An articulatory description of the liquids of urban East Norwegian, based on EPG and EMA

The thesis examines different articulatory aspects involved in the production of the four liquids of Urban East Norwegian (UEN), on the basis of Electropalatography (EPG) and Electromagnetic Articulography (EMA) data from seven speakers. The set of sounds labelled "liquids" contains laterals and rhotics, or "l-sounds" and "r-sounds". In UEN one finds four different liquids, all coronal; an apical and a laminal lateral approximant, a tap (the most common realisation of an underlying /r/) and a flap (aka. "tykk l" ("thick l")). In this thesis I have looked for the defining articulatory characteristics of these four sounds, and also paid attention to both inter- and intra-speaker variation, as well as contextual variation. The data for this study is taken from the Electronic database of Norwegian speech sounds. EPG is a method for visualizing and recording the contact pattern of the tongue on the palate. The speaker wears a custom made artificial palate with 62 implanted electrodes, which are activated as the tongue makes contact with the palate. EMA registers the movement of the active articulators. A helmet containing transmitter coils create a magnetic field around the speaker's head, and four receiver coils are glued to the speaker's tongue along the mid-sagital line. As the articulators move during speech, the alternating voltages in the receiver coils are converted into distance measures. The EMA and EPG signals are displayed and recorded on a computer and create a dynamic image of the articulation, showing how the contact pattern on the palate and the movements of the tongue change over time. This study investigates several different articulatory variables. I have looked at what part of the tongue it is that creates the constriction, as well as where on the palate this constriction occurs. I have compared the maximum velocity for the tongue movement, the sounds' duration, and the distance the tongue travels during the articulation of these sounds. Since some of these four liquids have been described as "retroflex", I have also looked at a value labelled the "r tip y" index. This is an index of how much the tongue tip is curled upwards and towards the back of the mouth. I found great variability in the location of the lingual-palatal contact between the seven speakers, and also between each speaker's repetitions. Despite the great diversity, all speakers show the same relative location among the four liquids. What separates the tap and the flap from the laterals is their short duration. This agrees with the general consensus in phonetic literature, that the tap/flaps are the consonants that involve the shortest possible contact between the articulators. There is nevertheless a great difference between the tap and the flap, as the absolute distance travelled by the tongue tip when articulating the flap is more than five times that of the tap. The flap also involves the highest velocity of all the liquids. The main differences between the two lateral approximants seem to be the active part of the tongue and the relative location of the constriction. The investigation of the "r tip y" value showed that retroflexion is a possible but not necessary feature for all the UEN liquids, and not only for those traditionally labelled "retroflex"