Content area
This study examined how phonology, specifically word stress, influences the masked processing of English-suffixed words by non-native speakers. The study included four prime types: TP+ (visualize-VISUAL), TP− (temptation-TEMPT), FP+ (example-EXAM), and FP− (entertain-ENTER). Primes in TP+ (semantically transparent and phonologically congruent) and FP+ (form and phonologically congruent) conditions matched their targets regarding primary stress, whereas primes in TP− (semantically transparent and phonologically incongruent) and FP− (form and phonologically incongruent) conditions exhibited phonological variations compared to their targets. Two groups of English-Chinese bilinguals with different levels of English proficiency (advanced vs lower-intermediate) participated in the study. The results indicated that advanced Chinese–English bilinguals exhibited significant priming effects across all conditions, with TP+ producing a stronger priming effect than TP− and FP+. In contrast, lower-intermediate Chinese–English bilinguals only displayed priming effects for two form-related conditions. Additionally, advanced Chinese–English bilinguals demonstrated more robust priming effects for TP+ than lower-intermediate bilinguals. These findings suggest that in bilingual masked morphological processing, phonological effects facilitate early visual word recognition, while morpho-semantic relationships and L2 proficiency moderate both morphological and phonological effects during early morphological decomposition. These findings challenge the localist view of morphology as a discrete unit in the mental lexicon and support the connectionist view of morphological representations being distributed across spelling, sound, and meaning.
Introduction
The early stages of visual word recognition involve breaking down morphologically complex words (worker) into their corresponding stems (WORK) and affixes (-er). Previous studies have consistently demonstrated the automatic pre-lexical stripping of affixes in masked priming across different linguistic contexts (see Rastle and Davis, 2008 for a review). Specifically, strong masked priming effects were consistently detected for both Transparent (worker-WORK) and Opaque (corner-CORN) conditions for native speakers and bilinguals1. However, the primary L1–L2 difference concerns whether purely orthographic overlap (think-THIN) constitutes a priming effect. Relevant studies have consistently found no formal priming in native speakers (Davis and Rastle, 2010; Feldman et al., 2009; McCormick et al., 2008; Rastle and Davis, 2008; Rastle et al., 2004); however, significant priming has been observed in bilinguals (Diependaele et al., 2011; Heyer and Clahsen, 2015; Li and Taft, 2019; Li et al., 2017; Li et al., 2019).
In addition, previous research has primarily focused on semantics and orthography in early masked priming, leaving the role of phonology in early morphological decomposition understudied2. This gap in the literature is further complicated by inconsistencies in extant studies on the role of phonology in recognizing native and non-native words (Colombo and Zevin, 2009; Lin et al., 2014; Ou, 2019; Protopapas et al., 2016; Schiller et al., 2004; Sulpizio et al., 2012). Therefore, the purpose of this study was to investigate the extent to which phonology facilitates masked morphological processing of English-suffixed words by non-native speakers. The study also examined how L2 proficiency and morpho-semantic relations moderate bilinguals’ phonological processing, using a masked priming experiment with four experimental conditions. Two groups of Chinese–English bilinguals with different English proficiency levels participated. This research is significant as it enhances our understanding of phonology’s role in early masked morphological processing and how bilinguals’ phonological processing is moderated by L2 proficiency and morpho-semantic relations. The findings of this study have practical implications for language teaching and learning.
Literature review
Localist vs connectionist views in masked morphological processing
The representation and access of morphologically complex words are central to studies of morphological processing (Marslen-Wilson, 2007). Two main perspectives (discrete vs non-discrete) have been proposed to explain the mental representation of morphologically complex words (Harley, 2013, p. 228). The discrete perspective (also known as the localist view) proposes that morphologically complex words are represented independently in terms of their morphological structures, including semantics, orthography, and phonology (Feldman and Milin, 2018). Three localist models (supralexical, sublexical, and hybrid) have been proposed to explain the role of morpho-semantic relations in masked morphological processing (see Diependaele et al., 2012 for a review). The supralexical model (form-with-meaning/morpho-semantic) assumes that the whole word representation is accessed prior to morphological decomposition (Feldman et al., 2010; Feldman et al., 2009; Feldman and Soltano, 1999; Giraudo and Grainger, 2000, 2003; Gu, 2022), while the sublexical model (form-with-meaning/morpho-semantic) assumes that morphologically complex words must be parsed into sublexical components prior to whole word recognition (Davis and Rastle, 2010; Rastle and Davis, 2008; Rastle et al., 2000; Rastle et al., 2004). The hybrid model suggests that morphological processing involves both morpho-semantic and morpho-orthographic pathways (Diependaele et al., 2011; Diependaele et al., 2012; Diependaele et al., 2005, 2009; Diependaele et al., 2010; Duñabeitia et al., 2013).
This discussion of localist models provides a foundation for understanding recent advances in explaining processing disparities between native speakers and bilinguals. For example, facilitatory effects from orthographic overlap are typically observed in non-native speakers but not in native speakers. Grainger and Beyersmann (2017) proposed the “edge-aligned embedded word activation mechanism” to explain the processing disparity between native speakers and bilinguals. They argue that when a word is embedded at the beginning or end of a prime, it receives pre-activation, which facilitates processing. However, if the prime is a real word, it competes with the embedded word, counteracting its pre-activation. For example, in the word “brothel”, the embedded word “BROTH” initially receives pre-activation, typically facilitating its recognition. However, this pre-activation is counteracted by competition from the entire word “brothel”, which more strongly engages the cognitive processing system and inhibits the recognition of “BROTH” as a separate entity. This mechanism explains why native speakers do not experience priming with form-related words. In contrast, form priming has been observed in bilinguals, possibly due to less effective lexical inhibition in bilingual morphological processing. (e.g., Diependaele et al., 2011; Heyer and Clahsen, 2015; Taft et al., 2021).
Despite advances in studying the effects of morpho-semantic relations and orthography on visual word recognition, the localist model does not explicitly address phonology’s role in early morphological decomposition. In contrast, the non-discrete view (also known as the connectionist) proposes that morphological representation emerges from weighted connections between neuron-like processing units (Gonnerman et al., 2007; O’Grady, 2008; Plaut and Gonnerman, 2000). The connectionist view predicts numerically graded priming effects based on incremental similarity in semantics, orthography, and phonology, assuming that morphological effects are additive to the overall facilitation effect (Tang and Witzel, 2020). Specifically, high degrees of linguistic similarity are expected to produce strong morphological effects, whereas low degrees of linguistic similarity are likely to result in weak facilitation effects. The present study aims to investigate whether phonological information is necessary for obtaining masked morphological priming in second language processing, providing a new test for the connectionist model.
Phonology in word recognition
Previous studies have examined how semantic opacity and orthographic overlap affect morphological processing (Davis and Rastle, 2010; Feldman et al., 2010; Feldman et al., 2012; Feldman et al., 2009; Rastle and Davis, 2008; Rastle et al., 2004). Since phonology is subordinate to semantics and orthography, it is also reasonable to assume that the pre-activation of phonological information from primes will influence the phonological encoding of the targets during spoken word recognition, ultimately facilitating lexical recognition (Clahsen and Veríssimo, 2016; Sulpizio and Colombo, 2017; Sulpizio and Job, 2015). While it is widely agreed that fast and automatic phonological activation underlies silent reading (Bacovcin et al., 2017; Clahsen and Veríssimo, 2016; Diependaele et al., 2010; Perry et al., 2007), opinions differ on how much speech affects early word recognition. Two main theoretical perspectives are generally discussed: weak phonology and strong phonology. The “weak phonology” view holds that phonology’s influence on word recognition is minimal and indirect compared to the direct influence of orthography (Coltheart et al., 2001; Tang and Witzel, 2020). In contrast, the “strong phonology” view argues that phonology plays a critical role in word recognition, as sublexical grapheme-to-phoneme conversion facilitates lexical access (Frost, 1998; Sulpizio et al., 2012).
The priming paradigm has been widely used to investigate how phonological information influences word recognition (Diependaele et al., 2010). Phonological priming occurs when participants respond faster to phonologically similar pairs (pear-BEAR) than to phonologically dissimilar pairs (mice-BEAR). Facilitation in phonological priming is believed to result from the phonological overlap between primes and targets, with primes pre-activating the lexical representation of targets and reducing recognition time (Rastle and Brysbaert, 2006). Moreover, the degree of phonological priming may vary with the presentation modality: the cross-modal paradigm typically taps into later stages of lexical access, while the masked visual paradigm investigates relatively early stages of morphological processing (Gor, 2018; Rastle et al., 2004). In the cross-modal paradigm, where longer processing time is allowed, phonological information is more accessible to participants, resulting in larger phonological effects. Additionally, in the cross-modal paradigm, primes are auditory and targets are visual, whereas both primes and targets are visual in the masked priming paradigm. In the cross-modal paradigm, phonological information is obligatory because the auditory prime is salient enough for perception. Thus, stronger phonological effects are expected in the cross-modal paradigm than in the masked visual paradigm. In summary, phonological effects are likely stronger in the cross-modal paradigm due to (1) longer processing time and (2) obligatory aural perception.
Phonological priming in native and bilingual processing
Previous studies on phonology’s role in processing morphologically complex words in L1 speakers have typically viewed phonological effects as additive to overall morphological processing (Cooper et al., 2002; Donselaar et al., 2005; Soto-Faraco et al., 2001). Gonnerman et al. (2007) conducted a cross-modal lexical decision study with variable prime durations (due to auditory primes) and found numerically graded priming effects for words with various changes: (1) no segmental changes (agreement-AGREE); (2) consonantal changes (absorption-ABSORB); (3) vocalic changes (criminal-CRIME); and (4) both consonantal and vocalic changes (introduction-INTRODUCE). These graded priming effects suggest that phonological information plays a crucial role in lexical recognition for L1 speakers. However, inconsistent findings have been reported. Tang and Witzel (2020) found similar priming effects in phonologically congruent and phonologically incongruent conditions using a masked priming lexical decision task with a 50 ms prime duration. They concluded that phonology has a minimal effect at the earliest stages of morphological processing for L1 speakers. In contrast, other studies suggest that phonological information may be activated during the early stages of visual word recognition, though with small effects (e.g., Ferrand and Grainger, 1992, 1994; Grainger et al., 2003; Grainger and Ferrand, 1994; Rastle and Brysbaert, 2006). Rastle and Brysbaert (2006) conducted a meta-analytic literature review on the role of phonology in early visual word recognition, reviewing L1 studies that used homophones, short prime duration (14 ms ≤ SOA ≤ 72 ms), and various tasks, including perceptual identification, reading aloud, lexical decision, and text reading. The results showed that masked phonological priming effects, though small (r = 0.22), were reliably detected across different tasks, indicating phonological effects in masked visual word recognition for L1 speakers.
In addition to segments, suprasegmental features like word stress play a role in word recognition, especially in stress-timed languages such as English and Spanish, where word stress is contrastive and helps differentiate between syntactic functions and word meanings (e.g., abSTRACT_verb, /æbˈstrækt/ vs ABstract_noun, /ˈæbstrækt/). Previous studies on the effect of stress in word recognition have focused on three main factors: stress congruency (sharing the same stress patterns) (e.g., Cooper et al., 2002), stress dominance/typicality (the most frequent stress type) (e.g., Arciuli et al., 2010), and stress neighborhood consistency (alignment of orthographic endings and stress patterns) (e.g., Sulpizio and Colombo, 2013). For example, L1 studies using cross-modal naming tasks have shown that prime-target pairs with congruent word stress are named faster than those with incongruent stress patterns (e.g., Colombo and Zevin, 2009; Shaw, 2013; Sulpizio et al., 2012). These studies generally suggest that consistent stress patterns facilitate word recognition. However, findings in the L1 literature are inconsistent, with some studies reporting no constraining or facilitating effects from congruent stress (Protopapas et al., 2016; Schiller et al., 2004; Slowiaczek et al., 2006). For example, Schiller et al. (2004) used a cross-modal naming task to examine stress priming effects in Dutch but found no significant stress effects, attributing this to Dutch’s predictable word stress, where initial stress is dominant.
In addition to primary stress, other degrees of word stress may also influence word recognition. Banzina et al. (2016) explored how native and non-native productions of English secondary-stressed (SS) syllables (e.g., “comprehension”, /ˌkɑmpriˈhɛnʃən/) and unstressed-unreduced (UU) syllables (e.g., “robot”, /ˈroʊˌbɑt/) affected phonological processing in native English speakers. The cross-modal priming lexical decision study involved native English-spoken words, Russian-spoken English words with modifications, and unmodified Russian-spoken English words. The results showed that facilitation was strongest for native productions, followed by non-native productions with modifications, and weakest for unmodified non-native productions. Additionally, non-native UU with greater vowel and duration reduction caused more interference than non-native SS with less vowel and duration reduction, indicating that both vowel quality and duration impact word stress processing. These inconsistencies may stem from methodological differences, as phonological information is often activated at later stages of word recognition (e.g., cross-modal task) when it becomes fully accessible (Sulpizio and Colombo, 2017). In contrast, tasks designed to tap early word processing stages (e.g., masked visual tasks) often fail to detect phonological effects (Tang and Witzel, 2020).
Phonological priming has been shown to impact bilingual morphological processing. Cooper et al. (2002) conducted a cross-modal priming lexical decision task to examine how stress congruence affects spoken word processing in native and non-native English speakers. The study featured two parallel experiments using two-syllable and monosyllabic fragments as primes, respectively. Results showed that consistent stress primes elicited faster responses than inconsistent ones for both L1 and L2 speakers, suggesting that word stress aids word recognition in both groups. In contrast, proficient and intermediate French-Spanish bilinguals struggled to discriminate minimal stress pairs in Spanish (Dupoux et al., 2008), which the authors attributed to L1 transfer, as European French lacks a contrastive stress system. The Stress Parameter Model (SPM) (Dupoux et al., 2010; Dupoux et al., 2008) explains cross-linguistic differences by typological distances between languages, suggesting that speakers encode contrastive stress early in life if their native language sets parameters for it. However, the “stress deafness” hypothesis has not been confirmed in other studies. For instance, Tremblay (2008) found that Canadian French speakers (advanced vs lower-intermediate French–English bilinguals) were able to use English word stress for visual word recognition. Higher L2 proficiency is correlated with stronger priming effects, suggesting that stress congruence may be a universal feature, independent of linguistic typology. Nonetheless, the different syllable contrasts in Canadian French, compared to European French, may also have contributed to the phonological effects in Tremblay (2008), explaining why Canadian French speakers recognized spoken English words more easily than their European counterparts in Dupoux et al. (2008).
Previous studies on stress priming in Chinese–English bilinguals
Mandarin Chinese is traditionally considered a tonal language, but it also exhibits contrastive lexical stress (Duanmu, 2007). It should be noted that Duanmu also discussed various perspectives on lexical stress in Chinese, noting that while some believe Chinese lacks lexical stress, others recognize patterns of lexical stress similar to other languages. In this study, we adopt the perspective that contrastive lexical stress is present in Mandarin Chinese, while acknowledging the debate within the linguistic community. For example, minimal pairs like “东西” (/tʊ̄ŋ’ɕī/, “east–west”) and “东西” (/tʊ̄ŋ’ɕi/, “thing”) demonstrate a remarkable phonological contrast. This contrast arises from the stress patterns in the second syllable: the second syllable in the first word is stressed (/ɕī/), whereas it is unstressed (/ɕi/) in the second word. According to the Stress Parameter Model (SPM), Chinese–English bilinguals should be sensitive to English contrastive word stress due to the activation of stress parameters for lexical contrasts in Mandarin Chinese. However, few studies have directly investigated stress priming in Chinese–English bilinguals. Lin et al. (2014) compared the lexical recognition of English spoken words by native English speakers, Korean–English bilinguals, and Chinese–English bilinguals in a cross-modal lexical decision task. Results showed that Chinese–English bilinguals outperformed Korean–English bilinguals in detecting nonwords with incorrect stress patterns, suggesting that L1–L2 phonological similarity contributes to L2 lexical recognition. Ou (2019) investigated how word stress affects spoken word recognition in native English speakers and advanced Chinese–English bilinguals using an auditory fragment task. The results indicated that both groups exhibited comparable priming effects in stress-congruent and stress-incongruent conditions, suggesting that word stress may not significantly impact spoken word recognition. Overall, these studies provide inconclusive evidence regarding the impact of word stress on lexical recognition in bilinguals.
Current research questions and hypotheses
Previous research on phonology’s role in lexical recognition has been inconclusive in several respects. First, most L1 and L2 studies have focused on the cross-modal paradigm, which examines phonology’s role in later stages of lexical recognition when primes are presented aurally. Second, masked priming studies have yielded contradictory results regarding phonologies in the early stages of morphological processing. Third, no studies have examined the moderating role of L2 proficiency in masked phonological priming during non-native processing. To address these gaps and conflicts, this study investigates how phonology affects Chinese–English bilinguals’ masked morphological processing of English-suffixed words. Additionally, the study examines how morpho-semantic relations and L2 proficiency moderate masked morphological processing. Therefore, the study addresses three main research questions:
Research question 1: Is there evidence of phonological effects, particularly word stress, in non-native masked morphological processing of English-suffixed words?
Hypotheses: If phonology significantly influences non-native masked morphological processing, stress-consistent primes should facilitate responses, while stress-inconsistent primes should impede them. Conversely, if phonological information has minimal impact, there should be no significant difference in responses between stress-consistent and stress-inconsistent primes.
Research question 2: Do morpho-semantic relations play a role in masked morphological processing?
Hypothesis: If morpho-semantic relations influence masked morphological processing, semantically-transparent words should produce greater priming effects than form-related words.
Research question 3: Does L2 proficiency affect non-native masked morphological processing of English-suffixed words?
Hypothesis: If proficiency influences masked morphological processing, advanced bilinguals should exhibit stronger priming effects than less proficient bilinguals.
Method
We would like to clarify that the current study reused some data from the previously published study by Gu (2022), specifically for the suffix words. In particular, 60 participants who took part in this study had also completed the suffix word conditions (Transparent and Form) from Gu (2022). However, the research objectives of the two studies differ: the present study examines the impact of phonology on derived word processing, while Gu (2022) focused on the effects of affixes.
Participants
The study invited two groups of Chinese–English bilingual undergraduate students from Sichuan University. All participants shared similar linguistic backgrounds, originating from Sichuan province and belonging to the Han ethnic group. They are considered native speakers of both Mandarin Chinese and Southwest Mandarin, having used these two languages since birth. Notably, Southwest Mandarin lacks the word stress contrasts found in Standard Chinese, which minimizes potential confounding effects in this study (Qin et al., 2017). The first group consisted of 40 participants who were considered “advanced” in English, while the second group included 40 “lower-intermediate” participants. We would like to clarify that 20 participants were new to this study and had not participated in Gu (2022), while the majority (N = 60) had previously participated in Gu (2022), which explored the effects of prefixation and suffixation on non-native morphological processing.
The College English Test Band 6 (CET 6), a high-stakes exam required for Chinese college students, was used to measure participants’ English proficiency. CET 6 consists of four modules (listening, reading, writing, and speaking) and has a passing score of 425 out of a total of 710. This study focuses on early morphological processing, primarily relying on implicit linguistic knowledge, making the overall language proficiency scores sufficient for our analysis. Furthermore, during the screening phase, there were no instances of extremely high or low overall scores or disproportionate module scores. Based on Zhang et al. (2017) categorization of intermediate bilinguals, advanced Chinese–English bilinguals were determined to score above 575, while lower-intermediate bilinguals scored no more than 425. The independent t-test confirmed the significant proficiency difference in English between the two language groups (t = 60.93, df = 78, p < 0.01). However, both groups were matched for age, age of acquiring English (AOA), and years of learning English (YOL) as closely as possible (see Table 1 for details). Independent t-tests confirmed the matches (all p > 0.05, See Table S1 in the Data Availability section for details). Lastly, all individuals involved in the study had either normal vision or vision corrected to normal, and none exhibited any cognitive or learning impairments.
Table 1. Participants’ information.
Group | N | Age | AOA | YOL | CET6 | ||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
M | SD | M | SD | M | SD | M | SD | Min | Max | ||
Advanced | 40 | 19.73 | 1.09 | 9.63 | 2.16 | 10.35 | 2.24 | 597 | 15.98 | 575 | 622 |
Lower-intermediate | 40 | 19.48 | 0.99 | 9.70 | 2.16 | 10.05 | 2.02 | 395 | 13.56 | 375 | 423 |
AOA age of acquisition, YOL years of learning English, CET6 College English test band 6; the mean (M) comes before the standard deviation (SD).
Materials
The study utilized 65 morphologically simple words as target items, each paired with one of four morphologically complex prime words. Examples of these pairings are shown in Table 2. TP+ primes (visualize) were semantically transparent and phonologically congruent with their corresponding targets, retaining the word stress of their target (VISUAL). In contrast, TP− primes (temptation) were semantically transparent but phonologically incongruent with the corresponding targets, as they did not retain the word stress of their target (TEMPT). FP+ primes (example) were form and phonologically congruent, maintaining the word stress of their target (EXAM), whereas FP− primes (ENTERTAIN) were form and phonologically incongruent, not retaining the word stress of their target (ENTER). All primes and targets preserved other phonological information as much as possible, except for some segmental changes, which were evenly distributed across conditions and not expected to significantly affect processing. TP+ and TP− primes were related to their target words at semantic, morphological, and orthographic levels, while FP+ and FP− primes were related to their target words only at the orthographic level. It should be noted that the TP+ and FP+ conditions in the current study correspond to the Transparent and Form conditions for suffixed words in Gu (2022), respectively.
Table 2. Examples of prime and target per prime type.
Prime type | Prime | IPA | Target | IPA |
|---|---|---|---|---|
TP+ | visual | /ˈvɪʒ.u.əl/ | VISUALIZE | /ˈvɪʒ.u.ə.laɪz/ |
TP− | tempt | /tɛmpt/ | TEMPTATION | /tɛmpˈteɪ.ʃən/ |
FP+ | exam | /ɪɡˈzæm/ | EXAMPLE | /ɪɡˈzæm.pəl/ |
FP− | enter | /ˈɛn.tər/ | ENTERTAIN | /ˌɛn.tərˈteɪn/ |
TP+ semantically-transparent and phonologically-congruent, TP− semantically-transparent and phonologically-incongruent, FP+ form and phonologically-congruent, FP− form and phonologically-incongruent.
To examine the prime-target semantic relations, latent semantic analysis (Günther et al., 2015) was utilized. The LSA algorithm evaluates word presence and co-occurrence in texts by vectorizing word frequencies and adjusting them with a weighting scheme. It then maps these vectors into a semantic space to measure similarities, using the cosine of vector angles (from −1 to 1) to assess semantic relatedness. This process includes identifying prime-target word pairs and determining their relatedness with the “LSAfun” package. Semantic transparency was visually confirmed, ensuring that semantically transparent pairs were more similar than opaque pairs. The study found that TP+ and TP− primes were semantically closely related to their targets (0.1 < r < 1), whereas FP+ and FP− primes had low semantic relations to their targets (0 < r < 0.1, excluding words with an LSA of 0.1). Unrelated primes, which had no semantic relation to their targets, were included as a control. Primes and targets were matched as closely as possible in terms of word length, frequency, familiarity (rated by independent peers on a 7-point Likert scale), and letter overlap (measured by Levenshtein distance), based on values obtained from the English Lexicon Project (Balota et al., 2007).
It should be noted that a two-way ANOVA (PrimeType*Relatedness) confirmed this match, except for semantic relatedness, where the related primes were significantly more associated with the targets than the unrelated primes. Nonword targets were generated using Wuggy software (Keuleers and Brysbaert, 2010) for distraction purposes. Half of these nonword targets were preceded by suffixed real-word primes (e.g., volcanic-VAFE), while the other half were preceded by non-suffixed real-word primes (e.g., satellite-ULTION). The generation process considered the length, transition frequency, and sub-syllabic segments of the target words used in this study. It should be emphasized that the generated words did not resemble the critical items on the lists. Two counterbalanced lists were created, each containing targets paired with either related or unrelated primes. Participants viewed only one list and read the targets once. The critical stimuli information is provided in the Data Availability section (Tables S2, S3, and S4).
Procedure
The study used E-Prime 3.0 software (Psychology Software Tools, 2016) to present stimuli and collect data in a dimly lit computer lab. Participants were instructed to make lexical judgments about the targets as quickly and accurately as possible. Each trial began with a string of hash marks for 500 ms, followed by a lowercase prime for 50 ms (three times the monitor’s refresh cycle) and an uppercase target for 2500 ms. Stimuli were presented in 40-pt bold Arial black font on a white background, with a 60-Hz refresh rate and a 16.66 ms refresh cycle. Participants placed their fingers on two fixed keys (left shift for “real” and right shift for “not real”) to minimize delays. Ten practice trials were completed before the actual experiment. We would like to clarify that the experimental procedures in this study are largely identical to those in Gu (2022), as both are part of the same larger project investigating non-native derived word processing.
Data analysis
This study preserved the dataset’s integrity by including all responses, regardless of speed or accuracy, except for cases of clear experimental error, such as missing data (See Table S5 in the Data Availability section). To justify this decision, we converted reaction times to Z-scores within each group to detect outliers, finding only five data points exceeding three standard deviations from the group means (Jiang, 2013). A visual inspection of the data distribution also indicated normality, making data transformation unnecessary. We used linear mixed-effects models to analyze reaction time data, employing the lmerTest package in R (Kuznetsova et al., 2017). This modeling approach is appropriate as it accounts for both fixed effects (e.g., key variables in the masked priming experiments) and random effects (e.g., subjects and items).
For model selection (See Table S6 in the Data Availability section), we used a bottom-up approach, starting with individual fixed effects (proficiency, prime type, and relatedness) and progressively adding two-way and three-way interaction terms. After identifying the optimal fixed effects structure, we incrementally introduced random intercepts for subjects and items, assessing model fit using ANOVA (for models without random effects), likelihood ratio tests (LRT) (for models with random effects), Akaike Information Criterion (AIC), and Bayesian Information Criterion (BIC). Random slopes for each fixed effect were then tested to refine the model. We then tested random slopes for each fixed effect to further refine the model. Finally, model criticism was conducted to check for non-linearity, non-normality of residuals, heteroscedasticity, multicollinearity, or the presence of outliers and influential points. All R codes have been extracted and are now provided in a .R file, available in the Data Availability section on the Open Science Framework.
Results
The final model included the main effects of all fixed variables (proficiency, prime type, and relatedness), their two-way and three-way interactions, and random intercepts for items. Model criticism confirmed that all assumptions were met. Fixed effects omnibus tests showed significant results for the following: (1) proficiency [F(1, 5123) = 1088.26, p < 0.001] and relatedness [F(1, 5123) = 68.75, p < 0.001]; (2) a two-way interaction between proficiency and prime type [F(3, 5132) = 12.06, p < 0.001]; (3) a three-way interaction among proficiency, prime type, and relatedness [F(3, 5132) = 6.39, p < 0.001].
Following the significant three-way interaction effect, we conducted a series of post-hoc comparisons to assess the priming effects by evaluating the simple effect of relatedness. Once significant priming effects were identified, we used interaction contrasts to measure differences in priming effects (e.g., TP+ vs TP−). To calculate the priming effects, we subtracted the related response time (RT) from the unrelated RT (as shown by the black rectangles in Figure S1 in the Data Availability section). We then compared the magnitude of priming between the following pairs: (1) TP+ vs TP−, FP+ vs FP−, TP+ vs FP+, TP− vs FP− (prime type effect, shown by the red dotted line); (2) advanced vs lower-intermediate (proficiency effect, shown by the blue dotted line). All p-values were adjusted for multiple comparisons using the Holm method (Ludbrook, 1998).
Table 3 presents the results for the simple effects of relatedness, with significant priming effects marked by an asterisk. Advanced Chinese–English bilinguals showed significant priming effects across all four prime types: TP+ (t = 5.23, SE = 19.57, p < 0.001), TP− (t = 4.35, SE = 19.57, p < 0.001), FP+ (t = 2.08, SE = 22.60, p = 0.038), and FP− (t = 2.46, SE = 27.68, p = 0.014). In contrast, lower-intermediate Chinese–English bilinguals exhibited significant priming effects only in FP+ (t = 5.09, SE = 22.60, p < 0.001) and TP− (t = 3.39, SE = 27.68, p < 0.001).
Table 3. By-group performance for each prime type.
Prime type | Relatedness | Advanced | Lower-intermediate |
|---|---|---|---|
TP+ | Unrelated | 899 (245) | 1171 (284) |
Related | 797 (229) | 1154 (311) | |
Effect | 102*** | 17 | |
TP− | Unrelated | 962 (249) | 1124 (317) |
Related | 877 (269) | 1123 (314) | |
Effect | 85*** | 1 | |
FP+ | Unrelated | 924 (246) | 1200 (297) |
Related | 878 (246) | 1085 (298) | |
Effect | 46* | 115*** | |
FP− | Unrelated | 901 (248) | 1207 (286) |
Related | 833 (248) | 1113 (311) | |
Effect | 68* | 94*** |
The “Effect” is calculated by subtracting the related RT from the unrelated RT, with the standard deviation shown in brackets.
***p < 0.001; **p < 0.01; *p < 0.05.
To examine the moderating effects of prime type (related to research questions 1 and 2 and their respective hypotheses) and proficiency (related to research question 3 and its corresponding hypotheses), we conducted interaction contrasts. The prime type effect revealed that advanced Chinese–English bilinguals had larger priming effects for TP+ compared to TP− (t = 2.72, SE = 27.64, p = 0.04) and FP+ (t = 3.29, SE = 29.86, p = 0.008). In contrast, lower-intermediate Chinese–English bilinguals produced equivalent priming effects across all four prime types (all ps > 0.05). For the proficiency effect, advanced Chinese–English bilinguals showed larger priming effects than lower-intermediate bilinguals for TP+ (t = 3.10, SE = 27.64, p = 0.002) and TP− (t = 3.01, SE = 27.64, p = 0.003), while both groups exhibited similar priming effects for FP+ and FP− (ps > 0.05).
Discussion
This study is the first to investigate whether bilingual speakers automatically activate phonological information during the early stages of visual word recognition when processing morphologically complex words. It focused on three main research questions: (1) whether phonological effects can be distinguished from morphological effects in the non-native masked morphological processing of English-suffixed words; (2) whether morpho-semantic relations influence this process; and (3) whether L2 proficiency moderates it. Methodologically, the study used a masked priming lexical decision task with two groups of Chinese–English bilingual speakers at different English proficiency levels. Four experimental conditions were included: TP+ (semantically-transparent and phonologically-congruent; e.g., visualize-VISUAL), TP− (semantically-transparent and phonologically-incongruent; e.g., temptation-TEMPT), FP+ (form and phonologically-congruent; e.g., example-EXAM), and FP− (form and phonologically-incongruent; e.g., entertain-ENTER).
Significant priming effects were found across all four experimental conditions (TP+, TP−, FP+, and FP−) for advanced Chinese–English bilinguals. Interaction contrasts revealed a prime-type effect only for this group, with priming differences observed in the TP+ vs TP− and TP+ vs FP+ comparisons exclusively among highly proficient Chinese–English bilinguals. The differences between TP+ and TP− suggest that word stress plays a crucial role in the non-native masked morphological processing of English-suffixed words. However, the absence of masked priming between FP+ and FP− among advanced speakers suggests that phonological effects may be influenced by morpho-semantic relations. Specifically, phonological effects were more pronounced for word pairs with morpho-semantic relations (e.g., worker-WORK), indicating that these relations can amplify phonological effects during masked priming.
Our study differed from previous research with native English speakers (e.g., Tang and Witzel, 2020) primarily in terms of the stimuli used. While Tang and Witzel (2020) focused on phonological segments, we investigated the role of word stress in masked priming. Stress variation may be more perceptually salient than segmental variation, which could explain why advanced bilinguals were more sensitive to stress changes between primes and targets during masked priming. Moreover, the differences observed between TP+ and FP+ for advanced Chinese–English bilinguals suggest that morphological priming arises not only from orthographic overlap but also from morpho-semantic connections. Lastly, the negligible differences found between TP− and FP− (compared to TP+ vs FP + ) for advanced Chinese–English bilinguals suggest that morpho-semantic effects may be diminished by phonological inconsistency.
In contrast to advanced Chinese–English bilinguals, lower-intermediate bilinguals exhibited significant priming effects only for FP+ and FP−. No significant differences were observed among the four experimental conditions for this group. The formal priming observed in lower-intermediate bilinguals suggests that less-proficient learners rely on orthographic overlap to process form-related words. Compared to monolingual English speakers in previous studies, this formal priming in lower-intermediate bilinguals may result from a lack of lexical competition in the non-native mental lexicon (Grainger and Beyersmann, 2017; Qiao et al., 2009). This ineffectiveness of lateral inhibition from formal sets to sufficiently suppress the bottom-up activation of embedded words within them (e.g., think-THIN) results in form priming among bilingual speakers (Diependaele et al., 2011; Heyer and Clahsen, 2015; Li and Taft, 2019).
Additionally, the lack of priming in the morphological set and the absence of differences between TP+ vs TP−, FP+ vs FP−, TP+ vs PF+, and TP− vs FP− suggest that, when masked, (1) less-proficient bilinguals lack access to morpho-semantic relations when processing morphological sets, and (2) phonological information is unavailable to lower-intermediate bilinguals. These findings suggest that language proficiency moderates the behavioral performance of non-native speakers (Gu, 2022).
The moderating effects of proficiency were evident in three ways. First, advanced Chinese–English bilinguals showed more robust priming effects compared to lower-intermediate Chinese–English bilinguals. Second, morpho-semantic and phonological effects were exclusive to advanced Chinese–English bilinguals. Third, advanced bilinguals exhibited stronger priming than lower-intermediate bilinguals for TP+ and TP−. The priming effects for semantically transparent words cannot be attributed solely to formal overlap, as the effect was greater for TP+ than FP+. Additionally, morpho-semantic facilitation was indirectly reflected in the comparable phonological priming observed between FP+ and FP−, unlike the larger phonological effects between TP+ and TP−. Conversely, the similar priming observed between TP− and FP− suggests that phonological consistency may offset the influence of morpho-semantics, leading to minimal differences between the two conditions.
Regarding the effect of modality, studies using experimental procedures that tap into later stages of morphological processing (e.g., cross-modal paradigms) have found significant phonological priming effects, whereas phonological effects are less commonly observed in masked visual priming during early word recognition. This suggests that phonology may not play a critical role in the early stages of morphological processing when phonological information is not yet fully accessible to the speakers. However, the present study found significant phonological effects (specifically, primary stress) using a masked priming lexical decision task in bilinguals. Along with previous studies (e.g., Gonnerman et al., 2007; Tang and Witzel, 2020) that identified phonological effects in both early and late stages of morphological processing, these results suggest that phonological effects may be independent of modality (visual vs auditory).
Furthermore, the salience of phonological features may influence phonological effects in masked priming. For instance, primary stress is acoustically more prominent than segments, making stress alternation more likely to capture speakers’ attention. Additionally, the duration of the prime and the length of the target may impact the likelihood of masked phonological priming. Specifically, masked phonological priming is more likely to occur with short primes presented for longer durations and shorter target words (e.g., Ferrand and Grainger, 1992, 1994; Grainger et al., 2003; Grainger and Ferrand, 1994; Rastle and Brysbaert, 2006), compared to shorter primes and longer target words (e.g., Tang and Witzel, 2020). This could explain the mixed findings regarding phonological priming across different masked priming studies.
This study challenges the view that morphemes are explicitly represented in the mental lexicon for bilingual morphological processing (Feldman et al., 2009; Rastle et al., 2004). Instead, it supports the connectionist model, which posits that morphemic representations are distributed across spelling, sound, and meaning (Gonnerman et al., 2007; Plaut and Gonnerman, 2000). Specifically, the findings suggest that phonology plays a crucial role, particularly in the early stages of recognizing visually presented morphologically complex words. Thus, alongside morpho-semantic information (Feldman et al., 2009) and morpho-orthographic information (Rastle et al., 2004), phonological information may be a key factor in morphological decomposition for bilingual speakers.
This study aligns with previous research on bilingual processing of word stress (e.g., Cooper et al., 2002; Lin et al., 2014; Tremblay, 2008) and supports the Speech Parameter Model (SPM), which predicts processing facilitation based on close linguistic distances (Dupoux et al., 2008). Since Mandarin Chinese employs similar weak-strong phonological contrasts to distinguish minimal pairs, as in English, Chinese–English bilinguals are expected to face minimal difficulty with English word stress. In this study, advanced Chinese–English bilinguals demonstrated stress priming (with stronger priming for TP+ than TP−), providing new evidence in favor of the SPM model. However, our findings contrast with previous studies, such as Ou (2019), which reported no stress priming effects for advanced Chinese–English bilinguals.
One possible explanation for these inconsistent results may lie in the research methods. Specifically, our study used whole English words as stimuli, whereas Ou (2019) used truncated forms of words. It is possible that word fragments do not provide full phonological access and, thereby fail to elicit phonological priming. This view is further supported by Lin et al. (2014), who used whole words and observed reliable stress priming effects in both native English speakers and advanced Chinese–English bilinguals. Beyond restricted phonological access, the lack of phonological priming in Ou (2019) may also result from the absence of morpho-semantic relations between truncated primes and targets. Our study found reliable stress priming effects for TP+ and TP− rather than FP+ and FP−, suggesting that morpho-semantic relations may influence phonological priming. Future research should carefully consider the experimental designs before making cross-study comparisons.
Conclusion
This study aimed to explore three research questions: (1) whether fast morphological effects are distinct from phonological information in bilingual morphological processing; (2) whether L2 proficiency moderates masked processing of non-native English suffixed words; and (3) whether morphological priming can be exclusively attributed to orthographic overlap. To address these questions, we conducted a masked priming task with four critical conditions, testing two groups of Chinese–English bilinguals with different levels of English proficiency (advanced vs lower-intermediate). Our findings revealed that advanced Chinese–English bilinguals exhibited a numerical advantage for semantically transparent and phonologically congruent primes, while lower-intermediate Chinese–English bilinguals displayed indistinguishable priming effects only for form-related words. These results suggest that bilingual morphological processing relies not only on orthographic overlap but also on phonological information. Moreover, the results indicate that L2 proficiency moderates these effects, with advanced bilinguals showing greater sensitivity to the morpho-semantic and phonological structures of English-suffixed words. Overall, the study suggests that discrete morphological representations may not play a central role in bilingual morphological processing. Instead, our findings align more closely with the connectionist model, which proposes that morphological representations are distributed over spelling, sound, and meaning.
Implications and limitations
The study suggests that for second language learners to effectively process morphologically complex words, attention to both phonological information and proficiency level is essential. Enhancing phonological awareness can improve their ability to acquire and process new vocabulary. Similarly, targeted instruction in morpho-semantic and phonological awareness may prove beneficial.
However, this study has several limitations. First, it focuses exclusively on Chinese–English bilinguals without including a native English speaker control group. Although the primary aim was not to compare L1–L2 differences in morphological processing, the absence of native speakers remains a limitation. Additionally, the use of a masked priming task may not fully capture natural reading processes, and the study’s limited set of primes and targets further constrains its scope. Moreover, individual differences in phonological awareness were not measured, and the cross-sectional design limits the ability to infer causal relationships.
Author contributions
The first author led the study, drafted the manuscript, and participated in its final review. The second author contributed to specific sections and assisted in refining the final draft.
Data availability
All materials referenced in this study are available on the Open Science Framework (OSF) at https://doi.org/10.17605/OSF.IO/REQFU.
Competing interests
The authors declare no competing interests.
Ethical approval
The study obtained ethical approval from the Survey and Behavioral Research Ethics Committee (SBRE) of the Faculty of Arts, Chinese University of Hong Kong, on March 19, 2020 (approval no. SBRE-20-356). All procedures adhered to the ethical principles specified in the Declaration of Helsinki.
Informed consent
Prior to participation, all participants were fully informed about the study’s objectives, procedures, data usage, their right to remain anonymous, and their right to withdraw at any time. Written informed consent was obtained by the first author from all participants on May 19, 2020.
It should note that in the Opaque condition, the pseudo-derived prime (corner) is related to the corresponding target (CORN) by the pseudo-suffix “-er”. Besides, opacity should be interpreted on a continuum (e.g., neither “ponytail” nor “cocktail” have anything to do with the animals used as first constituents, but a “ponytail” is nevertheless more of a tail than a “cocktail”).
2Morphological decomposition refers to the process by which complex words are broken down into smaller units called morphemes, which are the smallest units of meaning in a language.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
Arciuli, J; Monaghan, P; Seva, N. Learning to assign lexical stress during reading aloud: corpus, behavioral, and computational investigations. J Mem Lang; 2010; 63,
Bacovcin, H; Davies, A; Wilder, R; Embick, D. Auditory morphological processing: evidence from phonological priming. Cognition; 2017; 164, pp. 102-106. [DOI: https://dx.doi.org/10.1016/j.cognition.2017.03.011] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/28395153][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6953625]
Balota, DA; Yap, MJ; Hutchison, KA; Cortese, MJ; Kessler, B; Loftis, B; Neely, JH; Nelson, DL; Simpson, GB; Treiman, R. The English lexicon project. Behav Res Methods; 2007; 39,
Banzina, E; Dilley, LC; Hewitt, LE. The role of secondary-stressed and unstressed-unreduced syllables in word recognition: acoustic and perceptual studies with Russian learners of English. J Psycholinguist Res; 2016; 45,
Clahsen, H; Veríssimo, J. Investigating grammatical processing in bilinguals. Linguisti Approaches Biling; 2016; 6,
Colombo L, Zevin J (2009) Stress priming in reading and the selective modulation of lexical and sub-lexical pathways. PLoS ONE. https://doi.org/10.1371/journal.pone.0007219
Coltheart, M; Rastle, K; Perry, C; Langdon, R; Ziegler, J. DRC: a dual route cascaded model of visual word recognition and reading aloud. Psychol Rev; 2001; 108,
Cooper, N; Cutler, A; Wales, R. Constraints of lexical stress on lexical access in English: evidence from native and non-native listeners. Lang Speech; 2002; 45,
Davis, MH; Rastle, K. Form and meaning in early morphological processing: comment on Feldman, O’Connor, and Moscoso del Prado Martín (2009). Psychon Bull Rev; 2010; 17,
Diependaele, K; Duñabeitia, JA; Morris, J; Keuleers, E. Fast morphological effects in first and second language word recognition. J Mem Lang; 2011; 64,
Diependaele K, Grainger J, Sandra D (2012) Derivational morphology and skilled reading: an empirical review. In: Spivey MJ, McRae K, Joanisse MF (eds) The Cambridge handbook of psycholinguistics. Cambridge University Press, pp 311–332
Diependaele, K; Sandra, D; Grainger, J. Masked cross-modal morphological priming: Unravelling morpho-orthographic and morpho-semantic influences in early word recognition. Lang Cogn Process; 2005; 20,
Diependaele, K; Sandra, D; Grainger, J. Semantic transparency and masked morphological priming: the case of prefixed words. Mem Cogn; 2009; 37,
Diependaele, K; Ziegler, JC; Grainger, J. Fast phonology and the bimodal interactive activation model. Eur J Cogn Psychol; 2010; 22,
Donselaar, WV; Koster, M; Cutler, A. Exploring the role of lexical stress in lexical recognition. Q J Exp Psychol; 2005; 58,
Duanmu S (2007) The phonology of standard Chinese. Oxford University Press
Duñabeitia, JA; Dimitropoulou, M; Morris, J; Diependaele, K. The role of form in morphological priming: evidence from bilinguals. Lang Cogn Process; 2013; 28,
Dupoux, E; Peperkamp, S; Sebastián-Gallés, N. Limits on bilingualism revisited: stress ‘deafness’ in simultaneous French–Spanish bilinguals. Cognition; 2010; 114,
Dupoux, E; Sebastián-Gallés, N; Navarrete, E; Peperkamp, S. Persistent stress ‘deafness’: the case of French learners of Spanish. Cognition; 2008; 106,
Feldman, LB; Kostić, A; Basnight-Brown, DM; ĐURĐEVIĆ, DF; Pastizzo, MJ. Morphological facilitation for regular and irregular verb formations in native and non-native speakers: Little evidence for two distinct mechanisms. Biling Lang Cogn; 2010; 13,
Feldman, LB; Kostić, A; Gvozdenović, V; O’Connor, PA; del Prado Martín, FM. Semantic similarity influences early morphological priming in Serbian: a challenge to form-then-meaning accounts of word recognition. Psychon Bull Rev; 2012; 19,
Feldman LB, Milin P (2018) Psycholinguistic studies of word morphology and their implications for models of the mental lexicon and lexical processing. In: Berthiaume R, Daigle D, Desrochers A (eds) Morphological processing and literacy development: current issues and research. Routledge, pp 16–47
Feldman, LB; O’Connor, PA; del Prado Martín, FM. Early morphological processing is morphosemantic and not simply morpho-orthographic: a violation of form-then-meaning accounts of word recognition. Psychon Bull Rev; 2009; 16,
Feldman, LB; Soltano, EG. Morphological priming: the role of prime duration, semantic transparency, and affix position. Brain Lang; 1999; 68,
Ferrand, L; Grainger, J. Phonology and orthography in visual word recognition: evidence from masked non-word priming. Q J Exp Psychol; 1992; 45,
Ferrand, L; Grainger, J. Effects of orthography are independent of phonology in masked form priming. Q J Exp Psychol Sect; 1994; 47,
Frost, R. Toward a strong phonological theory of visual word recognition: true issues and false trails. Psychol Bull; 1998; 123,
Giraudo, H; Grainger, J. Effects of prime word frequency and cumulative root frequency in masked morphological priming. Lang Cogn Process; 2000; 15,
Giraudo H, Grainger J (2003) A supralexical model for French derivational morphology. In: Assink E, Sandra D (eds) Reading complex words. Kluwer, pp 139–157. https://doi.org/10.1007/978-1-4757-3720-2_7
Gonnerman, LM; Seidenberg, MS; Andersen, ES. Graded semantic and phonological similarity effects in priming: evidence for a distributed connectionist approach to morphology. J Exp Psychol Gen; 2007; 136,
Gor, K. Phonological priming and the role of phonology in nonnative word recognition. Biling Lang Cogn; 2018; 21,
Grainger J, Beyersmann E (2017) Edge-aligned embedded word activation initiates morpho-orthographic segmentation. In: Ross BH (ed) Psychology of learning and motivation, vol 67. Elsevier, pp 285–317. https://doi.org/10.1016/bs.plm.2017.03.009
Grainger, J; Diependaele, K; Spinelli, E; Ferrand, L; Farioli, F. Masked repetition and phonological priming within and across modalities. J Exp Psychol Learn Mem Cogn; 2003; 29,
Grainger, J; Ferrand, L. Phonology and orthography in visual word recognition: effects of masked homophone primes. J Mem Lang; 1994; 33,
Gu L (2022) The processing of English prefixed and suffixed words by Chinese–English bilinguals. Int J Biling. https://doi.org/10.1177/13670069221137454
Günther, F; Dudschig, C; Kaup, B. LSAfun—an R package for computations based on latent semantic analysis. Behav Res Methods; 2015; 47,
Harley TA (2013) The psychology of language: from data to theory. Psychology Press
Heyer, V; Clahsen, H. Late bilinguals see a scan in scanner AND in scandal: dissecting formal overlap from morphological priming in the processing of derived words. Biling Lang Cogn; 2015; 18,
Jiang N (2013) Conducting reaction time research in second language studies. Routledge
Keuleers, E; Brysbaert, M. Wuggy: a multilingual pseudoword generator. Behav Res Methods; 2010; 42,
Kuznetsova A, Brockhoff PB, Christensen RHB (2017) lmerTest package: tests in linear mixed effects models. J Stat Softw. https://doi.org/10.18637/jss.v082.i13
Li, J; Taft, M. The processing of English prefixed words by Chinese–English bilinguals. Stud Second Lang Acquis; 2019; 42,
Li, J; Taft, M; Xu, J. The processing of English derived words by Chinese–English bilinguals. Lang Learn; 2017; 67,
Li, J; Taft, M; Zhu, Z; Zhang, B. A difference in processing prefixed and suffixed words by Chinese EFL learners. Mod Foreign Lang; 2019; 42,
Lin, CY; Wang, M; Idsardi, WJ; Xu, Y. Stress processing in Mandarin and Korean second language learners of English. Biling Lang Cogn; 2014; 17,
Ludbrook, J. Multiple comparison procedures updated. Clin Exp Pharmacol Physiol; 1998; 25,
Marslen-Wilson W (2007) Processes in language comprehension. In: Gaskell MG (ed) The Oxford handbook of psycholinguistics. Oxford University Press, pp 495–524
McCormick, SF; Rastle, K; Davis, MH. Is there a ‘fete’in ‘fetish’? Effects of orthographic opacity on morpho-orthographic segmentation in visual word recognition. J Mem Lang; 2008; 58,
O’Grady, W. Innateness, universal grammar, and emergentism. Lingua; 2008; 118,
Ou, SC. The role of lexical stress in spoken English word recognition by listeners of English and Taiwan Mandarin. Lang Linguist; 2019; 20,
Perry, C; Ziegler, JC; Zorzi, M. Nested incremental modeling in the development of computational theories: the CDP+ model of reading aloud. Psychol Rev; 2007; 114,
Plaut, DC; Gonnerman, LM. Are non-semantic morphological effects incompatible with a distributed connectionist approach to lexical processing?. Lang Cogn Process; 2000; 15,
Protopapas, A; Panagaki, E; Andrikopoulou, A; Gutiérrez Palma, N; Arvaniti, A. Priming stress patterns in word recognition. J Exp Psychol: Hum Percept; 2016; 42,
Psychology Software Tools (2016) E-Prime 3.0. https://www.pstnet.com
Qiao, X; Forster, K; Witzel, N. Is banara really a word?. Cognition; 2009; 113,
Qin, Z; Chien, YF; Tremblay, A. Processing of word-level stress by Mandarin-speaking second language learners of English [Article]. Appl Psycholinguist; 2017; 38,
Rastle, K; Brysbaert, M. Masked phonological priming effects in English: Are they real? Do they matter?. Cogn Psychol; 2006; 53,
Rastle, K; Davis, MH. Morphological decomposition based on the analysis of orthography. Lang Cogn Process; 2008; 23,
Rastle, K; Davis, MH; Marslen-Wilson, WD; Tyler, LK. Morphological and semantic effects in visual word recognition: a time-course study. Lang Cogn Process; 2000; 15,
Rastle, K; Davis, MH; New, B. The broth in my brother’s brothel: morpho-orthographic segmentation in visual word recognition. Psychon Bull Rev; 2004; 11,
Schiller, NO; Fikkert, P; Levelt, CC. Stress priming in picture naming: an SOA study. Brain Lang; 2004; 90,
Shaw, JA. The phonetics of hyper-active feet: effects of stress priming on speech planning and production. Lab Phonol; 2013; 4,
Slowiaczek, LM; Soltano, EG; Bernstein, HL. Lexical and metrical stress in word recognition: Lexical or pre-lexical influences?. J Psycholinguist Res; 2006; 35,
Soto-Faraco, S; Sebastián-Gallés, N; Cutler, A. Segmental and suprasegmental mismatch in lexical access. J Mem Lang; 2001; 45, pp. 412-432. [DOI: https://dx.doi.org/10.1006/jmla.2000.2783]
Sulpizio, S; Colombo, L. Lexical stress, frequency, and stress neighbourhood effects in the early stages of Italian reading development. Q J Exp Psychol; 2013; 66,
Sulpizio, S; Colombo, L. Early markers of lexical stress in visual word recognition. Mem Cogn; 2017; 45,
Sulpizio, S; Job, R. The segment-to-frame association in word reading: early effects of the interaction between segmental and suprasegmental information. Front Psychol; 2015; 6, 1612. [DOI: https://dx.doi.org/10.3389/fpsyg.2015.01612] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/26539149][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4612140]
Sulpizio, S; Job, R; Burani, C. Priming lexical stress in reading Italian aloud. Lang Cogn Process; 2012; 27,
Taft M, Li J, Aryanti T (2021) Why do non-native English speakers show masked form priming when native speakers do not? J Second Lang Stud. https://doi.org/10.1075/jsls.21007.taf
Tang, R; Witzel, N. The role of phonology in processing morphologically complex words. Psychon Bull Rev; 2020; 27, pp. 379-384. [DOI: https://dx.doi.org/10.3758/s13423-019-01683-w] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31950457]
Tremblay, A. Is second language lexical access prosodically constrained? Processing of word stress by French Canadian second language learners of English. Appl Psycholinguist; 2008; 29,
Zhang, QG; Liang, LJ; Yao, PP; Hu, SS; Chen, BG. Parallel morpho-orthographic and morpho-semantic activation in processing second language morphologically complex words: evidence from Chinese–English bilinguals. Int J Biling; 2017; 21,
Copyright Palgrave Macmillan Dec 2025