Orthographic and root frequency effects in Arabic: Evidence from eye movements and lexical decision.

Abstract

One of the most studied and robust effects in the reading literature is that of word frequency.Semitic words (e.g., in Arabic or Hebrew) contain roots that indicate the core meaning to which the word belongs. The effects of the frequency of these roots on reading as measured by eye movements is much less understood. In a series of experiments, we investigated and replicated traditional word frequency effects in Arabic: Eye movement measures showed the expected facilitation for high-over low-frequency target words embedded in sentences (Experiment 1). The same was found in response time and accuracy in a lexical decision task (Experiment 3a). Using target words that were matched on overall orthographic frequency and other important variables, but that contained either high-or low-frequency roots, we found no significant influence of root frequency on eye movement measures during sentence reading (Experiment 2). Using the same target words in a lexical decision task (Experiment 3b), we replicated the absence of root frequency effects on real Arabic word processing. At first glance, the results may not appear to be in line with theoretical accounts that postulate early morphological decomposition and root identification when processing Semitic words.However, these results are compatible with accounts where morphological decomposition does occur but is followed by re-combination, and under certain conditions re-combination costs can eliminate or even reverse root frequency effects.

Insights from the study of Arabic reading

“…More support for the obligatory morphological decomposition of Semitic words into roots and patterns (e.g., Boudelaa, 2014; Frost et al, 1997), and for the delineation of the processes that take place after the morphological decomposition came from a recent study of Arabic word‐ and root‐frequency effects. Whereas Hermena et al (2019) replicated the word‐frequency effect observed in many world languages (see Rayner, 2009), frequency effects associated with Arabic roots were modulated by how often the roots occurred in combination with a particular pattern morpheme. Specifically, combining high‐frequency roots with patterns to produce low‐frequency words obliterated the potential benefit that might have otherwise been evident in the fixation durations during the reading of a high‐frequency root.…”

“…The results discussed in the previous paragraph were also interpreted as evidence that Semitic words are rapidly and obligatorily decomposed into their root and pattern morphemes in order to achieve complete word identification, and that this decomposition is not modulated by orthographic or semantic processing (Boudelaa, 2014; Frost et al, 1997; Hermena, Liversedge, Bouamama, & Drieghe, 2019). More support for the obligatory morphological decomposition of Semitic words into roots and patterns (e.g., Boudelaa, 2014; Frost et al, 1997), and for the delineation of the processes that take place after the morphological decomposition came from a recent study of Arabic word‐ and root‐frequency effects.…”

“…This expectation was evident if the parafoveal word did not carry diacritics or if the fixation prior to the word was too distant to afford perception of the diacritics. Finally, the third study suggested that the inclusion of diacritics to increase a word's orthographic transparency does not modulate the influence of its frequency, length or predictability, as the diacritized words are read more slowly, being the recipients of more, longer fixations and inflated re‐reading times (Hermena, Bouamama, Liversedge, & Drieghe, 2019). This pattern suggests that the presence of diacritics on a particular word will prompt readers to adopt a cautious reading strategy for the word and its surrounding region.…”

A dynamic adjustment account of word skipping in reading: Evidence from simulations and invisible boundary experiments

“…After the already discovered unimodal landing position distribution, it was not surprise that the anomalous previews affected saccade length through the dynamic adjustment mechanism. However, the visual anomalies had to be of sufficient size for being able to affect saccadic length, which seems to parallel Hermena et al (2018) finding no skipping effects for the absence of small diacritic marks in the reading of Arabic. Together, the available evidence suggests that any detectable information in blurred parafoveal vision is used to guide saccades in reading.…”

“…This finding suggests that the visual information that is available for saccadic modulation is functionally poorer than that later extracted via presaccadic attention. However, skipping rates are generally low in Arabic and also insensitive to parafoveal word frequency, making this complex orthography not an ideal testbed for studying skipping (Hermena et al, 2018).…”