Inflected Language Research Articles

Attitudes towards anglicisms and the English language in the Bosnian/Croatian/Serbian language used by IT professionals and students in Bosnia and Herzegovina

Anglicisms, words borrowed from the English language, which do not have to be of English origin, are observed in the Bosnian/Croatian/Serbian language in Bosnia and Herzegovina. In addition, attitudes towards the English language by a certain group of native Bosnian/Croatian/Serbian speakers were observed. The research includes insights into the adaptation of anglicisms as an important point in borrowing into an inflective language. The research observes the attitudes towards the usage of anglicisms in the Bosnian/Croatian/Serbian language and each of them separately. Additionally, attitudes towards the English language, familiarity with anglicisms and proper use of the observed anglicisms has been included in the research. The familiarity with anglicisms is observed through the understanding of the meaning and their proper usage. The participants are random native speakers of the Bosnian/Croatian/Serbian language who belong to the Internet Technology (IT) field profession and are based in Bosnia and Herzegovina. The results show that both the attitude towards anglicisms and the English language are vastly positive, although a total of 44% of the observed anglicisms are translated. The positive attitude towards anglicisms implies their frequent use, mainly adapted to the different systems of the Bosnian/Croatian/Serbian language. The usage is reflected in the two possibilities: English words used in their original form or adapted on different levels. In addition, as the need for it appeared along, the adaptation was also observed. The original or an adapted form and how far the adaptation goes in the system of Bosnian/Croatian/Serbian language (phonology and morphology) The research also displays the advisability and growing need for anglicisms in this field and confirms the significance of those words in the Bosnian/Croatian/Serbian language. The findings imply that there is probably another parameter or parameter (variable(s) influencing the attitude, which could easily be the trend and the global reputation of the English language.

A deep learning approach to building a framework for Urdu POS and NER

The study examines various studies on Named Entity Recognition (NER) and Part of Speech (POS) tagging for the Urdu language conducted by academics and researchers. POS and NER tagging for Urdu still faces obstacles in terms of increasing accuracy while lowering false-positive rates and labelling unknown terms, despite the efforts of numerous researchers. In addition, ambiguity exists when tagging terms with different contextual meanings within a sentence. Due to the fact that Urdu is an inflectional, derivational, morphologically rich, and context-sensitive language, the existing models, such as Linguistic rule application, N-gram Markov model, Tree Tagger, random forest (RF) tagger, etc., were unable to produce accurate experimental results on Urdu language data. The significance of this study is that it fills a gap in the literature concerning the lack of POS and NER tagging for the Urdu language. For Urdu POS and NER tagging, we propose a deep learning model with a well-balanced set of language-independent features as well as a survey of important Urdu POS/NER techniques. In addition, this is the first study to use residual biDirectional residual Long short-term memory (residual biLSTM) architecture trained on the Urmono dataset in conjunction with the randomly initialised word2vec, fastText and mBERT embeddings are utilised to generate word or character vectors.For each experiment, the paper also employs the evaluation methods of Macro-F1, precision, precision, and recall. The proposed method with mbert embedding as word vectors provides best results of F1 score for POS and NER at 91.11% and 99.11% respectively. Also, the accuracy, precision and recall for POS are reported at 94.85%, 91.79% and 90.77%. Similarly, the accuracy, precision and recall for NER of the proposed model are reported at 99.77%, 98.78% and 99.45% respectively, which are higher than baseline models.

Getting it right: Teaching undergraduate biology to undermine racial essentialism.

How we teach human genetics matters for social equity. The biology curriculum appears to be a crucial locus of intervention for either reinforcing or undermining students' racial essentialist views. The Mendelian genetic models dominating textbooks, particularly in combination with racially inflected language sometimes used when teaching about monogenic disorders, can increase middle and high school students' racial essentialism and opposition to policies to increase equity. These findings are of particular concern given the increasing spread of racist misinformation online and the misappropriation of human genomics research by white supremacists, who take advantage of low levels of genetics literacy in the general public. Encouragingly, however, teaching updated information about the geographical distribution of human genetic variation and the complex, multifactorial basis of most human traits, reduces students' endorsement of racial essentialism. The genetics curriculum is therefore a key tool in combating misinformation and scientific racism. Here, we describe a framework and example teaching materials for teaching students key concepts in genetics, human evolutionary history, and human phenotypic variation at the undergraduate level. This framework can be flexibly applied in biology and anthropology classes and adjusted based on time availability. Our goal is to provide undergraduate-level instructors with varying levels of expertise with a set of evidence-informed tools for teaching human genetics to combat scientific racism, including an evolving set of instructional resources, as well as learning goals and pedagogical approaches. Resources can be found at https://noto.li/YIlhZ5. Additionally, we hope to generate conversation about integrating modern genetics into the undergraduate curriculum, in light of recent findings about the risks and opportunities associated with teaching genetics.

Applying Distributional Semantic Models to a Historical Corpus of a Highly Inflected Language: the Case of Ancient Greek

So-called “distributional” language models have become dominant in research on the computational modelling of lexical semantics. This paper investigates how well such models perform on Ancient Greek, a highly inflected historical language. It compares several ways of computing such distributional models on the basis of various context features (including both bag-of-words features and syntactic dependencies). The performance is assessed by evaluating how well these models are able to retrieve semantically similar words to a given target word, both on a benchmark we designed ourselves as well as on several independent benchmarks. It finds that dependency features are particularly useful to calculate distributional vectors for Ancient Greek (although the level of granularity that these dependency features should have is still open to discussion) and discusses possible ways for further improvement, including addressing problems related to polysemy and genre differences.

Structural differences of syntactic models in Russian and Chinese

For many centuries of scientific existence linguistic researchers and philosophic leaders have been trying to establish a rigid determination of linguistic concepts. Theories of formal generative grammar provide the opportunity for logically proved abstract models. Such models allowing to reflect the typological features of Chinese and Russian syntactical structures in the most accessible way. According to the language morphological classification, the world's languages are divided into four morphological groups: a. inflectional languages, b. agglutinative languages, c. isolating languages, d. incorporating languages. The Chinese language is related to the morphological group of isolating languages as it produces poor methods of morphological inflection and strong significance of the word order in a sentence. The morphological structure of the Russian language which belongs to the inflectional morphological group is opposed to it. The current paper aims to present typological differences between the two comparable languages in terms of generative linguistics. The generated research produces two typological schemes of structural differences between the analyzed languages. The analysis is supplied with certain examples of language usage in both linguistic cultures.

B-NER: A Novel Bangla Named Entity Recognition Dataset With Largest Entities and Its Baseline Evaluation

Within the Natural Language Processing (NLP) framework, Named Entity Recognition (NER) is regarded as the basis for extracting key information to understand texts in any language. As Bangla is a highly inflectional, morphologically rich, and resource-scarce language, building a balanced NER corpus with large and diverse entities is a demanding task. However, previously developed Bangla NER systems are limited to recognizing only three familiar entities: person, location, and organization. To address this significant limitation, we introduce a novel Bangla NER dataset B-NER, which was created using 22,144 manually annotated Bangla sentences collected from Bangla newspapers and Bangla Wikipedia. This dataset includes a total of 9,895 unique words which were manually categorized into eight different entity types, such as a person, organization, event, artifact, time indicator, natural phenomenon, geopolitical entity, and geographical location. Inter-annotator agreement experiments were conducted to validate the quality of annotations performed by three annotators, resulting in a Kappa score of 0.82. In this paper, we provide an outline of the annotation guideline illustrated with examples, discuss the B-NER dataset properties, and present benchmark evaluations of the dataset. To establish that B-NER is more comprehensive and balanced in comparison to other publicly accessible datasets, we conducted cross-dataset modeling and validation, i.e. trained NER model on one dataset while tested on another, and found that the model trained on B-NER performed the best in that settings. Furthermore, we performed exhaustive benchmark evaluations based on Bidirectional LSTM with fastText embeddings and sentence transformer models. Among these models, fine-tuned IndicBERT achieved noticeable results with a Macro-F1 of 86%. This dataset and baseline results will be publicly available under a CC-BY 4.0 license in the CoNLL-2002 format to facilitate further research on Bangla NER.

Sentiment classification of annual reports in Slovak language using symbolic, subsymbolic and statistic methods

Background: Sentiment analysis as a tool for recognizing emotions from bank reports can serve to capture the intentions and opinions of corporations, change the emotional context during turbulent periods, and to verify the compliance of the context of the reports with banking conservatism. The purpose of the presented study is to describe the process of creating an own dataset from annual reports in the Slovak language with marked sentiment and the subsequent search for models capable of classifying the sentiment of reports as positive, negative, or neutral.Methods: Our methodology consists of a series of steps necessary for obtaining the text and its subsequent preprocessing to create a dataset. The study further deals with the iterative search for the best combination of hyperparameters for symbolic, subsymbolic, statistic and homogenous symbolic ensemble models in order to find the best classifying model.Results: The experimental results show that the subsymbolic methods are the most suitable for the classifying sentiment of Slovak annual reports. The proposed model can classify sentiment with 98% accuracy.Conclusion: The classification of sentiment in inflectional languages requires particularly precise pre-preparation to give. For the Slovak language, subsymbolic and statistical machine learning methods work best for the sentiment classification of annual reports.

TECHNOLOGY AND PROBLEMS OF AUTOMATIC LINGUOANALYSIS OF THE MORPHOLOGICAL COMPOSITION OF WORDS

The article discusses the technology of the morphological analyzer, which implements the program of morphological markup in the National corpus of the Kazakh language. The introduction of morphological markup is the most basic and important linguistic analysis of the National Corpus. It is also considered how language units in the intermediate layer are differentiated in the dictionary of basic words and the dictionary of grammatical forms (word forms) that we are developing. It is difficult for a computer to automatically divide words into roots and affixes and describe the relation of words to parts of speech and grammatical characteristics of affixes. However, in agglutinative languages, such as Kazakh, it is easier to automatically separate words and automatically analyze the composition of words than in inflectional languages. This is due to the fact that in agglutinative languages affixes are added in a certain system. The formal model of word forms is more understandable. The article discusses some difficulties of automatic parsing and analysis of words in the Kazakh language. The article also discusses problematic issues of functional affixes in the modeling of the morphological system of the Kazakh language, as well as issues related to categories and conditionally accepted codes included in the grammatical dictionary.

Generating a sentiment dictionary in R and dictionary-based sentiment analysis in Turkish texts

Abstract Dictionary-based sentiment analysis is a text mining application that allows comments about the sentimental states of the text or documents through the sentimental poles of the words. In recent years, it has become quite popular in many disciplines such as trade, health, education, usage for various purposes. It is applied in many languages depending on the sentiment dictionaries, which are the main component, are created. Unlike inflectional languages such as English and German, the number of dictionary-based sentiment analysis studies using the Turkish language, which is a sequencing language, is very limited. When the literature of sentiment analysis is examined separately in terms of the material used; it is seen that the number of studies on long-format literary texts, which have been functioning as sentiment transmission throughout history, is also quite limited. In the light of these knowledges, a comprehensive method of creating a domain-specific sentiment dictionary and applying dictionary-based sentiment analysis on Turkish texts is proposed in the R software.

Lithuanian priegaidė: a syllable or a word feature?

The place of priegaidė ‘syllable intonation’ in the phonological system of Lithuanian remains uncertain; so does its typological peculiarity. The article aims to discover the specific nature of this phonological feature taking into consideration the opposition between phonemic (inflectional) and syllabemic (isolating) language types. Phonetically similar prosodic means in these languages show noticeable functional differences determined by the structure of the meaningful units they belong to. Inflectional languages develop a system of suprasegmental phonological features, which integrate parts of their normally composite units, whereas isolating languages usually do with monosyllabic words (naturally, without stress) and monolithic morphemes with tones as distinctive features. Standard Lithuanian has priegaidė only on the stressed syllables, therefore its system can be called polyaccentual (not polytonic), i. e. possessing several types of accent. The dialects, however, show examples of relevant priegaidė besides syllables with primary stress. In this, more general, system priegaidė functions as a suprasegmental feature of a morpheme (exposed to fading in weak positions). The traditional differentiation of syllables according to priegaidė is secondary: a syllable provides conditions for the realization of priegaidė, but being part of the plane of expression, not connected with meaning, the syllable itself does not need differentiation.

Semantics and Ways of Expressing the Plural of Nouns in Russian and Japanese Languages

Purpose. The study compares the ways to express the plural, as well as categories connected with these ways, and the semantics of plurality in Russian and Japanese languages which are genetically and typologically unrelated: Russian is an inflectional language and Japanese is an agglutinative one.Results. There is a grammatical category of singular and plural in Russian. This category is missing in Japanese, but there is a lexical-semantic category. The concept of singularity and plurality is not represented in Japanese. Number may be actualized by the speaker if the object (animate or inanimate) enters the speaker’s personal zone or a personification takes place. There is only one way to express plurality in Russian – by using flexions of plurality. There are four ways to do that in Japanese which are markers of plurality, classifiers, reduplication and combined dvandva (a word with coordinative connection between components of the whole word). The usage of inflections to express plurality is mandatory in Russian, but the usage of plurality markers in Japanese is probabilistic.Conclusion. A common category for Russian and Japanese languages which is important for expressing plurality is the category of animation. The similar meanings of plurality both in Russian and in Japanese is collective, representative and indefinite discrete plurality expressed in different ways.

UTILIZING ROOTS AND PATTERNS TO IDENTIFY ARABIC NAMED ENTITIES

Named Entity Recognition NER is a subset of information extraction that seeks to recognize and categorize named things in text data into specified categories, such as people's names, organizations' names, geographic locations, and so on. This task has recently attracted a lot of attention due to the discovery it has the potential to boost the performance of a variety of NLP applications. In the domains of Question Answering and Summarization Systems, Information Retrieval and Extraction, Machine Translation, Video Annotation, Semantic Web Search, and Bioinformatics, the majority of difficulties require named entity recognition. Arabic is an inflectional language, which allows for non-concatenative morphological operations on the root. The purpose of this study is to extract and recognize entity names from Arabic articles. We proposed an algorithm for determining names from roots using patterns. We developed it in Python and leveraged the "pyqt5" visual package to see the results immediately, as well as modify and add patterns easily. To replicate the names, we used a random sample of 400 names and 45 different patterns. The algorithm correctly identified 370 names easily and quickly, yielding a success rate of 93%. All names with the same recognized names will be known in the same way by the method and do not need any manipulation in code or design. The names that are not recognized by our algorithm have no roots in the list of known Arabic roots. Our research shows that the approach can recognize names with roots with high speed and accuracy, but it is not possible to identify nouns that are not in the Arabic language using this method. As a result, we recommend using a hybrid method that incorporates multiple concepts.

О сходствах и различиях в употреблении пассива в русской и турецкой речи

The author of the article conducts a comparative study of the use of forms of the passive voice in Turkish and Russian speech. In the course of scientific analysis, striking differences are found in the use of verb forms expressing the meaning of the passive. The Russian language has several ways of expressing this, the Turkish language is more consistent in this respect and conveys the meaning of suffering in one way, and adheres to the regularity in the use of voice forms in speech. The author demonstrates various language strategies in coding the same mental content, which can help in understanding the language mechanisms themselves, as well as facilitate the perception of the complex grammatical meaning of the passive for speakers of opposite languages for more effective mastering of a foreign language. The author conducts research within the framework of system linguistics, believing that the reason for the language ways of expressing a particular meaning should be sought in the determinant of language types. The Russian language is an inflectional language, the Turkish language is included in the group of agglutinative languages. Different ways of expressing the meaning of the passive are rooted in a different grammatical principle: Russian speakers have a set of pre-prepared ways of expression, while Turkish speakers have a set of separate morphemes, from which they form a word form at the moment of speaking.

현대중국어 양상부사 연구 - 양상부사의 특징, 범위, 하위분류를 중심으로

This study mainly investigated modality adverbs in Modern Chinese. To investigate the characteristics, subclassification, and scope of Chinese modality adverbs, We have studied on ‘concepts of modality and mood’ and ‘subclassification of modality adverbs and mood adverbs’ through previous studies. After that, We presented the perspective of this paper, and reclassified 324 adverbs that presented in previous studies based on ‘modal meanings’. It has come to the following conclusions:BR First, in previous studies, the concepts of modality and mood are mixed. In the field of Western grammar, some scholars claim that modality is the semantic field, mood is the grammar field. This reflects the characteristics of ‘inflectional language’, but Chinese is ‘isolated language’. So at the time of accepting the concepts of modality, it can be seen that the concepts were mixed. In order to solve this problem, this paper distinguished concepts of modality(情態), mood(语氣₂), tone(口气).BR Second, in previous studies, each scholar’s classification for modality adverbs and mood adverbs didn’t match. However, researching on the adverbs suggested in previous studies, all of them can be distinguished as dynamic modality, deontic modality, epistemic modality, tone, etc., but tone doesn’t mean modality. So modality adverbs are generally divided into dynamic modality, deontic modality and epistemic modality. Also recently some scholars claim that ‘evaluative modality’ should be classified as a category of modality, but none of the adverbs can be found to indicate the meaning of evaluative modality.BR Third, after collecting a total of 324 adverbs of modality adverbs and mood adverbs in previous studies, 211 adverbs were excluded for two reasons(not included or presented in a different meaning in XIANDAI HANYU CIDIAN(『现代汉语词典(第7版)』), not indicated the meaning of modality). The remaining 113 adverbs were reclassified into dynamic modality(will), deontic modality, and epistemic modality. Adverbs with multiple modality in the same word were indicated by ‘Bi(必)₁, Bi(必)₂, Bi(必)₃’, and a total of 156 adverbs were classified in this paper. The results of the reclassification are as follows. There are a total of 29 dynamic modality adverbs accounting for about 19% of the total modality adverbs, and a total of 46 deontic modality adverbs accounting for about 29%, and a total of 81 epistemic modality adverbs accounting for about 52%. As the classification results, epistemic modality adverbs are the largest number of modality adverbs.

Towards Reliable Baselines for Document-Level Sentiment Analysis in the Czech and Slovak Languages

This article helps establish reliable baselines for document-level sentiment analysis in highly inflected languages like Czech and Slovak. We revisit an earlier study representing the first comprehensive formulation of such baselines in Czech and show that some of its reported results need to be significantly revised. More specifically, we show that its online product review dataset contained more than 18% of non-trivial duplicates, which incorrectly inflated its macro F1-measure results by more than 19 percentage points. We also establish that part-of-speech-related features have no damaging effect on machine learning algorithms (contrary to the claim made in the study) and rehabilitate the Chi-squared metric for feature selection as being on par with the best performing metrics such as Information Gain. We demonstrate that in feature selection experiments with Information Gain and Chi-squared metrics, the top 10% of ranked unigram and bigram features suffice for the best results regarding online product and movie reviews, while the top 5% of ranked unigram and bigram features are optimal for the Facebook dataset. Finally, we reiterate an important but often ignored warning by George Forman and Martin Scholz that different possible ways of averaging the F1-measure in cross-validation studies of highly unbalanced datasets can lead to results differing by more than 10 percentage points. This can invalidate the comparisons of F1-measure results across different studies if incompatible ways of averaging F1 are used.

Stylistic Fingerprints, POS-tags, and Inflected Languages: A Case Study in Polish

ABSTRACT In stylometric investigations, frequencies of the most frequent words (MFWs) and character n-grams outperform other style-markers, even if their performance varies significantly across languages. In inflected languages, word endings play a prominent role, and hence different word forms cannot be recognized using generic text tokenization. Countless inflected word forms make frequencies sparse, making most statistical procedures complicated. Presumably, applying one of the NLP techniques, such as lemmatization and/or parsing, might increase the performance of classification. The aim of this paper is to examine the usefulness of grammatical features (as assessed via POS-tag n-grams) and lemmatized forms in recognizing authorial profiles, in order to address the underlying issue of the degree of freedom of choice within lexis and grammar. Using a corpus of Polish novels, we performed a series of supervised authorship attribution benchmarks, in order to compare the classification accuracy for different types of lexical and syntactic style-markers. Even if the performance of POS-tags as well as lemmatized forms was notoriously worse than that of lexical markers, the difference was not substantial and never exceeded ca. 15%.

A human evaluation of English-Slovak machine translation

ABSTRACT The paper aims to obtain an error profile for machine translation (MT) from English into Slovak. We present an adjusted framework for MT evaluation, which is based on Vanko's categorical framework, but reflects machine translation peculiarities of synthetic and/or inflectional languages. Based on the framework, we analyse the errors generated by Google Translate and identify the most frequent categories of errors occurring in machine translation when translating newspaper articles from English into Slovak. While we have seen research on widely-spoken languages, such as English or other major official EU languages, little is known about Slovak, which is also an official EU language. This paper provides the first human MT evaluation study of English-Slovak machine translation using professional translators for a more detailed depiction of translation quality. Our research has revealed that the highest numbers of errors occurred in the sphere of lexical semantics, as well as in syntactic-semantic correlativeness, both being closely related. Additionally, based on the results of the Cochran Q test, we show how individual MT errors located in the examined categories differ in co-incidence and in how they impact translation quality.

REPRESENTATION OF SIMULATIVE ACTION MODALITY IN LANGUAGES WITH DIFFERENT STRUCTURES

The paper presents an analysis of the simulative action category and its representation in German and Udmurt. These languages are typologically different: German is an inflectional language, while Udmurt belongs to agglutinative languages. The modal meaning «simulative action» has its specific semantic features. Despite the fact that it is semantically close to the comparison meaning, it is also connected with the modality and influenced by subjective evaluation of a speaker. The representation of modal meaning «simulative action» covers all levels of the language system: morphological, lexical and syntactic. The simulative action meaning on the morphological level in the Udmurt language is expressed by the complex suffix -myask- and its combinations with the negative suffix -te-: -temyas ’ k-, -mteyas ’ k- . In German there are no specific affixes to express simulative action. Other linguistic means like comparative conjunctions kad ’ , vyllem in Udmurt or wie in German are used very rarely and can be considered as additional ones. Linguistic means of the lexical level demonstrate a large diversity: German verbs sich verstellen, sich stellen, sich anstellen, vortäuschen, sich aufspielen, heucheln, imitieren, simulieren, fingieren , Udmurt verbs analskyny, analtiskyny . In spite of wide range of semantically similar verbs, all of them except sich verstellen are used only in combination with other lexemes in restricted contexts. There are also some German lexemes derived from the verb scheinen . In both languages the semantics of the simulative action on the syntactic level can be expressed by syntactic units including an equivalent of the verb «to do» - karyny and tun. The authors specify the term «simulative action verbs» in Udmurt. They believe that the term is used incorrectly, when it comes to the verbal forms with suffixes -myas ’ k-, -temyas ’ k-, -mteyas ’ k- .

Finite-state super transducers for compact language resource representation in edge voice-AI

Finite-state transducers have been proven to yield compact representations of pronunciation dictionaries used for grapheme-to-phoneme conversion in speech engines running on low-resource embedded platforms. However, for highly inflected languages even more efficient language resource reduction methods are needed. In the paper, we demonstrate that the size of finite-state transducers tends to decrease when the number of word forms in the modelled pronunciation dictionary reaches a certain threshold. Motivated by this finding, we propose and evaluate a new type of finite-state transducers, called ‘finite-state super transducers’, which allow for the representation of pronunciation dictionaries by a smaller number of states and transitions, thereby significantly reducing the size of the language resource representation in comparison to minimal deterministic final-state transducers by up to 25%. Further, we demonstrate that finite-state super transducers exhibit a generalization capability as they may accept and thereby phonetically transform even inflected word forms that had not been initially represented in the original pronunciation dictionary used for building the finite-state super transducer. This method is suitable for speech engines operating on platforms at the edge of an AI system with restricted memory capabilities and processing power, where efficient speech processing methods based on compact language resources must be implemented.

Simplifying grammatical gender in inflectional languages: Odessa Russian and beyond

SummaryThis work aims to contribute to the analysis of the morphosyntactic processes of gender assignment and gender agreement in inflectional languages, through the study of gender variation in Russian. We will focus on the data of a special contact variety, Odessa Russian (OdR), and compare it to standard, dialectal, child Russian, as well as other contact varieties of Russian. OdR is a linguistic variety slightly decomplexified by non-native acquisition, which arises from a special language contact situation; it originated as alingua franca, and was passed on later to successive generations of speakers as a native variety. Some of the most striking features of OdR are its divergences in gender assignment and agreement with respect to Standard Russian. The specific processes analysed in this paper are classified in three groups: (i) loss of a gender value (neuter gender), by virtue if a strategy of gender (re)assignment; (ii) transfer or ‘migration’ of gender according to the phonological shape of the words involved (masculine into feminine and feminine into masculine); interestingly, formal rules in this group of processes can sometimes prevail over semantic rules (i. e. over natural gender); and (iii) disruptions of gender agreement (associated with disruption of grammatical case), which can be interpreted as the simplification of the corresponding syntactic tree by eliminating uninterpretable gender features in the language. We will show that these processes go beyond mere substrata effects, and proceed according to more general processes that partially take place also in other (contact and non-contact) varieties of Russian. More specifically: (i) occasional changes in the assignment and agreement of gender are reminiscent of dialectal Russian, but are more widespread in OdR (closer to child language) than in those varieties; (ii) even if gender in OdR was not lost as a grammatical category, some productions point to a partial loss of gender features, reminding us of pidgins and heritage languages.