Exception Words Research Articles

Sla-former: conformer using shifted linear attention for audio-visual speech recognition

Conformer-based models have proven highly effective in Audio-visual Speech Recognition, integrating auditory and visual inputs to significantly enhance speech recognition accuracy. However, the widely utilized softmax attention mechanism within conformer models encounters scalability issues, with its spatial and temporal complexity escalating quadratically with sequence length. To address these challenges, this paper introduces the Shifted Linear Attention Conformer, an evolved iteration of the conformer architecture. Shifted Linear Attention Conformer adopts shifted linear attention as a scalable alternative to softmax attention. We conducted a thorough analysis of the factors constraining the efficiency of linear attention. To mitigate these issues, we propose the utilization of a straightforward yet potent mapping function and an efficient rank restoration module, enhancing the effectiveness of self-attention while maintaining low computational complexity. Furthermore, we integrate an advanced attention-shifting technique facilitating token manipulation within attentional mechanisms, thereby enhancing information flow across various groups. This three-part approach enhances cognitive computations, particularly beneficial for processing longer sequences. Our model achieves exceptional Word Error Rates of 1.9% and 1.5% on the Lip Reading Sentences 2 and Lip Reading Sentences 3 datasets, respectively, showcasing its state-of-the-art performance in audio-visual speech recognition tasks.

The influence of oral vocabulary knowledge on individual differences in a computational model of reading

Studies have demonstrated systematic individual differences in the degree of semantic reliance (SR) when reading aloud exception words in adult skilled readers. However, the origins of individual differences in reading remain unclear. Using a connectionist model of reading, this study investigated whether oral vocabulary knowledge may affect the degree of SR as a potential source of individual differences in reading. Variety in oral vocabulary knowledge was simulated by training the model to learn the mappings between spoken and meaning word forms with different vocabulary sizes and quantities of exposure to these vocabularies. The model’s SR in the reading aloud task was computed. The result demonstrated that the model with varying amounts of oral exposure and vocabulary sizes had different levels of SR. Critically, SR was able to predict the performance of the model on reading aloud and nonword reading, which assimilated behavioural reading patterns. Further analysis revealed that SR was largely associated with the interaction between oral vocabulary exposure and oral vocabulary size. When the amount of exposure was limited, SR significantly increased with vocabulary sizes but decreased then with vocabulary sizes. Overall, the simulation results provide the first computational evidence of the direct link between oral vocabulary knowledge and the degree of SR as a source of individual differences in reading.

No exceptional words for Bernoulli percolation

Benjamini and Kesten introduced in 1995 the problem of embedding infinite binary sequences into a Bernoulli percolation configuration, known as percolation of words . We give a positive answer to their Open Problem 2: almost surely, all words are seen for site percolation on \mathbb{Z}^3 with parameter p = 1/2 . We also extend this result in various directions, proving the same result on \mathbb{Z}^d , d \geq 3 , for any value p \in (p_c^{\textup{site}}(\mathbb{Z}^d), 1 - p_c^{\textup{site}}(\mathbb{Z}^d)) , and for restrictions to slabs. Finally, we provide an explicit estimate on the probability to find all words starting from a finite box.

Unique, Shared, and Dominant Brain Activation in Visual Word Form Area and Lateral Occipital Complex during Reading and Picture Naming

Identifying printed words and pictures concurrently is ubiquitous in daily tasks, and so it is important to consider the extent to which reading words and naming pictures may share a cognitive-neurophysiological functional architecture. Two functional magnetic resonance imaging (fMRI) experiments examined whether reading along the left ventral occipitotemporal region (vOT; often referred to as a visual word form area, VWFA) has activation that is overlapping with referent pictures (i.e., both conditions significant and shared, or with one significantly more dominant) or unique (i.e., one condition significant, the other not), and whether picture naming along the right lateral occipital complex (LOC) has overlapping or unique activation relative to referent words. Experiment 1 used familiar regular and exception words (to force lexical reading) and their corresponding pictures in separate naming blocks, and showed dominant activation for pictures in the LOC, and shared activation in the VWFA for exception words and their corresponding pictures (regular words did not elicit significant VWFA activation). Experiment 2 controlled for visual complexity by superimposing the words and pictures and instructing participants to either name the word or the picture, and showed primarily shared activation in the VWFA and LOC regions for both word reading and picture naming, with some dominant activation for pictures in the LOC. Overall, these results highlight the importance of including exception words to force lexical reading when comparing to picture naming, and the significant shared activation in VWFA and LOC serves to challenge specialized models of reading or picture naming.

انفتاح اللغة العربية المعاصرة

This paper deals with the study and analysis of a phenomenon that has emerged - to a certain extent – in the contemporary Arabic Language, that is: It became gathering new types of the Feminine sound Plural that were not touched by the rules of the elders so that the abstract’s title was adopted. The study began with the debate that took place among the elders in favoring the term: "feminine sound plural" to the term: "plural ending with ((Alphabetical Letters: Alif (A) and Ta’a (T))), after that they made comparison between the strict conditions of the Sound muscular plural and the Non-strict conditions of the Sound Feminine Plural. Then, the Paper presented the varieties being recorded and grouped by Elders as Sound Feminine Plural together with exceptional words thereof at a limited extent. The Paper then went on to reveal the new species that the Contemporary Arabic has opened up to and have been applied by the Sound Feminine Plural group in aremarkable abundance. The Paper categorized the new types which the contemporary Arabic has opened up to, into three Classes comprised:Class (1) the masculine Words, either names of solids or Derivatives shifted to the nominal field.Class (2) the infinitives including the (3-lettered Verb) at a limited extent, and the infinitives of the (4-lettered Verbs) to a wider extent thereof, but the remarkable openness being onto the infinitives of 5-& 6- lettered verbs.Class (3) the Foreign Words which includes the terminologies word and the other exotic words implanted in the Arabic Language whereas they have been interested to apply the form of the Sound Feminine Plural for its simplicity as it require to only add (Alphabetical Letters: Alif (A) and Taa (T)) to the end of the nominal words.

관형사의 어순

The purpose of this study is to define the relative word orders of Korean adnominals. In order to accomplish this aim, I subclassify adnominals into nominal-adnominals and predicative-adnominals at first, then into the traditional subclassifications as demonstrative, numeric and attributive adnominals. I argue that my subclassification gives more accurate explanations for some exceptional word order problems among many adnominals which are not solved till now. That is, while nominal-adnominals have relatively fixed syntactic positions as demonstratives - numerics - attributives respectably in their order, predicative-adnominals are not the same as that. Some predicative-adnominals observe the relative word order, but others do not. They are free in their syntactic positions, and these kinds of adnominals also show peculiar syntactic behaves which are not mentioned till now. That is, they can be conjugated with adjectives, be replaced by proform, and be modified by adverbs. These extraordinary phenomena are explained also.

Semantic deficits in ALS related to right lingual/fusiform gyrus network involvement.

BackgroundThe clinicopathological continuity between amyotrophic lateral sclerosis (ALS) and frontotemporal lobar degeneration (FTLD) is well known. Although ALS demonstrates language symptoms similar to FTLD, including semantic dementia, word reading impairments in ALS have not been well studied. “Jukujikun” are Kanji-written words with irregular pronunciation comparable to “exception words” and useful for detecting semantic deficits in Japan. We conducted a cross-sectional study to investigate Jukujikun reading impairments and related network changes in ALS.MethodsWe enrolled 71 ALS patients and 69 healthy controls (HCs). Age-, sex-, and education matched HCs were recruited from another cohort study concurrently with patient registration. We examined neuropsychological factors including low frequency Jukujikun reading. We performed resting-state functional magnetic resonance imaging with voxel-based graph analysis on a subset of participants who agreed.FindingsLow frequency Jukujikun score was decreased in ALS (15·0[11·0–19·0](median[25–75 percentile])) compared with HCs (19·0[17·3–20·0]) (p < 0·001, effect size = 0·43). Fifty-two percent of ALS (N = 37) with low frequency Jukujikun score ≤ 5th percentile of HCs was classified as ALS with positive Jukujikun deficit (ALS-JD+). Compared with HCs, ALS-JD+ showed decreased degree centrality in the right lingual/fusiform gyrus, where connectivities with regions associated with word perception, semantic processing, or speech production were decreased. They also showed increased degree centrality in the left inferior/middle temporal gyrus, associated with increased connectivities involving semantic processing.InterpretationDysfunction of the “hub” in the right lingual/fusiform gyrus can affect semantic deficit in ALS. Considering neuropsychological symptoms as network impairments is vital for understanding various diseases.FundMHLW and MEXT, Japan.

Design and development of lemmatizer for Sindhi language in devanagri script

The Study of word formation is called as morphology. Morphological analysis is one of the important tasks of Natural Language Processing. IR and Machine Translation are two important applications of Morphological Analyzer. In order to search or retrieve any information from web we need, normalize form of a word .This task is accomplish by two morphological analyzer Stemmer and Lemmatizer . Stemming is a simple process of strip off the word endings to get the normalized form of a word, called stem. The process of stemming suffers from two types of errors i.e. over stemming and under stemming. We can handle Under stemming by making a database of exceptional words whereas Over stemming is handle by lemmatizer .Lemma is correct root word or dictionary word generated by applying some rules for affix removal and some additional rules for making a correct dictionary word. This paper presents design and development of lemmatizer for Sindhi language in Devnagri script. This is our first attempt to develop a lemmatizer for Sindhi language in devanagri script.

Semantic Deficits in ALS Related to Right Lingual/Fusiform Gyrus Network Involvement: A Case-Control Study

Background: The clinicopathological continuity between amyotrophic lateral sclerosis (ALS) and frontotemporal lobar degeneration (FTLD) is well known. Although patients with ALS demonstrate similar language symptoms to patients with FTLD, including semantic dementia, word reading impairments in ALS have not been well studied. are kanji-written with irregular pronunciation that applies to exception words and are useful for detecting semantic deficits in Japan. We investigated Jukujikun reading impairments and related changes in resting-state networks in patients with ALS. Methods: We enrolled 140 participants, including 71 ALS patients and 69 healthy controls. We examined demographic and comprehensive neuropsychological factors, including Jukujikun. We also performed resting-state functional magnetic resonance imaging with voxel-based graph analysis on a subset of participants who agreed to participate in the imaging study to investigate the regions and related networks associated with Jukujikun impairments in ALS. Findings: Fifty-two percent of ALS patients (N=37) exhibited a significantly reduced score in Jukujikun reading compared with controls. We classified those patients as ALS with positive Jukujikun deficit (ALS-JD+) and the other patients as ALS with negative deficit (ALS-JD-). Resting-state network analysis showed significantly decreased degree centrality in the right fusiform/lingual gyrus and increased degree centrality in the left inferior/middle temporal gyrus in ALS-JD+ compared with ALS-JD- and controls. Seed-based analysis also revealed that the dysfunction in the right fusiform/lingual gyrus depended on decreased connectivity with regions associated with visual word perception, semantic processing, and speech. On the other hand, the observed increased degree centrality in the left inferior/middle temporal gyrus was derived from regions related to semantic processing. Interpretation Dysfunction of the hub in the right fusiform/lingual gyrus can affect semantic deficit in ALS. Considering neuropsychological symptoms as network impairments is vital for understanding various diseases. Funding MHLW and MEXT, Japan. Funding Statement: MHLW and MEXT, Japan. Declaration of Interests: The authors declare no competing interests. Ethics Approval Statement: The study conformed to the Ethical Guidelines for Medical and Health Research Involving Human Subjects endorsed by the Japanese government. The study protocol was approved by the Ethics Review Committee of the Nagoya University Graduate School of Medicine, and written informed consent was obtained from all participants.

The Ventral Anterior Temporal Lobe has a Necessary Role in Exception Word Reading.

An influential account of reading holds that words with exceptional spelling-to-sound correspondences (e.g., PINT) are read via activation of their lexical-semantic representations, supported by the anterior temporal lobe (ATL). This account has been inconclusive because it is based on neuropsychological evidence, in which lesion-deficit relationships are difficult to localize precisely, and functional neuroimaging data, which is spatially precise but cannot demonstrate whether the ATL activity is necessary for exception word reading. To address these issues, we used a technique with good spatial specificity—repetitive transcranial magnetic stimulation (rTMS)—to demonstrate a necessary role of ATL in exception word reading. Following rTMS to left ventral ATL, healthy Japanese adults made more regularization errors in reading Japanese exception words. We successfully simulated these results in a computational model in which exception word reading was underpinned by semantic activations. The ATL is critically and selectively involved in reading exception words.

"To Name or Not to Name: That is the Question": The Role of Response Inhibition in Reading.

Reading is a complex process that includes the integration of information about letters (graphemes) and sounds (phonemes). In many circumstances, such as noisy environments, response inhibition is an additional factor that plays a marked role in successful oral reading. Response inhibition can take the form of task relevant inhibition (i.e., foils in a go/no-go task) and task irrelevant inhibition (i.e., distractor information). Here we investigated task relevant inhibition by having participants (N = 30) take part in two tasks: go/no-go naming with nonwords foils (GNG-NW) and go/no-go naming with pseudohomophones foils (GNG-PH). Also, we investigated the addition of task irrelevant inhibition by having participants (N = 28) take part in two tasks: GNG-NW + information masking and GNG-PH + information masking. We provide evidence that during a task relevant inhibition task, sub-word sound level information can be successfully inhibited, as evidenced by comparable response times for regular words and exception words, provided the foils do not contain familiar sound-based information (GNG-NW). In contrast, regular words were read aloud faster than exception words in a GNG-PH task, indicating that sub-word level interference occurs when the foils contain familiar sound-based information. The addition of task irrelevant inhibition (i.e., information masking at the phoneme level), served to increase response time overall, but did not impact the pattern of response times between regular words and exception words. Together these findings provide useful information regarding the role of response inhibition in word recognition and may be useful in computational models of word recognition and future work may benefit from accounting for the effects outlined in this paper.

One tract, two tract, old tract, new tract: A pilot study of the structural and functional differentiation of the inferior fronto-occipital fasciculus

Abstract The inferior fronto-occipital fasciculus (IFOF), a major ventral white matter pathway, has been shown to be a crucial component of semantic (Moritz-Gasser, Herbet & Duffau, 2013) and lexical/orthographic (Vandermosten, Boets, Polemans, Sunaert, Wouters & Ghesquiere, 2012) processing. However, recent anatomical studies of the brain have revealed at least two differentiable components of the IFOF: a dorsal component projecting from the frontal lobe to the superior parietal lobule and a ventral component connecting the frontal lobe with the inferior occipital gyrus and posterior temporal lobe (Martino, Brogna, Robles, Vergani & Duffau, 2010). We have replicated this anatomical division using a deterministic tractography protocol in DTI Studio with high inter-rater reliability (ICC > 0.9). Furthermore, we provide the first evidence of a functional distinction between these two components. We compared diffusion measures (fractional anisotropy [FA], as well as MD, AD, RD) with reaction times on five different reading tasks: basic naming of pure exception words, regular words, and mixed exception/regular words, and go/no-go tasks involving either pseudohomophone or nonword foils. We found a functional divide in the left IFOF, whereby dorsal FA was correlated with performance on tasks that required higher levels of visual attention and response selection (go/no-go and mixed naming tasks), while ventral FA was more broadly correlated with naming performance. This suggests that the anatomical distinction described by Martino et al. (2010) is indeed mirrored by a functional distinction, and indicates that future investigations of neuroanatomical models of reading and speech production should consider the dorsal and ventral IFOF as separate entities.

On-reading predominance over Kun-reading in progressive Gogi (Word-meaning) Aphasia: Characteristics of Kanji reading in the Japanese semantic dementia patients

Background: Progressive Gogi (word-meaning) aphasia is a Japanese manifestation of semantic dementia. Reading and writing impairments are similar to those in Western countries. Surface dyslexia and lexical agraphia (selective reading or writing impairment for irregular or exceptional words) are the cardinal features. We previously reported that On-reading preceded and Kun-reading was not recalled in reading of a single kanji character in a patient with semantic dementia.

On the dissociation between reaction time and response duration as a function of lexical and sublexical reading: An examination of phonetic decoding and computational models

ABSTRACTNeurobiological models of reading account for two ways in which orthography is converted to phonology: (1) familiar words, particularly those with exceptional spelling-sound mappings (e.g., shoe) access their whole-word lexical representations in the ventral visual stream, and (2) orthographically unfamiliar words, particularly those with regular spelling-sound mappings (i.e., pseudohomophones [PHs], which are orthographically novel but sound like real words; e.g., shue) are phonetically decoded via sublexical processing in the dorsal visual stream. The present study used a naming task in order to compare naming reaction time (RT) and response duration (RD) of exception and regular words to their PH counterparts. We replicated our earlier findings with words, and extended them to PH phonetic decoding by showing a similar effect on RT and RD of matched PHs. Given that the shorter RDs for exception words can be attributed to the benefit of whole-word processing in the orthographic word system, and the longer RTs for exception words to the conflict with phonetic decoding, our PH results demonstrate that phonetic decoding also involves top-down feedback from phonological lexical representations (e.g., activated by shue) to the orthographic representations of the corresponding correct word (e.g., shoe). Two computational models were tested for their ability to account for these effects: the DRC and the CDP+. The CDP+ fared best as it was capable of simulating both the regularity and stimulus type effect on RT for both word and PH identification, although not their over-additive interaction. Our results demonstrate that both lexical reading and phonetic decoding elicit a regularity dissociation between RT and RD that provides important constraints to all models of reading, and that phonetic decoding results in top-down feedback that bolsters the orthographic lexical reading process.

Reading difficulties in primary progressive aphasia in a regular language-speaking cohort of patients.

Reading impairment is an important feature in Primary Progressive Aphasia (PPA). The Spanish orthography entails completely regular spelling to sound correspondences, so reading disorders may be different to English. In the current study, reading, phonological and semantic abilities of 35 patients with the three variants of PPA, and 13 healthy volunteers were assessed. Brain metabolism was concomitantly obtained from each participant using 18F-fluorodeoxyglucose positron emission tomography imaging. Two main patterns of impairment were identified: difficulties in nonwords reading with preservation of exception words in agrammatic and logopenic aphasia, and the inverse pattern in semantic dementia. Left frontal and left parietotemporal regions were associated to nonwords reading, while the anterior temporal lobe was related to reading of exception words. These results support the usefulness of examining reading abilities in the differential diagnosis of PPA variants, and suggest potential types of words that could be used in Spanish to assess these patients.

The Role of the Left Anterior Temporal Lobe for Unpredictable and Complex Mappings in Word Reading.

The anterior temporal lobes (ATLs) have been consistently associated with semantic processing which, in turn, has a key role in reading aloud single words. This study aimed to investigate (1) the reading abilities in patients with the semantic variant of primary progressive aphasia (svPPA), and (2) the relationship between gray matter (GM) volume of the left ATL and word reading performance using voxel-based morphometry (VBM). Three groups of participants (svPPA, Alzheimer’s Disease, AD and healthy elderly adults) performed a reading task with exception words, regular words and pseudowords, along with a structural magnetic resonance imaging scan. For exception words, the svPPA group had a lower accuracy and a greater number of regularization errors as compared to the control groups of healthy participants and AD patients. Similarly, for regular words, svPPA patients had a lower accuracy in comparison with AD patients, and a greater number of errors related to complex orthography-to-phonology mappings (OPM) in comparison to both control groups. VBM analyses revealed that GM volume of the left ATL was associated with the number of regularization errors. Also, GM volume of the left lateral ATL was associated with the number of errors with complex OPM during regular word reading. Our results suggest that the left ATL might play a role in the reading of exception words, in accordance with its role in semantic processing. Results further support the involvement of the left lateral ATL in combinatorial processes, including the integration of semantic and phonological information, for both exception and regular words.

The Role of the Supplementary Motor Region in Overt Reading: Evidence for Differential Processing in SMA-Proper and Pre-SMA as a Function of Task Demands.

A differentiation in function between the pre-SMA (i.e., cognitive load) and the SMA-proper (i.e., motor execution) has been described (Zhang et al., Cereb Cortex 22:99-111, 2012). These differential SMA functions may be influential in overt reading tasks. The present study examined the relationships between various segments of the SMA and overt reading through the modulation of task demands in an effort to explore the complexity of the print-to-speech network. Skilled reading adults (N = 15) took part in five overt reading tasks: pure regular word reading, pure exception word reading, mixed regular word and exception word reading, go/no-go reading with nonword foils and go/no-go reading with pseudohomophone foils. Five regions of interest that spanned the pre-SMA to the SMA-proper were isolated. Behaviour-function relationships were tested to examine the associations between performance (response time) and brain activity (percent signal change). Further, the coherence between feedforward (SMA) and feedback (supramarginal gyrus) regions were explored to further refine the print-to-speech network. We found that the pre-SMA was related to cognitively demanding tasks (go/no-go with pseudohomophones), whereas the SMA-proper was related to an automatized task (pure regular words). Notably, only those tasks that required information from the feedback system (i.e., mixed word lists, go/no-go tasks) showed connections between SMA regions and the supramarginal gyrus, which is in line with the role of feedback and feedforward systems in the print-to-speech network. Together, these results support the notion that the pre-SMA and SMA-proper are sensitive to reading tasks that differentially invoke higher cognitive resources (mixed word lists, go/no-go) versus automatized articulation (pure lists), respectively. We discuss our findings in the context of print-to-speech neural networks.

Generalization from newly learned words reveals structural properties of the human reading system.

Connectionist accounts of quasiregular domains, such as spelling-sound correspondences in English, represent exception words (e.g., pint) amid regular words (e.g., mint) via a graded "warping" mechanism. Warping allows the model to extend the dominant pronunciation to nonwords (regularization) with minimal interference (spillover) from the exceptions. We tested for a behavioral marker of warping by investigating the degree to which participants generalized from newly learned made-up words, which ranged from sharing the dominant pronunciation (regulars), a subordinate pronunciation (ambiguous), or a previously nonexistent (exception) pronunciation. The new words were learned over 2 days, and generalization was assessed 48 hr later using nonword neighbors of the new words in a tempo naming task. The frequency of regularization (a measure of generalization) was directly related to degree of warping required to learn the pronunciation of the new word. Simulations using the Plaut, McClelland, Seidenberg, and Patterson (1996) model further support a warping interpretation. These findings highlight the need to develop theories of representation that are integrally tied to how those representations are learned and generalized. (PsycINFO Database Record

The Role of Semantic Knowledge in Learning to Read Exception Words

In research and clinical practice, oral and written language skills have often been treated as separate domains. Yet they do not exist independently. Reading skills are contingent upon earlier acquired oral language skills, and the process of reading fosters growth in oral language. The importance of semantic knowledge for reading comprehension is well-documented, but there is growing evidence that it also plays a significant role in word reading. In English, a distinction can be made between regular words that follow predictable spelling-sound mappings, and exception words that do not. Oral language knowledge may be particularly important for the latter as it functions to supplement partial decoding. For speech- language pathologists (SLPs), it is important to consider how remediation targeted at improving oral language skills may also elicit benefits for reading development, and conversely how reading might be used to support oral language development. Practitioners should be aware of the pattern of literacy impairments that they are likely to encounter in children with developmental language disorder, and how this relates to their oral language profiles. The purpose of this paper is to enable practitioners to generalize their knowledge and skills across the artificial boundaries that have traditionally separated these two domains.

Do You Read How I Read? Systematic Individual Differences in Semantic Reliance amongst Normal Readers.

The extent to which meaning is involved in reading aloud has proven an area of longstanding debate, and current computational models differ on this dimension. The connectionist triangle model proposes that normal individuals rely on semantic information for correct reading of words with atypical spelling-sound relationships, but to varying degrees. This proposed individual difference would account for the varying stage of decline at which patients with semantic dementia first show the reading impairment known as surface dyslexia. Recent neuroimaging data has provided validation of this view, showing that individual differences in degree of semantic reliance during exception word reading predict the amount of activation in left anterior temporal regions associated with semantic processing. This study aimed to establish the cognitive correlates of individual differences in semantic reliance during exception word reading. Experiment 1 used a subgrouping approach with 32 participants and found larger imageability and semantic priming effects specifically for exception word reading amongst high relative to low semantic reliance readers. High semantic reliance readers also tended to read nonwords more slowly than low semantic reliance readers. A second experiment used a regression approach with 129 readers and confirmed the relationship of degree of semantic reliance both to imageability effects in exception word reading and speed of nonword reading. Further, while the performance of the higher semantic readers revealed no significant association with semantic processing tasks, there was a negative relationship with rhyme processing tasks. We therefore speculate that differences in phonological abilities may be responsible for varying degrees of semantic reliance in reading aloud. This proposal accords with the results of functional imaging showing that higher semantic reliance during exception word reading corresponds to lower activation in left pre-central gyrus, an area associated with direct spelling sound mapping and phonological processing. Our results therefore establish the nature of systematic individual differences in degree of semantic involvement amongst normal readers, and suggest directions for future neuroimaging and computational modeling research to uncover their origins.