Brown Corpus Research Articles

A Method for Measuring Word Sequence Complexity of Text

ABSTRACT In linguistics and natural language processing, measuring text features is crucial for representing and revealing the properties of texts in terms of topics, genres, sentiment and more. Current methods predominantly rely on the frequency of linguistic units and rarely consider the syntagmatic properties of these units, which can reflect deeper linguistic characteristics (e.g. syntax and pragmatics). This paper proposes a general method for calculating the sequence complexity of text. By combining the features of word length and word frequency, two specific implementations of this method are provided. Using these two formulas, word sequence complexity of text based on the Brown Corpus and the Gutenberg Corpus is calculated. The results show that word sequence complexity of text indicates a characteristic of gradual stabilization. By comparing random texts with natural texts, it is found that word sequence complexity is influenced by syntactic rules rather than sentence order, and natural texts tend to alternate between words of varying lengths and frequencies. Classification experiment results indicate that the proposed word sequence complexity outperforms commonly used quantitative indicators for representing the genre of text, such as TTR, information entropy, Zipf’s law parameters and motifs.

TF-IDF combined rank factor Naive Bayesian algorithm for intelligent language classification recommendation systems

With the continuous improvement of smart language systems, a large amount of language text data has emerged. How to efficiently and accurately process these text data has become an important challenge. Therefore, a speech classification recommendation system based on improved Naive Bayesian algorithm is proposed. The system first adopts the traditional Bayesian algorithm to classify language texts. The Term Frequency-Inverse Document Frequency and rank factor are combined to increase the weight of feature languages. Then, the classified language texts are combined with the improved algorithm for language classification recommendation. Finally, performance testing and simulation applications are conducted on the system. From the results, in the Gutenberg corpus, the research algorithm had the highest accuracy and completeness, with 98.5 % and 91.6 %, respectively, and the lowest values were 92.6 % and 89.4 %. The average values were 95.5 % and 91.1 %, with an F1 value of about 92.6 %. In the Brown corpus, the average accuracy, completeness, and F1 value of the designed algorithm were 96.2 %, 91.2 %, and 93.2 %, respectively. When the number of online customers reached 1000, the response time of the designed Chinese system was 1.15 s, the classification recommendation accuracy was 95 %, and the system stability was about 83 % on average. The response time of the English system was 0.64 s, the classification recommendation accuracy was 96 %, and the system stability was about 90 % on average. It shows that the designed method can significantly enhance the operation accuracy of the classification recommendation system.

Text Segmentation Via Processes that Count the Number of Different Words Forward and Backward

ABSTRACT The paper is developing a new statistical approach to automatic partitioning of texts into parts belonging to different authors. It is based on the analysis of processes that counts the number of different words forward and backward. The theoretical study of the processes is based on the assumptions of an elementary probability model with a change point. We prove consistence of our statistical estimate of the point of concatenation in the case when the concatenated texts have different Zipf exponents. This method is being tested on the Brown corpus and also on newspaper texts in different languages. Testing shows a good estimate of the concatenation point. This method can be used in parallel with other text segmentation methods.

ВИКОРИСТАННЯ ЛІНГВІСТИЧНИХ КОРПУСІВ У ВИВЧЕННІ СЛОВОБУДОВИ ДІАЛЕКТІВ АМЕРИКАНСЬКОГО ВАРІАНТУ СУЧАСНОЇ АНГЛІЙСЬКОЇ МОВИ

The article deals with the principles of using the concepts of corpus linguistics in studying American English dialects. The origins of corpus linguistics of American English dialects are considered, and developments in corpus linguistics by W. Francis and G. Kuchera are investigated (Brown Corpus, 1964). The concept of «corpus linguistics» is defined, and some problematic aspects are described. The qualifications for using the corpus-linguistic approach in the study of dialect grammatical features of American English are defined. The grammatical differences between British and American English are outlined. "Spoken English" is defined as how the latter was dialectical in the United States. The American dialects and Standardized American English are distinguished. Attention is focused on comparative and legal developments in corpus linguistics of British and American English (Brown and Lancaster Corps). Discographic studies of spoken American (dialectical) are mentioned: Lancaster/IBM SEC and Corpus of Spoken American English. The peculiarities of using the International Corpus of English, which studies British and American English in philological and methodological unity, are outlined. The processes of expanding the language corpus of the American English language from 1990 till the present (through the prism of statistics from The Bank of English) are determined. The main methodological developments in the systematization of the corpus of American English from the point of view of the theory of Corpus Linguistics are indicated: software corpus; problem-parametric industry selections; text transcription; search and system work; description encoding of "preserved" dialect words and word formations.

An approach based on deep learning for Indian sign language translation

PurposeAccording to the Indian Sign Language Research and Training Centre (ISLRTC), India has approximately 300 certified human interpreters to help people with hearing loss. This paper aims to address the issue of Indian Sign Language (ISL) sentence recognition and translation into semantically equivalent English text in a signer-independent mode.Design/methodology/approachThis study presents an approach that translates ISL sentences into English text using the MobileNetV2 model and Neural Machine Translation (NMT). The authors have created an ISL corpus from the Brown corpus using ISL grammar rules to perform machine translation. The authors’ approach converts ISL videos of the newly created dataset into ISL gloss sequences using the MobileNetV2 model and the recognized ISL gloss sequence is then fed to a machine translation module that generates an English sentence for each ISL sentence.FindingsAs per the experimental results, pretrained MobileNetV2 model was proven the best-suited model for the recognition of ISL sentences and NMT provided better results than Statistical Machine Translation (SMT) to convert ISL text into English text. The automatic and human evaluation of the proposed approach yielded accuracies of 83.3 and 86.1%, respectively.Research limitations/implicationsIt can be seen that the neural machine translation systems produced translations with repetitions of other translated words, strange translations when the total number of words per sentence is increased and one or more unexpected terms that had no relation to the source text on occasion. The most common type of error is the mistranslation of places, numbers and dates. Although this has little effect on the overall structure of the translated sentence, it indicates that the embedding learned for these few words could be improved.Originality/valueSign language recognition and translation is a crucial step toward improving communication between the deaf and the rest of society. Because of the shortage of human interpreters, an alternative approach is desired to help people achieve smooth communication with the Deaf. To motivate research in this field, the authors generated an ISL corpus of 13,720 sentences and a video dataset of 47,880 ISL videos. As there is no public dataset available for ISl videos incorporating signs released by ISLRTC, the authors created a new video dataset and ISL corpus.

A Machine Translation System from Indian Sign Language to English Text

Sign language recognition and translation is a crucial step towards improving communication between the deaf and the rest of the society. According to the Indian Sign Language Research and Training Centre (ISLRTC), India has around 300 certified human interpreters. With such a shortage of human interpreters, an alternative service is desired that helps people to achieve smooth communicate with deaf. In this study, an approach is presented that translates ISL sentences in English text using MobileNetV2 model and neural machine translation (NMT). The system features ISL corpus created from Brown corpus using ISL grammar rules. The approach converts the ISL videos into ISL gloss sequence using MobileNetV2 model and recognised ISL gloss sequence is then fed to machine translation module. MobileNetV2 was proven best-suited model for recognition of ISL sentences and NMT gives better result than statistical machine translation (SMT) to convert ISL gloss sequence into English text. The automatic and human evaluation of the proposed approach gives 83.3% and 86.1% accuracy, respectively.

Conformal Prediction for Text Infilling and Part-of-Speech Prediction

Modern machine learning algorithms are capable of providing remarkably accurate point-predictions; however, questions remain about their statistical reliability. Unlike conventional machine learning methods, conformal prediction algorithms return confidence sets (i.e., set-valued predictions) that correspond to a given significance level. Moreover, these confidence sets are valid in the sense that they guarantee finite sample control over type 1 error probabilities, allowing the practitioner to choose an acceptable error rate. In our paper, we propose inductive conformal prediction (ICP) algorithms for the tasks of text infilling and part-of-speech (POS) prediction for natural language data. We construct new ICP-enhanced algorithms for POS tagging based on BERT (bidirectional encoder representations from transformers) and BiLSTM (bidirectional long short-term memory) models. For text infilling, we design a new ICP-enhanced BERT algorithm. We analyze the performance of the algorithms in simulations using the Brown Corpus, which contains over 57,000 sentences. Our results demonstrate that the ICP algorithms are able to produce valid set-valued predictions that are small enough to be applicable in real-world applications. We also provide a real data example for how our proposed set-valued predictions can improve machine generated audio transcriptions.

An auxiliary Part‐of‐Speech tagger for blog and microblog cyber‐slang

AbstractThe increasing impact of Web 2.0 involves a growing usage of slang, abbreviations, and emphasized words, which limit the performance of traditional natural language processing models. The state‐of‐the‐art Part‐of‐Speech (POS) taggers are often unable to assign a meaningful POS tag to all the words in a Web 2.0 text. To solve this limitation, we are proposing an auxiliary POS tagger that assigns the POS tag to a given token based on the information deriving from a sequence of preceding and following POS tags. The main advantage of the proposed auxiliary POS tagger is its ability to overcome the need of tokens' information since it only relies on the sequences of existing POS tags. This tagger is called auxiliary because it requires an initial POS tagging procedure that might be performed using online dictionaries (e.g., Wikidictionary) or other POS tagging algorithms. The auxiliary POS tagger relies on a Bayesian network that uses information about preceding and following POS tags. It was evaluated on the Brown Corpus, which is a general linguistics corpus, on the modern ARK dataset composed by Twitter messages, and on a corpus of manually labeled Web 2.0 data.

Multi-word units (and tokenization more generally): a multi-dimensional and largely information-theoretic approach

It has been argued that most of corpus linguistics involves one of four fundamental methods: frequency lists, dispersion, collocation, and concordancing. All these presuppose (if only implicitly) the definition of a unit: the element whose frequency in a corpus, in corpus parts, or around a search word are counted (or quantified in other ways). Usually and with most corpus-processing tools, a unit is an orthographic word. However, it is obvious that this is a simplifying assumption borne out of convenience: clearly, it seems more intuitive to consider because of or in spite of as one unit each rather than two or three. Some work in computational linguistics has developed multi-word unit (MWU) identification algorithms, which typically involve co-occurrence token frequencies and association measures (AMs), but these have not become widespread in corpus-linguistic practice despite the fact that recognizing MWUs like the above will have a profound impact on just about all corpus statistics that involve (simplistic notions of) words/units. In this programmatic proof-of-concept paper, I introduce and exemplify an algorithm to identify MWUs that goes beyond frequency and bidirectional association by also involving several well-known but underutilized dimensions of corpus-linguistic information: frequency: how often does a potential unit (like in_spite_of) occur?; dispersion: how widespread is the use of a potential unit?; association: how strongly attracted are the parts of a potential unit?; entropy: how variable is each slot in a potential unit? The proposed algorithm can use all these dimensions and weight them differently. I will (i) present the algorithm in detail, (ii) exemplify its application to the Brown corpus, (iii) discuss its results on the basis of several kinds of MWUs it returns, and (iv) discuss next analytical steps.

Using Human Intelligence to Test the Impact of Popular Preprocessing Steps and Feature Extraction in the Analysis of Human Language

More than half a century has passed since Chomsky’s theory of language acquisition, Green and colleagues’ first natural language processor Baseball, and the Brown Corpus creation. Throughout the early decades, many believed that once computers became powerful enough, the development of A.I. systems that could understand and interact with humans using our natural languages would quickly follow. Since then, Moore’s Law has basically held; computer storage and performance has kept pace with our imaginations. And yet, 60 years later, even with these dramatic advances in computer technology, we still face major challenges in using computers to understand human language. The authors suggest that these same exponential increases in computational power have led current efforts to rely too much on techniques designed to exploit raw computational power and, in so doing, efforts have been diverted from advancing and applying the theoretical study of language to the task. In support of this view, the authors provide empirical evidence exposing the limitations of techniques – such as n-gram extraction – used to pre-process language. In addition, the authors conducted an analysis comparing three leading natural-language processing question-answering systems to human performance, and found that human subjects far outperformed all question answering-systems tested. The authors conclude by advocating for efforts to discover new approaches that use computational power to support linguistic and cognitive approaches to natural language understanding, as opposed to current techniques founded on patterns of word frequency.

Semantic role labeling for knowledge graph extraction from text

This paper introduces TakeFive, a new semantic role labeling method that transforms a text into a frame-oriented knowledge graph. It performs dependency parsing, identifies the words that evoke lexical frames, locates the roles and fillers for each frame, runs coercion techniques, and formalizes the results as a knowledge graph. This formal representation complies with the frame semantics used in Framester, a factual-linguistic linked data resource. We tested our method on the WSJ section of the Peen Treebank annotated with VerbNet and PropBank labels and on the Brown corpus. The evaluation has been performed according to the CoNLL Shared Task on Joint Parsing of Syntactic and Semantic Dependencies. The obtained precision, recall, and F1 values indicate that TakeFive is competitive with other existing methods such as SEMAFOR, Pikes, PathLSTM, and FRED. We finally discuss how to combine TakeFive and FRED, obtaining higher values of precision, recall, and F1 measure.

Traces of American English in Pakistani English: A Comparative Multi-dimensional Study of Press Editorials

This study explores the linguistic variation of periodized data of Pakistani press editorials in comparison with American press editorials. Most of the previous research studies conducted on Pakistani English, in general, and press editorial, in particular, compare Pakistani English with British English. No study compares Pakistani press editorials from a period to American press editorials. To trace the influence of American English on Pakistani English, this study explores the phases of Pakistani press editorials that resemble or differ from American press editorials. Biber’s (1988) multi-dimensional modal was used as a theoretical framework for this study. It exploited a diachronic corpus of Pakistani press editorials which was divided into three temporally distanced phases: 1947-1951, 1971-1975, and 2012- 2016. The Biber tagger was employed to annotate grammatical features. To draw a comparison, Nini’s (2015) MAT results of the American (Brown) corpus were utilized. The regression method was used for the computation of factor scores and analysis of variance (ANOVA) was used to draw a comparison between the diachronic data of Pakistani press editorials and American press editorials. The findings indicate that phase 3 of Pakistani press editorials is the closest to American press editorials in producing informational and explicit discourse. Phase 1 of Pakistani press editorials and American press editorials are quite close in non-narrative and abstract style in their discourse production. Moreover, phase 2 of Pakistani press editorials and American press editorials with a small difference in mean scores produce overt expression of persuasion.

Part of Speech Tagging Using Hidden Markov Models

Abstract In this paper, we present a wide range of models based on less adaptive and adaptive approaches for a PoS tagging system. These parameters for the adaptive approach are based on the n-gram of the Hidden Markov Model, evaluated for bigram and trigram, and based on three different types of decoding method, in this case forward, backward, and bidirectional. We used the Brown Corpus for the training and the testing phase. The bidirectional trigram model almost reaches state of the art accuracy but is disadvantaged by the decoding speed time while the backward trigram reaches almost the same results with a way better decoding speed time. By these results, we can conclude that the decoding procedure it’s way better when it evaluates the sentence from the last word to the first word and although the backward trigram model is very good, we still recommend the bidirectional trigram model when we want good precision on real data.

The interaction of various temporal devices in the use of past followed by temporal nouns

Followed by a temporal noun, past can be synonymous with last, but not with non-deictically anchored previous (e.g., ‘I've not been feeling very well for the past/last/*previous few days’). Most dictionaries provide examples in which past occurs with the present perfect, giving the impression that past is incompatible with other tenses. A close examination of the token past retrieved from the Brown Corpus, the Frown Corpus, and the Corpus of Contemporary American English (coca), however, has revealed that it is not uncommon for past to occur with the simple past or even with the past perfect (e.g., ‘His wife had left him the past year for a wealthy Wall Street broker’) and that past occurring with the past perfect can be explained as instances of free indirect style, discourse freezing or difference in reality (i.e., the three discourse principles allowing a shift in viewpoint with last). Research on temporal reference has thus far been ‘strongly biased towards certain devices’ such as tense, Aktionsart, and aspect ( Klein, 2009 : 41). This study shows that our understanding of how time is encoded in language can benefit from research studies dealing with the relationship between tense-aspect, temporal adverbials and discourse principles.

Presuppositions and Assertions: A Case Study on Acquisition of Most Common Presupposition Triggers in Early Childhood

This research aims to find out the acquisition of lexical presupposition triggers in early childhood. Children acquire semantic knowledge later than acquiring phonemes and uttering words. At what age the acquisition of lexical presupposition triggers is acquired is the question answered through this research. I also identify the incorrect and sensible usage along with the acquisitional challenge. In order to do so, Brown corpus is utilised. Out of three children Adam, Sara and Eve, Adam’s speech is analysed for the presence of lexical presupposition triggers. Again, a lexical presupposition trigger occurred 189 times in Adam’s speech. The methodology involved two phases. In the first phase, I sifted each utterance carrying again according to age, context, and the status of the utterance. The findings are presented in the form of a table. The next phase involves analysis of the instances collected. Examples are analysed in terms of mere repetitions, frequent use and acquiring a sensible usage of the lexical trigger again. I used comparison tool to analyse the examples further making clear distinction between the phenomenon of assertion and presupposition. The results show that the phenomenon of presupposition is acquired by the age of three and half years. By year four, Adam starts acquiring different syntactic structures involving the usage of the lexical trigger again and shows sensible usage by age five. The acquisitional challenge is the age and lack of interaction with the outer environment. The implications of this research would be that it will prompt to make further research on other children and other lexical triggers in future.

English Raising Predicates and (Non-)Finite Clauses

In this paper, we present a diachronic and synchronic analysis of raising and extraposition constructions in the historical Brown Corpus and the more contemporary English Web Corpus 2015. We begin by establishing two diachronic facts: first, raising constructions are used much more frequently than their semantically equivalent extraposition variants, and second, the distribution of raising and extraposition remains – rather exceptionally in comparison to other structures allowing for finite/non-finite variation – diachronically consistent from the beginning of the 20th century to 2015. We then supplement this unique diachronic distribution with an analysis of the most recent corpus data, which shows that the choice between the two semantically equivalent constructions is governed by distinct structural factors unique to each construction. Concretely, we show that the raising construction is frequently used as a relative clause, whereas the extraposition variant generally resists such a syntactic role. By contrast, we show that a prominent factor in favour of extraposition relates to the negative marker, which is placed with similar frequency both in the matrix and in the embedded clause of the extraposition construction in contrast to the raising variant, which uses the negative marker almost exclusively in the matrix clause. Lastly, we show that extraposition constructions contain modal verbs in the matrix clause more frequently than the raising variants and we tie this observation to the idea that the clausal composition of the extraposition construction is structurally more suited for expressing tentativeness.

Functional Text Dimensions for the annotation of web corpora

This paper presents an approach to classifying large web corpora into genres by means of Functional Text Dimensions (FTDs). This offers a topological approach to text typology in which the texts are described in terms of their similarity to prototype genres. The suggested set of categories is designed to be applicable to any text on the web and to be reliable in annotation practice. Interannotator agreement results show that the suggested categories produce Krippendorff's α at above 0.76. In addition to the functional space of eighteen dimensions, similarity between annotated documents can be described visually within a space of reduced dimensions obtained through t-distributed Statistical Neighbour Embedding. Reliably annotated texts also provide the basis for automatic genre classification, which can be done in each FTD, as well as as within the space of reduced dimensions. An example comparing texts from the Brown Corpus, the BNC and ukWac, a large web corpus, is provided.

Zu theoretischen und praktischen Aspekten des Fachübersetzens

Abstract In the past few decades, it has extensively been written about corpus linguistics, which has owned its upswing mainly to the use of electronic corpora since the 1960s (Brown Corpus). Meanwhile, an increasing number of fields within general and applied linguistics (e.g. computational linguistics, discourse analysis, contrastive linguistics, diachronic and synchronic linguistics, language teaching and learning research, lexicology and lexicography, psycholinguistics, sociolinguistics, translation studies) have been using corpus linguistic methods. In linguistic research, the empirical and descriptive character of corpus-based linguistic analysis has also been given an emphasis. Thanks to the digital revolution of the 20th and 21st centuries the creation and provision of digital linguistic corpora is becoming accessible for smaller nations and language communities as well as for scientists. Nowadays, linguistic corpora cannot only be regarded as a tool to support language research and Translation Studies, but they also contribute to the enrichment of cultural diversity. The article focuses on international examples as well as on the most significant Hungarian corpora. The paper also discusses the criteria of corpus creation and several cultural aspects of corpus linguistics.

CRNN: A Joint Neural Network for Redundancy Detection

This article proposes a novel framework for detecting redundancy in supervised sentence categorisation. Unlike traditional singleton neural network, our model incorporates character-aware convolutional neural network (Char-CNN) with character-aware recurrent neural network (Char-RNN) to form a convolutional recurrent neural network (CRNN). Our model benefits from Char-CNN in that only salient features are selected and fed into the integrated Char-RNN. Char-RNN effectively learns long sequence semantics via sophisticated update mechanism. We compare our framework against the state-of-the-art text classification algorithms on four popular benchmarking corpus. For instance, our model achieves competing precision rate, recall ratio, and F1 score on the Googlenews data-set. For twenty-news-groups data stream, our algorithm obtains the optimum on precision rate, recall ratio, and F1 score. For Brown Corpus, our framework obtains the best F1 score and almost equivalent precision rate and recall ratio over the top competitor. For the question classification collection, CRNN produces the optimal recall rate and F1 score and comparable precision rate. We also analyse three different RNN hidden recurrent cells’ impact on performance and their runtime efficiency. We observe that MGU achieves the optimal runtime and comparable performance against GRU and LSTM. For TFIDF based algorithms, we experiment with word2vec, GloVe, and sent2vec embeddings and report their performance differences.

Prepositions from the Perspective of Cognitive Study

Prepositions as function words, plays very important role in English. Prepositions are deemed as very difficult items both in English learning and teaching, because the senses of them are quite complex and abundant and the usage and collocation of them are quite flexible and complex. According to the English corpus statistics, there is a preposition in every eight word in any English text. In American English Brown Corpus, 12.21% of the word is a preposition; in its corresponding LOB corpus of British English, there are 12.34% of the word is preposition. Thus, the preposition plays an indispensable role in the English vocabulary. At the same time, prepositions research plays an important role in cognitive linguistics.