The Mental Lexicon - Volume 19, Issue 1, 2024

Neuropsychological assessments depend on language-based measures of cognitive functioning and the proper diagnosis of certain disorders relies on patterns of impairment shown on these measures (Lezak et al., 2004). The current project was motivated by the relative lack of literature integrating psycholinguistic experimental findings and clinical neuropsychological research on tests of verbal memory, specifically list learning. It has been well documented that word-level characteristics impact language processing and memory (see Yap & Balota, 2015 for a review). Therefore, it is critical that neuropsychologists begin to understand how current measures can be confounded by the underlying lexical and semantic characteristics of the stimuli and how, if used properly, those characteristics could aid in diagnostic specificity. The current study examined the structure of popular list learning tests and analyzed the influence of several psycholinguistic variables on the performance of healthy undergraduate participants. Results demonstrated that (1) age of acquisition, emotional valence, semantic neighborhood density, and imageability predicted recall accuracy of items from neuropsychological tests and (2) only one of the ten clinical test lists examined adequately controlled for these influential variables. Thus, clinicians could be missing clinically relevant data by ignoring psycholinguistic contributions to patient performance.

Several cross-linguistic studies have reported an impairment of verbal morphosyntax in People with Aphasia (PWA), highlighting a greater impact on tense morphology in contrast to agreement-inflectional morphology (Benedet, Christiansen & Goodglass, 1998; Friedmann & Grodzinsky, 1997, 2000; Gavarrò & Martínez-Ferreiro, 2007; Kok et al., 2007; Wenzlaff & Clahsen, 2004). The Diacritical Encoding and Retrieval Hypothesis (DERH, Faroqi-Shah & Thompson, 2007) postulates that this deficit is due to a specific difficulty projecting the semantic time information to the morphosyntax of verbs, related to processing limitations in PWA. This study aimed to assess the processing of the tense and agreement inflection in Spanish-speaking PWA. A group of 9 PWA completed four tasks designed to study verbal inflection production and comprehension in Spanish: Sentence Completion, Sentence Elicitation, Grammaticality Judgements and Sentence-Picture Matching. As expected, PWA showed a significatively greater impairment in tense inflection compared to agreement inflection. Furthermore, the analysis showed that this differential impairment manifested when PWA were asked to encode semantic time information and retrieve the corresponding verb morphology. These findings are consistent with the postulates of the DERH.

In various English lexical decision tasks (LDTs), bi-/multilinguals have evinced shorter response times (RTs) for cognates (i.e., words with the same meaning in two languages, e.g., the Dutch-English water) and longer RTs for interlingual homographs (IHs; words with distinct meanings in two languages, e.g., the Dutch-English map) compared to monolingual controls (e.g., Biloushchenko, 2017). This suggests that multilinguals automatically activate lexical representations from multiple languages (Dijkstra et al., 1998). To further investigate language (non-)selectivity, in our English LDTs, we compare the processing of cognates and IHs to intralingual words that are similar but only exist in English (i.e., cognates to metonyms like chicken, which can refer to the animal and the closely-related sense “chicken meat”, and IHs to homonyms like bat, which has two meanings: “baseball bat” and “nocturnal flying animal”). Half of our cognates and IHs only exist in our native Dutch participants’ non-native languages (English-French) to avoid any potentially confounding effects of the supposed “special status” (Midgley et al., 2011) of L1. Significant inhibition was found for homonyms and significant facilitation for metonyms and native (Dutch-English) cognates but not for non-native (English-French) cognates. These results are discussed in relation to the language non-selective hypothesis (Dijkstra et al., 1998).

Bilinguals’ decision-making may be affected by the linguistic context, depending on whether the scenario to be evaluated is presented in their first or second language. This phenomenon is known as the (moral) Foreign Language Effect (FLe/MFLE). To the best of our knowledge, the present study is the first to investigate whether a difference can also be observed between decisions taken by multilinguals in their second (L2) vs third languages (L3). Expectedly, akin to the L2 Status Factor hypothesis, the results of a survey experiment showed no significant distinctions in scenario evaluations across non-native languages, suggesting that the “foreign status” of these renders similar emotionality and/or exemption from social normativity. The few nuances that do emerge are traceable to the specific scenarios used in FLe research.

Semantic dimensions such as context availability, imageability and valence, form core components of many theoretical accounts of lexical processing. Typically, normative data for such semantic dimensions are drawn from subjective ratings, however, questions have been raised regarding the reliability and validity of these ratings. In this paper, we discuss this issue with a focus on context availability norms. Using data collected for another study, we show that context availability ratings required significantly higher rates of data exclusions at the level of both participants and items compared to other variables. In addition, high standard deviations at the item level, indicated a substantial degree of disagreement between participants. This suggested that some participants may have had difficulty understanding the concept of context availability which hindered their ability to complete the ratings. We provide recommendations for future research focusing on context availability and for norm collection procedures more broadly in order that the validity of such norms can be improved. In particular, we suggest that clear guidelines are required for data cleaning in order that the reliability of such norms is maximised and to facilitate replication across studies.

Pursuing cognitively stimulating activities, such as engaging with art, is crucial to a healthy lifestyle. The current work simulates visits to an art museum in a laboratory setting. Using eye tracking, we explored how linguistically guided visual search may increase attention, enjoyment and retention of information when viewing art. Two groups of adults, young (under 35 years) and older (over 65 years) viewed ten paintings on a computer screen presented either with or without an accompanying audio-guide, while having their eye movements recorded. Audio-guides referred to specific areas of the painting, marked as Interest Areas (IA). Across age groups, as attested by gaze fixations, the audio-guides increased attention to these areas compared to free-viewing. Audio-guided viewing did not lead to a significantly increase over free-viewing in information recall accuracy or feelings of enjoyment and engagement. Overall, older adults did report feeling more positively about both audio-guided and free viewing than young adults. Thus, the use of audio-guides, specifically the gamification through linguistically guided visual search, may be a useful tool to promote meaningful attentional interactions with art.

This study examines the applicability of the Cloze test, a widely used tool for assessing text comprehension proficiency, while highlighting its challenges in large-scale implementation. To address these limitations, an automated correction approach was proposed, utilizing Natural Language Processing (NLP) techniques, particularly word embeddings (WE) models, to assess semantic similarity between expected and provided answers. Using data from Cloze tests administered to students in Brazil, WE models for Brazilian Portuguese (PT-BR) were employed to measure the semantic similarity of the responses. The results were validated through an experimental setup involving twelve judges who classified the students’ answers. A comparative analysis between the WE models’ scores and the judges’ evaluations revealed that GloVe was the most effective model, demonstrating the highest correlation with the judges’ assessments. This study underscores the utility of WE models in evaluating semantic similarity and their potential to enhance large-scale Cloze test assessments. Furthermore, it contributes to educational assessment methodologies by offering a more efficient approach to evaluating reading proficiency.

Dual-route models of typing assume two pathways to retrieve a word’s spelling: a direct route connecting word to letter representations, and an indirect route via sound representations. The individual contribution of each route may depend on the modality of language acquisition: the first language (L1) is acquired sequentially in spoken and written modality respectively, whereas the second language (L2) is often acquired simultaneously in both modalities. We investigated whether sequential bilinguals rely more on the direct route during L2 vs. L1 typing. French-English bilinguals performed a typed picture-word interference task in their L1 and L2. We compared facilitation in naming for distractors that were phonologically (P) related, phonologically + orthographically (PO) related, or unrelated to the target. We predicted more facilitation by PO vs. P distractors in the L2 than in the L1. Participants showed significant facilitation by PO distractors, but not by P distractors, suggesting that orthographic overlap (together with phonology) helped retrieving the target spelling, whereas phonological overlap alone did not. The magnitude of this effect was similar across L1 and L2, contrary to our predictions. However, the absence of mere phonological facilitation suggests that phonology only contributes to typing when supported by orthography.

An emerging trend among young Cantonese speakers is to script-mix morphographic Chinese characters with Latin graphemes in social media exchanges, uncommon in traditional Chinese contexts. Results of a self-paced reading experiment with Cantonese speakers are reported to determine whether script-mixing incurs processing costs, and if so, whether these can be attributed to Inhibitory Control of one of the two scripts or to Dual Activation of both scripts but with slower lexical access within the non-dominant script. Sentences were presented either entirely in Chinese characters or had one region presented in Latin graphemes. Processing costs arose only at the switch from Latin graphemes back to Chinese characters, pointing to the involvement of Inhibitory Control. Further, these costs only appeared in a subset of grammatical categories, potentially coinciding with parsing uncertainties. As such, a combination of script-mixing and parsing complexities could be seen to result in processing costs in certain sentential positions.

Typing is a complex activity involving identifying and encoding the to-be-typed word, planning, and executing the keystrokes. Different mechanisms have been proposed to explain how contextual information about meaning influences the processing of a target word (for example, semantic priming), and it is unclear how this information influences the typing output of the target word. When the interval between the prime and target is short, the prime is thought to automatically activate the target. With a longer interval, the facilitation may be more strategic. The influence of the prime on the output of the target may thus depend on the interval between the two. We found that at both short and long intervals, related semantic primes facilitated the speed of the first keystroke of the target word. However, there was no effect on how quickly the non-initial letters of the target were entered, suggesting that information from semantic primes influences the planning and initiation of typing, but not the execution of remaining keystrokes. Interestingly, the initial and non-initial keystrokes were faster when the interstimulus interval was long, suggesting participants encode the letters of the prime which could interfere with encoding and typing the target if the interval between them is short.

The Bouba/Kiki (BK) effect is observed when a linguistic sound is associated with a shape. People usually associate the nonword bouba with a round shape, and kiki with a sharp shape (Ramachandran & Hubbard, 2001). In 2011, Nielsen and Rendall found that certain English letters (/k/, /p/, and /t/) and (/b/, /l/, /m/, and /n/) were associated with sharp and round shapes respectively. The BK effect was investigated in depth for the first time in Arabic in 2022 (Nassereddine) using Arabic Analogs to the English letters. Arabic participants’ performance was not consistent with previous research (Nielsen & Rendall, 2011). The goal of the present study was to determine the roundest and sharpest Arabic letters by presenting all letters both visually and auditorily to Arabic speakers and have them say whether a letter shape or sound best maps on to the standard bouba and kiki shapes. The results revealed that Arabic does have both round and sharp letters, and that there is a strong influence of phonological features on this BK effect.

This study investigates word processing in Brazilian Portuguese, focusing on blends, which juxtapose or overlap (W)ords and/or (C)lips (e.g. portunhol = (portu)guês ‘Portuguese’ + espa(nhol) ‘Spanish’). Blends present intriguing theoretical and empirical challenges to models of morphological analysis, morphological processing, lexical access, and the mental lexicon. Most research on blends has been conducted in languages other than Portuguese. This study addresses this gap by exploring the processing of blends in Brazilian Portuguese through a behavioral lexical decision experiment. We manipulated blends in constituent structure and grammatical structure, considering (H)ead and (M)odification. Additionally, we compared blends against words with various morphological structures, such as derived complex words containing prefixes (e.g., [des]acordo ‘disagreement’) or suffixes (e.g., cozinh[eiro] ‘cook’), and monomorphemic simplex words. We also included simplex and complex pseudowords (e.g., [acont]arago; dador[eiro]) and nonwords (e.g., sfaricrelj) in the experiment. Accuracy and reaction time results suggest that blends are accepted and processed differently from simplex and complex words, resembling pseudowords. This study contributes to a deeper understanding of blend description and processing, providing valuable insights into lexical access, enhancing theoretical and empirical comprehension of morphological processing.

Organizational policies for passwords and passphrases require certain criteria, such as minimum length or uppercase letters, to be met, often resulting in a tradeoff between complexity and ease of typing. Uppercase letters, specifically, lead to slower and more error prone entries. Our present study examined their influence on the typing of three-word passphrases. We were interested in whether uppercase letter location, which should not influence passphrase security, would influence its typing. Passphrases with no uppercase letter were typed more accurately and quickly than passphrases with an uppercase letter. Importantly, passphrases with an uppercase letter in the second word were more likely to be typed incorrectly, and were typed more slowly when entered correctly. Our findings are consistent with the linguistic information of adjacent words influencing the output of the word being typed, where an altered second word interfered with the output of both the first and second words.

Temperature adjectives serve as linguistic vehicles to convey sensory perceptions of physical temperature and metaphorical extensions that resonate across various spheres of communication. Languages vary widely in their vocabulary related to temperature, how they classify these terms, and where they fit within grammatical structures. Despite this diversity, there are shared features in how temperature adjectives convey both temperature and metaphorical meanings across languages. However, European Portuguese (EP) research on temperature adjectives remains scarce. To better understand temperature adjectives in EP, we conducted a corpus-based analysis of eight adjectives -“gélido”, “gelado”, “frio”, “fresco”, “morno”, “tépido”, “quente”, “escaldante” (icy, frozen, cold, cool, lukewarm, tepid, hot, scorching) — to investigate their literal and metaphorical meanings using a corpus of 2920 fragments from the Reference Corpus of Contemporary Portuguese and a quantitative and qualitative approach for the analysis. Results show that all adjectives have metaphorical meanings in addition to basic temperature interpretations, but their distribution varies. They primarily appear post-nominally and in attributive positions, responding differently to degree quantifiers. While the nouns with which the adjectives combine are relevant, the alternation between temperature and metaphorical readings also depends on context.

This study investigates the usage of 笑死 (xiaosi hereafter, ‘laugh-die; die laughing’) in Facebook posts retrieved during Taiwan’s level-3 COVID-19 alert period (5/15/2021–7/26/2021). The analysis of xiaosi revealed that it serves two major functions: (a) as its literal meaning, ‘to laugh to death’ (die laughing), and (b) as a discourse marker. The use of this discourse-marker function demonstrates that it has acquired a new pragmatic function in social media contexts beyond its literal meaning. Specifically, xiaosi can express a speaker’s stance or provide an evaluative comment on the situation, thus introducing a new conversational function distinct from its original meaning.