730 F26, 10/15/26, Class 6.2 Tone and intonation

• Some perspectives:
  • Michaud (2008) ISSP paper
  • See also Michaud and Vaissière (2015)
  • Xu and Wang (1997)

Xu and Wang (1997): What can tone studies tell us about intonation?

• Note: mostly based on work on ("Standard") Mandarin

• What are the primitives of intonation?

See also Xu et al. (2015, p. 506-507):

One of the greatest difficulties in studying prosody is what can be referred to as ‘the lack of reference problem’ (Pierrehumbert 1980, 2000, Xu 2011a). That is, due to the general absence of orthographic representations of prosody other than punctuation, which itself may be due to a general difficulty in judging prosodic meanings by native speakers, there is little to fall back on when it comes to identifying prosodic units, whether in terms of their temporal location, scope, phonetic property or communicative function.

• Idea: let's start with something that is implemented via f0 contours that can be manipulated independently of intonation: lexical tone

treating lexical tones as contributors to the surface intonation form and using the observed variations of the their f0 contours as clues to the intonation primitives and the basic mechanisms that may underlie both tone and intonation (Xu and Wang 1997, p. 337)

• "Fundamental principle": "a tone is always aligned with the syllable it is associated with" (p. 337), and when not, it is because of...

  • late alignment: strong "carryover" coarticulation; early portion of syllable's f0 is strongly influenced by f0 contour from immediately preceding syllable
  • dynamic overshoot: since the early part of the syllable is taken, rising and falling in f0 contours is fast and starts near the end of the syllable and often continues into next syllable
  • anticipatory raising: If a high tone is followed by a low tone, it gets raised (see also pre-low raising work by Albert Lee and colleagues, in readings/w6).

• Articulatorily motivated cross-linguistic patterns/tendencies:

  • Tone spreading comes from late alignment and dynamic overshoot
  • Downstep comes from late alignment and anticipatory raising
  • Peak delay comes from dynamic overshoot

• Implications for intonational mechanisms

  • Declination cannot be physiological: see for instance, sequence of all Hs in Mandarin--it's just flat; it's local shapes of intonational melodies that can result in what looks like declination---like focus!
  • Focus: focal raising, post-focal lowering, not as much effect before focused region (at least in Mandarin)
  • IP/utterance-initial f0 peaks (introducing a new topic?) in English: big drop to next peak, and even higher than something in focus later. But no such non-linear f0 decline in Mandarin, Yoruba or Swedish.
  • Downstep: is general downtrend the result of successive downstep? Doesn't look like it at least in MAndarin--it's a local effect?

Michaud (2008): Tones and intonation: some current challenges

Tonal vs. non-tonal intonation

The daring theoretical move which consists in modelling lexical tones and intonation with the same phonological concept – tone – receives support from some tonal languages. There are well-established cases where intonation is encoded by tones that are treated on a par with lexical tones (section 3.1 provides examples); these cases will be referred to below as ‘tonal intonation’.1 However, a typological overview suggests that tonal intonation is relatively infrequent.” (Michaud, 2008, p. 1)

• Michaud's distinction between "tonal" and "non-tonal" intonation differs from Hyman's which was based more on use of f0 vs. other phonetic properties
  • Reminder: don't equate linguistic tone with f0, e.g., voice quality in Hanoi Vietnamese

In some languages, certain junctures of the utterance are characterised by the addition of boundary tones, which, though introduced by post-lexical rules, are integrated into the tone sequence of the utterance on a par with lexical tones. Such phenomena will be referred to here as tonal intonation (a phrase used in a different sense by [4]: see footnote to p. 1). L. Hyman (p.c.) points out that such phenomena are “rampant in African tone systems”. (Michaud, 2008, p. 2008)

• Examples:

  • boundary tone insertion (which could spread, changing phonological identity of other tones)
  • Naxi focus by adding a H tone at right edge of words with L or M tone on final syllable, changing tone of last syllable to a rise

• Tendency for languages with inventory of multiple levels to develop tonal intonation--contrast preservation, can't play with f0 range like in Mandarin, Vietnamese

Non-tonal intonation

non-tonal intonation is here defined as that part of intonation which is not implemented by the addition of tones.” (Michaud, 2008, p. 2008)

• no modification of underlying phonological sequence of tones---no insertion or deletion of tones

Level vs. contour tones, and contour tones decomposed into sequences of level tones

“Surprisingly, in the Austroasiatic and Tai-Kadai language families, no convincing evidence is found for the decomposition of contour tones into simpler units.” (Michaud, 2008, p. 2)

• Evidence for unitary contour hypothesis providing additional generalizations that a decomposed analysis cannot? Look to diachronic changes
  • Level tones: categorical phonological changes, e.g., high tone anticipation perhaps phonologized into an extra high tone in Moba upon loss of word-final vowels, cf. high tone raising in Gulmancema when followed by vowel with L tone.
  • Contour tones: defined "in terms of an overall shape", undergo gradual change
    • e.g., diachronic change in Thai in Pittayaporn (2007), see figure 2 below

• What about contour tones in intonation? AM theory is all about sequences of level tones...