Not Everybody Sees the Ness in the Darkness: Individual Differences in Masked Suffix Priming.

Joyse Medeiros,Jon Andoni Duñabeitia

doi:10.3389/fpsyg.2016.01585

Joyse Medeiros, Jon Andoni Duñabeitia

Open Access

https://doi.org/10.3389/fpsyg.2016.01585

Copy DOI

Abstract

The present study explores the role of individual differences in polymorphemic word recognition. Participants completed a masked priming lexical decision experiment on suffixed words in which targets could be preceded by suffix-related words (words sharing the same suffix) or by affixed primes with a different suffix. Participants also completed a monomorphemic word lexical decision and were divided in two groups (fast and slow readers) according to their performance in this task. When the suffix priming data were analyzed taking into consideration participants' reading speed as a proxy for their greater reliance on orthography or on semantics, a significant interaction between reading speed and the magnitude of the masked suffix priming effects emerged. Only slow participants showed significant priming effects, whereas faster participants showed negligible masked suffix priming effects. These results demonstrate that different reading profiles modulate the access to morphological information in a qualitatively different manner and that individual differences in reading determine the manner in which polymorphemic words are processed.

Highlights

Since the seminal study by Taft and Forster (1975), many studies in different languages have supported the view of a morphological decomposition process mediating lexical access for polymorphemic words, and together with sub-lexical and lexico-semantic variables, the morphological richness of words is a key factor in visual word recognition (e.g., Baayen et al, 2006)
We followed an approach based on generalized linear mixed-effect models (GLMM), using participants’ mean reaction time (RT) in the monomorphemic lexical decision task as a continuous fixed factor
Instead of using linear mixed-effect models (LMM), we opted for GLMM given that they are better suited for investigating individual differences by satisfying normality assumptions without requiring data transformation

Summary

Introduction

Since the seminal study by Taft and Forster (1975), many studies in different languages have supported the view of a morphological decomposition process mediating lexical access for polymorphemic words (see Rastle and Davis, 2008; Amenta and Crepaldi, 2012, for reviews), and together with sub-lexical and lexico-semantic variables, the morphological richness of words is a key factor in visual word recognition (e.g., Baayen et al, 2006). The evidence gathered from numerous masked priming studies has reinforced the assumption of automatic decomposition of morphologically complex words (e.g., Grainger et al, 1991; Taft, 2003; Rastle et al, 2004; Taft and Kougious, 2004; Longtin and Meunier, 2005). It is well known that prime-target pairs sharing their root morpheme (e.g., walker-WALK or revive-SURVIVE) activate each other, demonstrating that affixed words are decomposed into their corresponding morphemes (e.g., walk+er; e.g., Rastle et al, 2000; Longtin et al, 2003; Pastizzo and Feldman, 2004). Polymorphemic words sharing derivational suffixes (e.g., walker-DREAMER) activate each other, yielding masked suffix priming effects that emerge from the automatic decomposition of polymorphemic words (e.g., Duñabeitia et al, 2008). Compound words sharing one of their constituent lexemes (e.g., milkman-FIREMAN) have been shown to activate each other, demonstrating that morphemic parsing extends to other forms of polymorphemic words too (e.g., Duñabeitia et al, 2009; Crepaldi et al, 2013).

Objectives

Methods

Results

Conclusion