Does adding an accent mark hinder lexical access? Evidence from Spanish

Abstract

Recent research has shown that omitting the accent mark in a Spanish word, which is a language in which these diacritics only indicate lexical stress, does not cause a delay in lexical access (e.g., cárcel [prison] ≈ carcel; cárcel-CÁRCEL ≈ carcel-CÁRCEL). This pattern has been interpreted as accented and nonaccented vowels sharing the abstract letter representations in Spanish. However, adding an accent mark to a nonaccented Spanish word appears to produce a reading cost in masked priming paradigms (e.g., féliz-FELIZ [happy] > feliz-FELIZ). We examined whether adding an accent mark to a non accented Spanish word slows down lexical access in two semantic categorization experiments to solve this puzzle. We added an accent mark either on the stressed syllable (Experiment 1, e.g., cébra for the word cebra [zebra]) or an unstressed syllable (Experiment 2, e.g., cebrá). While effect sizes were small in magnitude, adding an accent mark produced a cost relative to the intact words, especially when the accent mark was added on an unstressed syllable (cebrá > cebra). These findings favor the view that letter identity and (to a lesser extent) accent mark information are encoded during word recognition in Spanish. We also examined the practical implications of these results.

Are goats chèvres, chévres, chēvres, and chevres? Unveiling the orthographic code of diacritical vowels.

“…Even though this post hoc analysis should be treated with caution, these findings seem to suggest that the bulk of the reading cost occurred mainly for those words in which the extra diacritical mark was placed on the letter e (i.e., there was a reading cost when reconstructing e from è in chèval). Instead, for those words in which the diacritical mark was placed on a letter other than e, the reading cost was small (8 ms; e.g., as in loup [lu] vs. loûp [lu]) and similar to that reported recently in Spanish (Labusch, Gómez, & Perea, 2022).…”

“…The focus of the analysis was to compare the intact words and the words with the omitted diacritic (or replaced/added diacritic in other experiments). For completeness, as in prior research (see Labusch, Gómez, & Perea, 2022; Labusch, Kotz, & Perea, 2022; Perea, Labusch, & Marcet, 2022), we also reported the effect of the category (animal vs. Nonanimal) and the two-way interaction between the two factors—note that “animals” and “nonanimals” require different responses. Thus, we prefer to analyze and report all outcomes…”

“…Notably, in a recent semantic categorization experiment in Spanish (i.e., a language in which diacritical and nondiacritical vowels presumably share their abstract letter units), Labusch, Gómez, and Perea (2022) found a small, but reliable cost of adding a diacritical mark in the unstressed syllable in Spanish (around 7-9 ms). For instance, the intact word cebra [ˈhebɾa] (zebra) was responded to faster than cebrá [hebɾˈa].…”

“…Second, in a language like Spanish, diacritical vowels do not modify individual phonemes (e.g., a and á are always pronounced /a/, and the same occurs for e-é /e/, i-í /i/, o-ó /o/, and u-ú /u/). Instead, they indicate lexical stress following some accentuation rules (see Appendix A in Labusch, Gómez, & Perea, 2022, for a detailed description 1 ). In this scenario, there would be no reason why diacritical and nondiacritical vowels would be represented as separate units in the mental lexicon (see Chetail & Boursain, 2019;Schwab, 2015).…”

“…Comparison of the Mean (SD) Characteristics of the Word Items in Experiment 1 words and the words with the omitted diacritic (or replaced/added diacritic in other experiments). For completeness, as in prior research (seeLabusch, Gómez, & Perea, 2022;Labusch, Kotz, & Perea, 2022; …”