social action relationship is context-sensitive and mediated by different orders of organization (Couper-Kuhlen & Selting, 2017). In this line, Kaimaki (2010) argues that in response-to-summons turns in Greek, falling and rising tones are in free variation; likewise, research on different kinds of interrogatives in interaction (e.g., Couper-Kuhlen, 2012) have debunked prior beliefs around the relationship between rising and falling tones and yes/no and wh-questions respectively.

Additional Related Entries:

• Period intonation
• Candidate understanding
• Question intonation
• Intonation contour
• Try-marking
• Comma intonation

Cited References:

Additional References: