The dynamics of intonation: categorical and continuous variation in an attractor-based model

Abstract

The framework of dynamical systems offers powerful tools to understand the relation between stability and variability in human cognition in general and in speech in particular. In the current paper, we propose a dynamical systems approach to the description of German nuclear pitch accents in focus marking to account for both the categorical as well as the continuous variation found in intonational data. We report on results from 27 native speakers and employ an attractor landscape to represent pitch accent types in terms of f0 measures in a continuous dimension. We demonstrate how the same system can account for both the categorical variation (relative stability of one prosodic category) as well as the continuous variation (detailed modifications within one prosodic category). The model is able to capture the qualitative aspects of focus marking such as falling vs. rising pitch accent types as well as the quantitative aspects such as less rising vs. more rising accents in one system by means of scaling a single parameter. Furthermore, speaker group specific strategies are analysed and modelled as differences in the scaling of this parameter. Thus, the model contributes to the ongoing debate about the relation between phonetics and phonology and the importance of variation in language and speech.

Evaluating generalised additive mixed modelling strategies for dynamic speech analysis

“…The simulations are run using three different data types: (i) a set of real F2 contours representing the diphthong /aI/ in New Zealand English (Sóskuthy et al, 2019); (ii) a set of pitch (F0) contours representing nuclear accented syllables produced with contrastive focus in German (Roessig et al, 2019); and (iii) a set of simulated F2 contours. In type I error simulations, each contour is randomly assigned to one of two different groups (referred to as A and B), resulting in a situation where the true curves underlying the data do not differ across the groups.…”

“…All simulations are carried out using the mgcv (Wood, 2017) and itsadug (van Rij et al, 2017) packages in R (R Core Team, 2013). Several different data sets are used: a set of formant measurements for the diphthong /aI/ in New Zealand English (Sóskuthy et al, 2019); a simulated data set modelled after these formant measurements; and pitch contours representing contrastive focus in German (Roessig et al, 2019). While real data are not commonly used in simulation studies, they play a crucial role in the current paper.…”

Modeling sonority in terms of pitch intelligibility with the Nucleus Attraction Principle

“…We assume here that perception and cognition are continuous processes that can activate categorical representations (see Case et al, 1995;Gafos, 2006;Lancia & Winter, 2013;Roessig, Mücke, & Grice, 2019 for derivation of linguistic categories from continuous inputs). These categorical representations are the learned symbols of the system in abstract terms, readily equated with co-activation of neural populations in the brain in physical terms (see, e.g., Friederici, 2011;Mesgarani, Cheung, Johnson, & Chang, 2014;Tang, Hamilton, & Chang, 2017).…”

“…In essence, works like those in Tuller et al (1994), Gafos (2006), Gafos and Benus (2006) and Roessig et al (2019) present a single model that incorporates continuous and symbolic aspects via the unifying mathematics of attractor landspaes. This is a necessary component in a system where continuous and symbolic entities are assumed to interact, modeling the manner in which the two types of entities can relate to each other.…”

Categorical perception of lexical tone contrasts and gradient perception of the statement–question intonation contrast in Zhumadian Mandarin

“…meaning, a claim for a morpheme with affective meaning has been made for a polar boundary tone in Dutch, which expresses positive affective meanings if its value is opposite to the upcoming tone of a pitch accent (Grabe, Gussenhoven, Haan, Marsi, & Post, 1998;Gussenhoven, 2004, p. 88) (see Table 1). Both gradient and discrete inputs to the speech signal are captured in the way Roessig, Mücke, and Grice (2019) model variation across as well as within intonational categories in German. In their dynamical system, stable states ('attractors') reflect phonological categories, such that within-category and between-category variation may affect attractor strengths as well as shifts in their definitions, which have been shown to vary as a function of semantics, giving different 'prototypes ' (Gili Fivela, 2013).…”