Social Media and Language Processing: How Facebook and Twitter Provide the Best Frequency Estimates for Studying Word Recognition

Corpus‐based word frequencies are one of the most important predictors in language processing tasks. Frequencies based on conversational corpora (such as movie subtitles) are shown to better capture the variance in lexical decision tasks compared to traditional corpora. In this study, we show that frequencies computed from social media are currently the best frequency‐based estimators of lexical decision reaction times (up to 3.6% increase in explained variance). The results are robust (observed for Twitter‐ and Facebook‐based frequencies on American English and British English datasets) and are still substantial when we control for corpus size.     

Translation norms for English and Spanish: The role of lexical variables,word class,and L2 proficiency in negotiating translation ambiguity

We present a set of translation norms for 670 English and 760 Spanish nouns, verbs and class ambiguous items that varied in their lexical properties in both languages, collected from 80 bilingual participants. Half of the words in each language received more than a single translation across participants. Cue word frequency and imageability were both negatively correlated with number of translations. Word class predicted number of translations: Nouns had fewer translations than did verbs, which had fewer translations than class-ambiguous items. The translation probability of specific responses was positively correlated with target word frequency and imageability, and with its form overlap with the cue word. Translation choice was modulated by L2 proficiency: Less proficient bilinguals tended to produce lower probability translations than more proficient bilinguals, but only in forward translation, from Ll to L2. These findings highlight the importance of translation ambiguity as a factor influencing bilingual representation and performance. The norms can also provide an important resource to assist researchers in the selection of experimental materials for studies of bilingual and monolingual language performance. These norms may be downloaded from www.psychonomic.org/archive.     

Moving beyond Kučera and Francis: A critical evaluation of current word frequency norms and the introduction of a new and improved word frequency measure for American English

Word frequency is the most important variable in research on word processing and memory. Yet, the main criterion for selecting word frequency norms has been the availability of the measure, rather than its quality. As a result, much research is still based on the old Kučera and Francis frequency norms. By using the lexical decision times of recently published megastudies, we show how bad this measure is and what must be done to improve it. In particular, we investigated the size of the corpus, the language register on which the corpus is based, and the definition of the frequency measure. We observed that corpus size is of practical importance for small sizes (depending on the frequency of the word), but not for sizes above 16–30 million words. As for the language register, we found that frequencies based on television and film subtitles are better than frequencies based on written sources, certainly for the monosyllabic and bisyllabic words used in psycholinguistic research. Finally, we found that lemma frequencies are not superior to word form frequencies in English and that a measure of contextual diversity is better than a measure based on raw frequency of occurrence. Part of the superiority of the latter is due to the words that are frequently used as names. Assembling a new frequency norm on the basis of these considerations turned out to predict word processing times much better than did the existing norms (including Kučera & Francis and Celex). The new SUBTL frequency norms from the SUBTLEXUS corpus are freely available for research purposes from http://brm.psychonomic-journals.org/content/supplemental, as well as from the University of Ghent and Lexique Web sites.     

Contextual diversity, not word frequency, determines word-naming and lexical decision times

Word frequency is an important predictor of word-naming and lexical decision times. It is, however, confounded with contextual diversity, the number of contexts in which a word has been seen. In a study using a normative, corpus-based measure of contextual diversity, word-frequency effects were eliminated when effects of contextual diversity were taken into account (but not vice versa) across three naming and three lexical decision data sets; the same pattern of results was obtained regardless of which of three corpora was used to derive the frequency and contextual-diversity values. The results are incompatible with existing models of visual word recognition, which attribute frequency effects directly to frequency, and are particularly problematic for accounts in which frequency effects reflect learning. We argue that the results reflect the importance of likely need in memory processes, and that the continuity between reading and memory suggests using principles from memory research to inform theories of reading.     

Is the go/no-go lexical decision task an alternative to the yes/no lexical decision task?

In the go/no-go lexical decision task (LDT), participants are instructed to respond as quickly as they can when a word is presented and not to respond if a nonword is presented. By minimizing part of the response selection process in the experimental task, the impact of response decision time on the obtained lexical decision time is probably reduced relative to the standard yes/no LDT (Gordon, 1983). Experiments 1 and 2 show that the go/no-go LDT is sensitive to the effects of word frequency and associative priming--the magnitude of these effects is similar with the two tasks. More important, the go/no-go LDT has a number of advantages with respect to the "standard" yes/no LDT: It offers faster response times, more accurate responding, and fewer processing demands than does the yes/no task. Accordingly, the go/no-go task appears to be an excellent alternative to the standard yes/no task.     

Transfer effects from language processing to visual attention dynamics: The impact of orthographic transparency

The consistency between letters and sounds varies across languages. These differences have been proposed to be associated with different reading mechanisms (lexical vs. phonological), processing grain sizes (coarse vs. fine) and attentional windows (whole words vs. individual letters). This study aimed to extend this idea to writing to dictation. For that purpose, we evaluated whether the use of different types of processing has a differential impact on local windowing attention: phonological (local) processing in a transparent language (Spanish) and lexical (global) processing of an opaque language (English). Spanish and English monolinguals (Experiment 1) and Spanish–English bilinguals (Experiment 2) performed a writing to dictation task followed by a global–local task. The first key performance showed a critical dissociation between languages: the response times (RTs) from the Spanish writing to dictation task was modulated by word length, whereas the RTs from the English writing to dictation task was modulated by word frequency and age of acquisition, as evidence that language transparency biases processing towards phonological or lexical strategies. In addition, after a Spanish task, participants more efficiently processed local information, which resulted in both the benefit of global congruent information and the reduced cost of incongruent global information. Additionally, the results showed that bilinguals adapt their attentional processing depending on the orthographic transparency.     

Interactions of reading and semantics along the ventral visual processing stream

Converging evidence supports a distributed-plus-hub view of semantic processing, in which there are distributed modular semantic sub-systems (e.g., for shape, colour, and action) connected to an amodal semantic hub. Furthermore, object semantic processing of colour and shape, and lexical reading and identification, are processed mainly along the ventral stream, while action semantic processing occurs mainly along the dorsal stream. In Experiment 1, participants read a prime word that required imagining either the object or action referent, and then named a lexical word target. In Experiments 2 and 3, participants performed a lexical decision task (LDT) with the same targets as in Experiment 1, in the presence of foils that were legal nonwords (NW; Experiment 2) or pseudohomophones (PH; Experiment 3). Semantic priming was similar in effect size regardless of prime type for naming, but was greater for object primes than action primes for the LDT with PH foils, suggesting a shared-stream advantage when the task demands focus on orthographic lexical processing. These experiments extend the distributed-plus-hub model, and provide a novel paradigm for further research.     

Pretrial EEG coherence as a predictor of semantic priming effects

Twenty subjects were presented 240 randomized trials of a lexical decision task (LDT) while connected to a 16-electrode EEG montage. EEG data were acquired for 2 sec prior to the onset of each LDT trial. Alpha (9-13 Hz) coherence, alpha power, and 3- to 30-Hz power from five different combinations of EEG electrodes were used to predict semantic effects in the LDT trials. High pretrial coherence in the Right Hemisphere and Left Posterior-Parietal regions was associated with greater semantic facilitation (faster reaction times) in related-prime trials and greater semantic inhibition (slower reaction times) in unrelated-prime trials, relative to neutral-prime trials. No power variables were associated with semantic effects. Results support an information flow model of the functional significance of EEG coherence to cognitive processing.     

Checking and bootstrapping lexical norms by means of word similarity indexes

In psychology, lexical norms related to the semantic properties of words, such as concreteness and valence, are important research resources. Collecting such norms by asking judges to rate the words is very time consuming, which strongly limits the number of words that compose them. In the present article, we present a technique for estimating lexical norms based on the latent semantic analysis of a corpus. The analyses conducted emphasize the technique??s effectiveness for several semantic dimensions. In addition to the extension of norms, this technique can be used to check human ratings to identify words for which the rating is very different from the corpus-based estimate.     

Word-frequency effects in long-term semantic priming and false memory

Several studies have used the lexical decision task (LDT) with the Deese-Roediger-McDermott (DRM) false-memory paradigm to investigate whether long-term semantic priming (LTSP) occurs following presentation of lists of items (e.g., bed, dream, snore) for related non-presented lure words (e.g., sleep). However, results have been mixed, with some studies observing priming, whilst others have not. The present study had four goals: (i) to investigate the existence of LTSP in the LDT; (ii) to investigate effects of LTSP on standard effects of word frequency on LDT performance; (iii) to investigate the effect, if any, of word frequency on true and false recall; and (iv) to compare LDT performance with performance on a subsequent free-recall task. The findings showed (i) a significant effect of LTSP on LDT performance; (ii) no effect of LTSP on standard effects of word frequency on LDT performance; (iii) no effect of word frequency on either true or false free recall; and (iv) a significant relationship between LDT and free-recall performance. Implications of these findings for understanding LTSP and false memories are discussed.     

Semantic Ambiguity Effects in L2 Word Recognition

The present study examined the ambiguity effects in second language (L2) word recognition. Previous studies on first language (L1) lexical processing have observed that ambiguous words are recognized faster and more accurately than unambiguous words on lexical decision tasks. In this research, L1 and L2 speakers of English were asked whether a letter string on a computer screen was an English word or not. An ambiguity advantage was found for both groups and greater ambiguity effects were found for the non-native speaker group when compared to the native speaker group. The findings imply that the larger ambiguity advantage for L2 processing is due to their slower response time in producing adequate feedback activation from the semantic level to the orthographic level.     

Grasping the invisible: Semantic processing of abstract words

The problem of how abstract word meanings are represented has been a challenging one. In the present study, we extended the semantic richness approach (e.g., Yap, Tan, Pexman, & Hargreaves in Psychonomic Bulletin & Review 18:742–750, 2011) to abstract words, examining the effects of six semantic richness variables on lexical–semantic processing for 207 abstract nouns. The candidate richness dimensions were context availability (CA), sensory experience rating (SER), valence, arousal, semantic neighborhood (SN), and number of associates (NoA). The behavioral tasks were lexical decision (LDT) and semantic categorization (SCT). Our results showed that the semantic richness variables were significantly related to both LDT and SCT latencies, even after lexical and orthographic factors were controlled. The patterns of richness effects varied across tasks, with CA effects in the LDT, and SER and valence effects in the SCT. These results provide new insight into how abstract meanings may be grounded, and are consistent with a dynamic, multidimensional framework for semantic processing.     

Cognitive processing and motor execution in the lexical decision task: a developmental study

We investigated lexical decision making in children and adults by analyzing spatiotemporal characteristics of responses involving a hand movement. Children’s and adults’ movement trajectories were assessed in three tasks: a lexical decision task (LDT), a pointing task that involved minimal cognitive processing, and a symbol task requiring a simple binary decision. Cognitive interference on motor performance was quantified by analyzing movement characteristics in the LDT and symbol task relative to the pointing task. Across age groups, movements in the LDT were less smooth, slower, and more strongly curved to the opposite response option, and these interference effects decreased steadily with age. Older children showed stronger interference effects than did adults, even though their reaction times were similar to adults’ performance. No comparable effects were found in the symbol task, indicating that task characteristics such as response mapping and decision selection alone are not able to explain the developmental differences observed in the LDT. Our results indicate substantial overlap between cognitive processing and motor execution in the LDT in children that is not captured by computational models of visual word recognition and cognitive development.     

Disentangling Context Availability and Concreteness in Lexical Decision and W ord Translation

This study examines contrasting predictions of the dual coding theory and the context availability hypothesis regarding concreteness effects in monolingual and bilingual lexical processing. In three experiments, concreteness was controlled for or confounded with rated context availability. In the first experiment, bilingual subjects performed lexical decision in their native language (Dutch, L1). In the second experiment, lexical decision performance of bilinguals in their second language (English, L2) was examined. In the third experiment, bilinguals translated words 'forwards' (from L1 to L2) or 'backwards' (from L2 to L1). Both monolingual and bilingual tasks showed a concreteness effect when concreteness was confounded with context availability. However, concreteness effects disappeared when abstract and concrete words were matched on context availability, and even occasionally reversed. Implications of these results for theories that account for concreteness effects, particulary in bilingual processing, are discussed.     

Effects of frequency on visual word recognition tasks: where are they?

Compared the effect of frequency on lexical decision time (LDT) with that on reaction time (RT) in four other tasks, for the same words and subjects. Exp. 1 yielded an effect on semantic categorization RT (person vs. thing) similar in size and form to the effect on LDT. Exp. 2 yielded a substantial effect for syntactic categorization (noun vs. adjective), although weaker than the effect on LDT. In Exp. 3, the effect on naming RT for stress-final disyllabic words was identical to that on LDT, whereas the effect for stress-initial words was weaker. Exp. 4 showed no effect of frequency on delayed naming RT. The data undermine recent arguments for a (mainly) postidentification task-specific locus of frequency effects but are compatible with the older assumption (also characteristic of new PDP learning models) that lexical identification is a major locus of frequency effects (perhaps together with retrieval of meaning or phonology). But effects at that locus may be masked or diluted by other processes.     

Homophone effects in visual word recognition depend on homophone type and task demands.

This experiment examined how the characteristics of homophones and their mates influence homophone effects, as a function of task demands. Two types of homophones were presented: 1) low-frequency homophones with higher-frequency mates that are not animal names (e.g., maid--made), and 2) low-frequency homophones with mates that are, on average, of equivalent frequency and are animal names (e.g., foul--fowl). We observed a double dissociation: In the lexical decision task (LDT), there was a homophone effect for the first type of homophones but not for the second, whereas in the semantic categorization task (SCT) the opposite was true. These results suggest that in these tasks the effects of homophony arise when the homophone's mate creates competition in terms of the type of processing emphasized in the task, namely, orthographic processing in the LDT and semantic processing in the SCT.     

Unpredictability and complexity of print-to-speech correspondences increase reliance on lexical processes: more evidence for the orthographic depth hypothesis

The Orthographic Depth Hypothesis [Katz, L., & Frost, R. (1992). The reading process is different for different orthographies: The orthographic depth hypothesis. In R. Frost & L. Katz (Eds.), Orthography, phonology, morphology, and meaning (pp. 67–84). Amsterdam: Elsevier Science] proposes cross-linguistic differences in the involvement of lexical processing during reading. In orthographies with complex, inconsistent, and/or incomplete sublexical correspondences, decoding is more difficult and therefore slower. This gives more time to the lexical route to retrieve information, and leads to a greater ratio of lexical processing. We test whether this mechanism applies both for words with inconsistent (in English) and for words with complex (in French) correspondences. As complex correspondences are sufficient to derive a correct pronunciation, an increase in lexical processing may not occur. In a reading-aloud task, we used the frequency effect to measure lexical processing. The data showed stronger involvement of lexical processing for inconsistent compared to consistent words, and for complex compared to simple words. The results confirm that Katz and Frost’s proposed mechanism applies to different sources of orthographic depth.     

Responses on a lateralized lexical decision task relate to both reading times and comprehension

Research over the last few years has shown that the dominance of the left hemisphere in language processing is less complete than previously thought [Beeman, M. (1993). Semantic processing in the right hemisphere may contribute to drawing inferences from discourse. Brain and Language, 44, 80–120; Faust, M., & Chiarello, C. (1998). Sentence context and lexical ambiguity resolution by the two hemispheres. Neuropsychologia, 36(9), 827–835; Weems, S. A., & Zaidel, E. (2004). The relationship between reading ability and lateralized lexical decision. Brain and Cognition, 55(3), 507–515]. Engaging the right brain in language processing is required for processing speaker/writer intention, particularly in those subtle interpretive processes that help in deciphering humor, irony, and emotional inference. In two experiments employing a divided field or lateralized lexical decision task (LLDT), accuracy and reaction times (RTs) were related to reading times and comprehension on sentence reading. Differences seen in RTs and error rates by visual fields were found to relate to performance. Smaller differences in performance between fields tended to be related to better performance on the LLDT in both experiments and, in Experiment 1, to reading measures. Readers who can exploit both hemispheres for language processing equally appear to be at an advantage in lexical access and possibly also in reading performance.     

Repetition in visual word identification: Benefits and costs

University students performed lexical tasks with visually presented target words after the presentation of an identical or unrelated prime, at short (80–120?ms) or longer (410–710?ms) prime–target stimulus onset asynchronies (SOAs). Experiment 1 showed perceptual identification benefits in vocal responding at a short SOA that were reduced (accuracy) or reversed (latency) at a longer SOA. Experiment 2 showed a transition from a repetition benefit to a cost over 3 SOAs in a target-masked version of the lexical decision task (LDT; target displayed for only 141?ms). In Experiment 3 the repetition cost was replicated at a 530-ms SOA in the LDT with masked targets, but a repetition benefit was observed in the conventional LDT (target displayed until response). The dependence of repetition costs on target masking is more consistent with biases based on episodic confusions than refractoriness of lexical representations.