Slide Slam S18 Sandbox Series
Undoable: computing hierarchical morphological structures in Aphasia
Kyan Salehi1, Caitlyn Antal1, Alexa Ray Falcone1, Laura Pissani1, Roberto G. de Almeida1; 1Concordia University
INTRODUCTION Linguistic productivity relies on the ability to compute morphologically complex hierarchical structures. This ability is mostly determined by accessing knowledge of selectional restrictions of roots and affixes. For instance, in a word such as “unsinkable” the prefix “un-“ attaches to the complex adjective sinkable, not to the verb “sink” (thus, ruling out “unsink”). Conversely, in the case of “unlockable”, both morphological structures can be computed: [un[lockable]] “not able to be locked” or [[unlock]able] “able to be unlocked”. As such, the correct parsing of these trimorphemic structures directly determines the derived meaning. Few experimental studies have investigated the parsing and interpretation of these types of words in isolation and in context (de Almeida & Libben, 2005; Libben, 2003; Libben, 2006; Pollatsek, Drieghe, Stockall, & de Almeida, 2010). Results have shown either right- or left-branching preference, with factors such as context and frequency affecting later rather than initial stages of analysis. We investigated morphological parsing in individuals with aphasia to understand (a) whether there is a default parsing strategy, (b) how sentential-semantic context influences parsing preferences, and (c) the breakdown of morpho-semantic processing across different clinical groups of aphasia. METHOD Participants were 12 individuals with aphasia (3 fluent [FL], 2 mixed [MX], 2 mixed but predominantly non-fluent [MN], 5 non-fluent [NF]). Controls were 30 healthy individuals matched to the clinical groups in age, sex, and education. All participants were native speakers of English. Stimuli consisted of 48 sentences containing ambiguous trimorphemic words (e.g., unlockable), with 24 biasing towards the left-branching and 24 towards the right-branching analysis of the trimorphemic word (e.g., ‘When the zookeeper went to unlock/lock the cage, he found it was unlockable’). In addition, materials included 24 sentences containing left-branching words ([[refill]able]) and 24 sentences containing right-branching words ([un[sinkable]]). These sentences were divided into two booklets, with each participant completing one booklet. Participants (a) rated how good each sentence was on a 5-point scale (Rating task: 1-bad, 5-good), and (b) indicated, by drawing a vertical line, where a separation could be made on target words (Parsing task) taken from each sentence and presented after the rating scale. RESULTS AND DISCUSSION Correct parsing was analyzed by items, considering word type (right-branching ambiguous, left-branching ambiguous, right-branching unambiguous, left-branching unambiguous) and group (controls, FL, MX, MN, NF), with repeated measures on the second factor. A cut before the suffix in the case of left-branching words ([[unlock]able], [[refill]able]) and a cut after the prefix for right-branching words ([un[lockable]], [un[sinkable]]) were considered correct. Results showed a main effect of word type and an interaction. Both the MX and the NF groups differed significantly from the control group across all word types, with the exception of a marginal difference for the MX group on right-branching words. Results are consistent with online experiments (de Almeida & Libben, 2005; Pollatsek et al., 2010) suggesting that the right-branching parse is preferred early in morphological analysis regardless of context. Notably, the NF group showed the inverse effect, indicating that the morphological parser can be affected in non-fluent aphasia.