Malay Lexicon Project 3: The impact of orthographic-semantic consistency on lexical decision latencies.

Theories of word processing propose that readers are sensitive to statistical co-occurrences between spelling and meaning. Orthographic-semantic consistency (OSC) measures provide a continuous estimate of the statistical regularities between spelling and meaning. Here we examined Malay, an Austronesian language that is agglutinative. In Malay, stems are often repeated in other words that share a related meaning (e.g., sunyi/quiet; ke-sunyi-an/silence; makan/eat; makan-an/foods). The first goal was to expand an existing large Malay database by computing OSC estimates for 2,287 monomorphemic words. The second goal was to explore the impact of root family size and OSC on lexical decision latencies for monomorphemic words. Decision latencies were collected for 1,280 Malay words of various morphological structures. Of these, data from 1,000 monomorphemic words were analysed in a series of generalised additive mixed models (GAMMs). Root family size and OSC were significant predictors of decision latencies, particularly for lower frequency words. We found a facilitative effect of root family size and OSC. Furthermore, we observed an interaction between root family size and OSC in that an effect of OSC was only apparent in words with larger root families. This interaction has not yet been explored in English but has the potential to be a new benchmark effect to test distributional models of word processing.

Cross-language semantic parafoveal preview benefits in bilinguals.

The aim of this project was to identify factors contributing to cross-language semantic preview benefits. In Experiment 1, Russian-English bilinguals read English sentences with Russian words presented as parafoveal previews. The gaze-contingent boundary paradigm was used to present sentences. Critical previews were cognate translations of the target word (CTAPT-START), noncognate translations (CPOK-TERM), or interlingual homograph translations (MOPE-SEA). A semantic preview benefit (i.e., shorter fixation durations for related than unrelated previews) was observed for cognate and interlingual homograph translations, but not for noncognate translations. In Experiment 2, English-French bilinguals read English sentences with French words used as parafoveal previews. Critical previews were interlingual homograph translations of the target word (PAIN-BREAD) or interlingual homograph translations with a diacritic added (PÁIN-BREAD). A robust semantic preview benefit was found only for interlingual homographs without diacritics, although both preview types produced a semantic preview benefit in the total fixation duration. Our findings suggest that semantically related previews need to have substantial orthographic overlap with words in the target language to produce cross-language semantic preview benefits in early eye fixation measures. In terms of the Bilingual Interactive Activation+ model, the preview word may need to activate the language node for the target language before its meaning is integrated with that of the target word. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Malay Lexicon Project 2: Morphology in Malay word recognition.

Morphological processing in visual word recognition has been extensively studied in a few languages, but other languages with interesting morphological systems have received little attention. Here, we examined Malay, an Austronesian language that is agglutinative. Agglutinative languages typically have a large number of morphemes per word. Our primary aim was to facilitate research on morphological processing in Malay by augmenting the Malay Lexicon Project (a database containing lexical information for almost 10,000 words) to include a breakdown of the words into morphemes as well as morphological properties for those morphemes. A secondary goal was to determine which morphological variables influence Malay word recognition. We collected lexical decision data for Malay words that had one prefix and one suffix, and first examined the predictive power of 15 morphological and four lexical variables on response times (RT). Of these variables, two lexical and three morphological variables emerged as strong predictors of RT. In GAMM models, we found a facilitatory effect of root family size, and inhibitory effects of prefix length and prefix percentage of more frequent words (PFMF) on RT. Next, we explored the interactions between overall word frequency and several of these predictors. Of particular interest, there was a significant word frequency by root family size interaction in which the effect of root family size is stronger for low-frequency words. We hope that this initial work on morphological processing in Malay inspires further research in this and other understudied languages, with the goal of developing a universal theory of morphological processing.

The effect of speaker age on the perception of ironic insults.

We investigated a cue that readers may use in determining whether a remark such as "You are so helpful!" is intended as a compliment or as an ironic insult. The cue was the age of the speaker. Remarks were preceded by a sentence that either invited a literal or ironic interpretation of the remark. Data were collected on the familiarity of the remark as an ironic statement, and the incongruity of the remark with the prior context. In Experiment 1, participants were asked to rate the intent of the speaker as to how ironic, mocking, polite, and funny they intended their remark to be. In Experiment 2, participants read the scenarios as their eye movements were tracked. The results showed that age of the speaker had an impact on first pass reading times when statements were not familiar as ironic statements. Our younger adult participants did not appear to immediately activate a nonliteral interpretation of an ambiguous remark made by an older adult unless they had evidence from past experience that the remark is often used as an insult. However, ratings of the ironic intent of the statements were unaffected by speaker age; the age of the speaker affects the ease of interpretation but not the final outcome. The results are consistent with constraint-based theories of sentence comprehension. (PsycInfo Database Record (c) 2021 APA, all rights reserved).

Is letter position coding when reading in L2 affected by the nature of position coding used when bilinguals read in their L1?

Using four-character Chinese word targets, Yang, Chen, Spinelli, and Lupker (Journal of Experimental Psychology: Learning, Memory, and Cognition, 45(8), 1511-1526, 2019) and Yang, Hino et al. (Journal of Memory and Language, 113, 104017, 2020) demonstrated that backward primes (Roman alphabet example-dcba priming ABCD) produce large masked priming effects. This result suggests that character position information is quite imprecisely coded by Chinese readers when reading in their native language. The present question was, If Chinese readers have evolved a reading system not requiring precise position information, would Chinese-English bilinguals show more extreme transposed letter priming effects when processing English words than both English monolinguals and other types of bilinguals whose L2 is English? In Experiment 1, Chinese-English bilinguals, but not English monolinguals, showed a clear backward priming effect in a lexical decision task. In Experiment 2, the parallel backward priming effect was absent for both Spanish-English and Arabic-English bilinguals. Apparently, the orthographic coding system that Chinese-English bilinguals use when reading in their L2 leans heavily on the flexible/imprecise position coding process that they develop for reading in their L1.

Early activation of cross-language meaning from phonology during sentence processing.

The current study investigated whether shared phonology across languages activates cross-language meaning when reading in context. Eighty-five bilinguals read English sentences while their eye movements were tracked. Critical sentences contained English members of English-French interlingual homophone pairs (e.g., mow; French homophone mate mot means "word") or they contained spelling control words (e.g., mop). Only the meaning of the unseen French homophone mate fit the context (e.g., Hannah wrote another mow/mop on the blackboard for the spelling test). Differences in fixation durations between homophone errors and spelling control errors provided evidence for cross-language activation that extended to semantic representations. When the unseen French homophone was of high frequency, shorter first fixations and gaze durations were observed on English interlingual homophones than on English control words, providing evidence that the French meaning associated with the shared phonology was activated during early stage word identification. Individual differences analyses showed that these effects were larger when bilinguals were using the nontarget language (i.e., French) more regularly in daily life. Results provide evidence that cross-language activation of phonology can be sufficiently strong to activate corresponding semantic representations during single language sentence processing. (PsycInfo Database Record (c) 2020 APA, all rights reserved).

Figuring Out How Verb-Particle Constructions Are Understood During L1 and L2 Reading.

The aim of this paper was to investigate first-language (L1) and second-language (L2) reading of verb particle constructions (VPCs) among English-French bilingual adults. VPCs, or phrasal verbs, are highly common collocations of a verb paired with a particle, such as eat up or chew out, that often convey a figurative meaning. VPCs vary in form (eat up the candy vs. eat the candy up) and in other factors, such as the semantic contribution of the constituent words to the overall meaning (semantic transparency) and frequency. Much like classic forms of idioms, VPCs are difficult for L2 users. Here, we present two experiments that use eye-tracking to discover factors that influence the ease with which VPCs are processed by bilingual readers. In Experiment 1, we compared L1 reading of adjacent vs. split VPCs, and then explored whether the general pattern was driven by item-level factors. L1 readers did not generally find adjacent VPCs (eat up the candy) easier to process than split VPCs (eat the candy up); however, VPCs low in co-occurrence strength (i.e., low semantic transparency) and high in frequency were easiest to process in the adjacent form during first pass reading. In Experiment 2, we compared L2 reading of adjacent vs. split VPCs, and then explored whether the general pattern varied with item-level or participant-level factors. L2 readers generally allotted more second pass reading time to split vs. adjacent forms, and there was some evidence that this pattern was greater for L2 English readers who had less English experience. In contrast with L1 reading, there was no influence of item differences on L2 reading behavior. These data suggest that L1 readers may have lexicalized VPC representations that are directly retrieved during comprehension, whereas L2 readers are more likely to compositionally process VPCs given their more general preference for adjacent particles, as demonstrated by longer second pass reading time for all split items.

Effects of lexical ambiguity on perception: A test of the label feedback hypothesis using a visual oddball paradigm.

We used a visual oddball paradigm to investigate whether a shared verbal label makes two objects belonging to different conceptual categories less perceptually distinct. In Experiment 1, the critical images shared a label as well as some perceptual features (orange, referring to the color and the fruit), and in Experiment 2, the critical images shared a label but no perceptual features (bat, referring to the animal and the sports equipment). In both experiments comparison images were similar to each of the critical images but they did not share a label. A reduced deviant-related negativity (DRN) was observed for critical images compared with comparison images in both experiments, suggesting that the critical image pairs were perceived as less distinct than comparison pairs. These results extend previous research using the visual oddball paradigm that has shown that images from the same conceptual category are perceived as more distinct when they have different labels, and provide further support for the label-feedback hypothesis (Lupyan, 2012) in which language is assumed to modulate perception online. (PsycINFO Database Record (c) 2018 APA, all rights reserved).

Reading homophone puns: Evidence from eye tracking.

We investigated how readers make sense of homophone puns (e.g., The butcher was very glad we could meat up) by tracking their eye movements as they read. Comparison sentences included homophone-error sentences in which the presented homophone was also not correct (e.g., The lawyer was very glad we could meat up) and sentences in which the homophone was correct for the context (e.g., The butcher was very glad to chop meat up for the stew). An effect of the frequency of the unpresented homophone mate (e.g., meet) was found on first-pass reading times for homophones, indicating that participants activated the meaning of the homophone mate through shared phonology. First-fixation and gaze durations on the homophones were longer in puns than in correct-context sentences, indicating that participants immediately noticed that the homophone was incongruous with the adjacent context (e.g., glad we could meat) in puns, but total reading times did not differ, suggesting that the incongruity was quickly resolved. Immediate reading times on homophone in puns and homophone-error sentences did not differ, but total reading times did, suggesting that the impact of the critical context word (e.g., butcher) is delayed. Further analyses examined the resolution process in more detail. Ratings of the funniness of the puns were most strongly related to the strength of the association between the homophone and the critical context word (e.g., butcher). (PsycINFO Database Record

The effect of semantic transparency on the processing of morphologically derived words: Evidence from decision latencies and event-related potentials.

Decomposition theories of morphological processing in visual word recognition posit an early morpho-orthographic parser that is blind to semantic information, whereas parallel distributed processing (PDP) theories assume that the transparency of orthographic-semantic relationships influences processing from the beginning. To test these alternatives, the performance of participants on transparent (foolish), quasi-transparent (bookish), opaque (vanish), and orthographic control words (bucket) was examined in a series of 5 experiments. In Experiments 1-3 variants of a masked priming lexical-decision task were used; Experiment 4 used a masked priming semantic decision task, and Experiment 5 used a single-word (nonpriming) semantic decision task with a color-boundary manipulation. In addition to the behavioral data, event-related potential (ERP) data were collected in Experiments 1, 2, 4, and 5. Across all experiments, we observed a graded effect of semantic transparency in behavioral and ERP data, with the largest effect for semantically transparent words, the next largest for quasi-transparent words, and the smallest for opaque words. The results are discussed in terms of decomposition versus PDP approaches to morphological processing. (PsycINFO Database Record

Skilled adult readers activate the meanings of high-frequency words using phonology: Evidence from eye tracking.

We examined whether highly skilled adult readers activate the meanings of high-frequency words using phonology when reading sentences for meaning. A homophone-error paradigm was used. Sentences were written to fit 1 member of a homophone pair, and then 2 other versions were created in which the homophone was replaced by its mate or a spelling-control word. The error words were all high-frequency words, and the correct homophones were either higher-frequency words or low-frequency words-that is, the homophone errors were either the subordinate or dominant member of the pair. Participants read sentences as their eye movements were tracked. When the high-frequency homophone error words were the subordinate member of the homophone pair, participants had shorter immediate eye-fixation latencies on these words than on matched spelling-control words. In contrast, when the high-frequency homophone error words were the dominant member of the homophone pair, a difference between these words and spelling controls was delayed. These findings provide clear evidence that the meanings of high-frequency words are activated by phonological representations when skilled readers read sentences for meaning. Explanations of the differing patterns of results depending on homophone dominance are discussed.

Cross-script orthographic and phonological preview benefits.

The present experiment examined the use of parafoveally presented first-language (L1) orthographic and phonological codes during reading of second-language (L2) sentences in proficient Russian-English bilinguals. Participants read English sentences containing a Russian preview word that was replaced by the English target word when the participant's eyes crossed an invisible boundary located before the preview word. The use of English and Russian allowed us to manipulate orthographic and phonological preview effects independently of one another. The Russian preview words overlapped with English target words in (a) orthography (ВЕЛЮР [vʲɪ'lʲʉr]-BERRY), (b) phonology (БЛАНК [blank]-BLOOD), or (c) had no orthographic or phonological overlap (КАЛАЧ [kɐ'lat͡ɕ]-BERRY; ГЖЕЛЬ [ɡʐϵlʲ]-BLOOD). The results of this study showed a clear and strong benefit of the parafoveal preview of Russian words that shared either orthography or phonology with English target words. This study is the first demonstration of cross-script orthographic and phonological parafoveal preview benefit effects. Bilinguals integrate orthographic and phonological information across eye fixations in reading, even when this information comes from different languages.

Phonological activation of word meanings in grade 5 readers.

Three experiments examined the role of phonology in the activation of word meanings in Grade 5 students. In Experiment 1, homophone and spelling control errors were embedded in a story context and participants performed a proofreading task as they read for meaning. For both good and poor readers, more homophone errors went undetected than spelling control errors. In Experiments 2 and 3, homophone and spelling control errors were in sentence contexts. Experiment 2 used an online sentence verification task, and found that both good and poor readers were less accurate when sentences contained a homophone error than a spelling control error. Furthermore, a difference between the 2 types of sentences was observed even when participants were concurrently performing an articulation task. In Experiment 3, initial reading times were shorter on homophone errors than on spelling controls, and participants were less likely to make a regression from homophone errors than spelling controls. These experiments provide clear evidence that phonology makes an important contribution to the activation of word meanings in Grade 5 readers.

The role of lexical expertise in reading homophones.

In Experiment 1, university students classified on lexical expertise on the basis of spelling plus nonword pronunciation accuracy made lexical decisions to homophones and control words. Homophones were accepted as words more slowly than control words, but lexical experts showed a smaller homophone cost than the less skilled group. In Experiment 2, similarly classified groups showed a large difference in their ability to detect homophones, with the low-expertise group showing a yes bias to high-frequency words, and having difficulty detecting homophones when mate-frequency was low. The results suggest superior use of orthography in the lexical experts and more reliance on semantic information in nonexperts, and support the importance of facility with orthography-phonology mappings in lexical expertise.

Phonological processing dynamics in bilingual word naming.

The current study investigated phonological processing dynamics in bilingual word naming. English-French and French-English bilinguals named interlingual heterophonic homographs (i.e., words that share orthography but not meaning or pronunciation across languages), heterophonic cognates (i.e., words that share both orthography and meaning across languages, but not pronunciations), interlingual homophones (i.e., words that share pronunciation, but not orthography or meaning across languages), and single-language matched control words in both English and French naming tasks. Cross-language phonological activation was strongest in bilinguals' second language. The results provided evidence for feedforward activation of phonological representations in the nontarget language, as well as feedback activation of these phonological representations from semantic representations. Results are interpreted within the more recent Bilingual Interactive Activation (BIA+) framework.

Cross-language phonological activation: evidence from masked onset priming and ERPs.

The goal of the present research was to provide direct evidence for the cross-language interaction of phonologies at the sub-lexical level by using the masked onset priming paradigm. More specifically, we investigated whether there is a cross-language masked onset priming effect (MOPE) with L2 (English) primes and L1 (Russian) targets and whether it is modulated by the orthographic similarity of primes and targets. Primes and targets had onsets that overlapped either only phonologically, only orthographically, both phonologically and orthographically, or did not have any overlap. Phonological overlap, but not orthographic overlap, between primes and targets led to faster naming latencies. In contrast, the ERP data provided evidence for effects of both phonological and orthographic overlap. Finally, the time-course of phonological and orthographic processing for our bilinguals mirrored the time-course previously reported for monolinguals in the ERP data. These results provide evidence for shared representations at the sub-lexical level for a bilingual's two languages.

Discrimination of English and French orthographic patterns by biliterate children.

We investigated whether young English-French biliterate children can distinguish between English and French orthographic patterns. Children in French immersion programs were asked to play a dictionary game when they were in Grade 2 and again when they were in Grade 3. They were shown pseudowords that contained either an English spelling pattern or a French spelling pattern, and they were asked to decide whether each pseudoword should go in an English dictionary or a French dictionary if it became a real word. Comparison groups of monolingual English children, monolingual French children, and English-French bilingual university students were also tested on the task. French immersion students in both grades were above chance in discriminating between the two types of pseudowords but were well below adult performance on the task. Measures obtained in kindergarten showed that early print knowledge had some ability to predict later ability to discriminate between the orthographic patterns of the two languages. Further analyses indicated that exposure to print in each language in Grades 1 to 3 was strongly related to discrimination performance. The findings are interpreted as being consistent with the statistical learning hypothesis.

The processing advantage and disadvantage for homophones in lexical decision tasks.

Studies using the lexical decision task with English stimuli have demonstrated that homophones are responded to more slowly than nonhomophonic controls. In contrast, several studies using Chinese stimuli have shown that homophones are responded to more rapidly than nonhomophonic controls. In an attempt to better understand the impact of homophony, we investigated homophone effects for Japanese kanji words in a lexical decision task. The results indicated that, whereas a processing disadvantage emerged for homophones when they have only a single homophonic mate (as in the English experiments), a processing advantage occurred for homophones when they have multiple homophonic mates (as in the Chinese experiments). On the basis of these results, we discuss the nature of the processes that may be responsible for producing the processing advantages and disadvantages for homophones.

Cross-language message- and word-level transfer effects in bilingual text processing.

The present study examined the nature of the mental representations bilinguals form when reading a text and to what extent they are language specific. English-French bilinguals read five pairs of passages in succession while their eye movements were tracked. Dependent measures were overall reading times on second passages and fixation latencies on target cognates embeddedin second passages. The first passage w as (1) identical tothe second passage in the pair, (2) related in content only (i.e., a translation), (3) related in content and some words (i.e., translation with cognates), (4) related in words only (i.e., different content with the same cognates), or (5) unrelated. There was substantial cross-language facilitation for passages that shared meaning, but the amount of transfer was less than that for identical passages, indicating that memory representations are largely meaning based but do contain some information about surface form. Cross-language transfer for cognates was observed but depended on the skill of the bilinguals in their second language, the direction of transfer, and whether the passages shared meaning. These results are discussed in relation to Raney's (2003) model of text representation.