Articulatory Phonology1 What is a speech “gesture”?2 How are gestures coordinated with each other?3 How does this help explain phonological patterns?Note: this lecture is an overview of work done primarily by Catherine Browman and Louis Goldstein (Haskins Labs & Yale University).What is a speech gesture?1. Coordinative structure - articulators (and the muscles that move them) are organized to produce vocal tract constrictions.2. Dynamic system - using a spring-mass model to simulate the internal dynamics of articulator motion.note: the system we are discussing here is an implemented model of phonology. i.e. it is tied to a speech synthesizer and produces speech. It is a research system which is not complete. It is intended to point the way toward a complete model by outlining an approach that, in principle, can be expanded upon as more knowledge about the articulators is acquired and better speech synthesis methods are developed.Coordinative structures:Lip constriction (two gestural specifications)LP - lip protrusionLA - lip apertureTongue tip constrictionTTCL - tongue tip constriction locationTTCD - tongue tip constriction degreeTongue body constrictionTBCL - tongue body constriction locationTBCD - tongue body constriction degreeVelo-pharyngeal port constrictionVEL - velic apertureGlottal constrictionGLO - glottal apertureDynamics. The motions of the articulators are calculated as a damped mass-spring model:€ m˙ ˙ x + b˙ x + k( x − u) = 0m = mass (set to 0 in this implementation)b = damping (set to critical damping so there is no oscillation)k = the spring stiffnessx = the current location of the mass (€ ˙ x = velocity, € ˙ ˙ x = acceleration)u = the new (target) location of the massmkbxu2 How are gestures coordinated with each other?• A “gestural score” coordinates the gestures.• Relative timing specified in terms of phase relations among gesturesactivation intervalspecification of u and k for TDCL “narrow” and TBCD “phar”The gestural score (with its specifications of dynamic system parameters) then produces smooth time-varying “movement curves” of the articulators.3 How does this help explain phonological patterns?The gestural approach to phonology is good at (offers important insight into):• assimilation• deletion• resyllabification• metathesis (maybe)But not good at:• morphophonemics• dissimilation• perceptually driven sound change“perfect memory” spoken with an aspirated [t]Gestural hiding can cause deletion.“perfect memory” spoken with a deleted [t]but the tongue blade gesture is still there!Another case of “deletion” via gestural hiding.“nabbed most” pronounced as [næbmost],but the alveolar closure gesture is still there.“seven plus seven” with an alveolar [n] between [v] and [b].dip in lower lipvertical locationGestural hiding can cause assimilation.“seven plus seven” with an assimilated [m].but the tongue blade gesture is still there!Gestural blending can result in assimilation.Movement of the tongue tip in /l/, /t/ and /ld/ in Castilian Spanish.Place of articulation is backer for /l/ than for /t/ (and presumably also /d/.Place of articulation in the /ld/ sequence is a blend of the two.Romero, Joaquín (1996) Articulatory blending of lingual gestures. Journal of Phonetics 24:99–111.Gestural reorganization can lead to resyllabification.“kiss ted” spoken more and more quickly - at the fastest rate the [t] becomes [t] as if the syllable break moved [ks.td] -> [k.std].Phase of single glottal opening gesture in [st] is timed as for the cluster, losing the normal phasing for [t ].[s]Glottalaperture[t][s][s][st][t][t]Assimilation is different from gestural blending.Consider the /s∫/ sequence in “tops shop”Is it [∫], [∫∫], or a gesturally blended [s∫]?Notice that the /∫s/ sequence in “fish soup” behaves a little differently – with less apparent assimilation of [∫] to [s].Francis Nolan, Tara Holst and Barbara Kühnert (1996) Modelling [s] to [∫] accommodation in English. J. Phon. 24, 113-137Browman, C. P., & Goldstein, L. (1986). Towards an articulatory phonology. Phonology Yearbook, 3, 219-252.Browman, C. P., & Goldstein, L. (1989). Articulatory gestures as phonological units. Phonology, 6, 201-251.Browman, C. P., & Goldstein, L. (1990a). Gestural specification using dynamically-defined articulatory structures. Journal of Phonetics, 18, 299-320. Some