Refine
Year of publication
Document Type
- Article (23)
- Preprint (7)
- Part of a Book (5)
- Conference Proceeding (5)
- Doctoral Thesis (1)
- Working Paper (1)
Language
- English (42) (remove)
Keywords
- Korpus <Linguistik> (16)
- Sprachstatistik (15)
- Deutsch (8)
- Sprachwandel (5)
- Wortschatz (5)
- language complexity (4)
- time series analysis (4)
- Benutzer (3)
- Computerunterstützte Lexikographie (3)
- Englisch (3)
Publicationstate
- Veröffentlichungsversion (21)
- Zweitveröffentlichung (6)
- Postprint (5)
- Preprint (2)
Reviewstate
Publisher
- Cornell University (4)
- de Gruyter (4)
- MDPI (3)
- De Gruyter (2)
- IDS-Verlag (2)
- Oxford University Press (OUP) (2)
- Springer Nature (2)
- Benjamins (1)
- Buske (1)
- Institute for Specialised Communication and Multilingualism (1)
A comparison between morphological complexity measures: typological data vs. language corpora
(2016)
Language complexity is an intriguing phenomenon argued to play an important role in both language learning and processing. The need to compare languages with regard to their complexity resulted in a multitude of approaches and methods, ranging from accounts targeting specific structural features to global quantification of variation more generally. In this paper, we investigate the degree to which morphological complexity measures are mutually correlated in a sample of more than 500 languages of 101 language families. We use human expert judgements from the World Atlas of Language Structures (WALS), and compare them to four quantitative measures automatically calculated from language corpora. These consist of three previously defined corpus-derived measures, which are all monolingual, and one new measure based on automatic word-alignment across pairs of languages. We find strong correlations between all the measures, illustrating that both expert judgements and automated approaches converge to similar complexity ratings, and can be used interchangeably.
In this paper, an exploratory data-driven method is presented that extracts word-types from diachronic corpora that have undergone the most pronounced change in frequency of occurrence in a given period of time. Combined with statistical methods from time series analysis, the method is able to find meaningful patterns and relationships in diachronic corpora, an idea that is still uncommon in linguistics. This indicates that the approach can facilitate an improved understanding of diachronic processes.
In this paper, a method for measuring synchronic corpus (dis-)similarity put forward by Kilgarriff (2001) is adapted and extended to identify trends and correlated changes in diachronic text data, using the Corpus of Historical American English (Davies 2010a) and the Google Ngram Corpora (Michel et al. 2010a). This paper shows that this fully data-driven method, which extracts word types that have undergone the most pronounced change in frequency in a given period of time, is computationally very cheap and that it allows interpretations of diachronic trends that are both intuitively plausible and motivated from the perspective of information theory. Furthermore, it demonstrates that the method is able to identify correlated linguistic changes and diachronic shifts that can be linked to historical events. Finally, it can help to improve diachronic POS tagging and complement existing NLP approaches. This indicates that the approach can facilitate an improved understanding of diachronic processes in language change.
One of the fundamental questions about human language is whether all languages are equally complex. Here, we approach this question from an information-theoretic perspective. We present a large scale quantitative cross-linguistic analysis of written language by training a language model on more than 6500 different documents as represented in 41 multilingual text collections consisting of ~ 3.5 billion words or ~ 9.0 billion characters and covering 2069 different languages that are spoken as a native language by more than 90% of the world population. We statistically infer the entropy of each language model as an index of what we call average prediction complexity. We compare complexity rankings across corpora and show that a language that tends to be more complex than another language in one corpus also tends to be more complex in another corpus. In addition, we show that speaker population size predicts entropy. We argue that both results constitute evidence against the equi-complexity hypothesis from an information-theoretic perspective.
Classical null hypothesis significance tests are not appropriate in corpus linguistics, because the randomness assumption underlying these testing procedures is not fulfilled. Nevertheless, there are numerous scenarios where it would be beneficial to have some kind of test in order to judge the relevance of a result (e.g. a difference between two corpora) by answering the question whether the attribute of interest is pronounced enough to warrant the conclusion that it is substantial and not due to chance. In this paper, I outline such a test.
In the first volume of Corpus Linguistics and Linguistic Theory, Gries (2005. Null-hypothesis significance testing of word frequencies: A follow-up on Kilgarriff. Corpus Linguistics and Linguistic Theory 1(2). doi:10.1515/ cllt.2005.1.2.277. http://www.degruyter.com/view/j/cllt.2005.1.issue-2/cllt.2005. 1.2.277/cllt.2005.1.2.277.xml: 285) asked whether corpus linguists should abandon null-hypothesis significance testing. In this paper, I want to revive this discussion by defending the argument that the assumptions that allow inferences about a given population – in this case about the studied languages – based on results observed in a sample – in this case a collection of naturally occurring language data – are not fulfilled. As a consequence, corpus linguists should indeed abandon null-hypothesis significance testing.
In the first volume of Corpus Linguistics and Linguistic Theory, Gries (2005. Null-hypothesis significance testing of word frequencies: A follow-up on Kilgarriff. Corpus Linguistics and Linguistic Theory 1(2). doi:10.1515/cllt.2005.1.2.277. http://www.degruyter.com/view//cllt.2005.1.issue-2/cllt.2005.1.2.277/cllt.2005.1.2.277.xml: 285) asked whether corpus linguists should abandon null-hypothesis significance testing. In this paper, I want to revive this discussion by defending the argument that the assumptions that allow inferences about a given population – in this case about the studied languages – based on results observed in a sample – in this case a collection of naturally occurring language data – are not fulfilled. As a consequence, corpus linguists should indeed abandon null-hypothesis significance testing.
This thesis consists of the following three papers that all have been published in international peer-reviewed journals:
Chapter 3: Koplenig, Alexander (2015c). The Impact of Lacking Metadata for the Measurement of Cultural and Linguistic Change Using the Google Ngram Data Sets—Reconstructing the Composition of the German Corpus in Times of WWII. Published in: Digital Scholarship in the Humanities. Oxford: Oxford University Press. [doi:10.1093/llc/fqv037]
Chapter 4: Koplenig, Alexander (2015b). Why the quantitative analysis of dia-chronic corpora that does not consider the temporal aspect of time-series can lead to wrong conclusions. Published in: Digital Scholarship in the Humanities. Oxford: Oxford University Press. [doi:10.1093/llc/fqv030]
Chapter 5: Koplenig, Alexander (2015a). Using the parameters of the Zipf–Mandelbrot law to measure diachronic lexical, syntactical and stylistic changes – a large-scale corpus analysis. Published in: Corpus Linguistics and Linguistic Theory. Berlin/Boston: de Gruyter. [doi:10.1515/cllt-2014-0049]
Chapter 1 introduces the topic by describing and discussing several basic concepts relevant to the statistical analysis of corpus linguistic data. Chapter 2 presents a method to analyze diachronic corpus data and a summary of the three publications. Chapters 3 to 5 each represent one of the three publications. All papers are printed in this thesis with the permission of the publishers.
Frimer et al. (2015) claim that there is a linear relationship between the level of prosocial language and the level of public disapproval of US Congress. A re-analysis demonstrates that this relationship is the result of a misspecified model that does not account for first-order autocorrelated disturbances. A Stata script to reproduce all presented results is available as an appendix.
It was recently suggested in a study published in Nature Human Behaviour that the historical loosening of American culture was associated with a trade-off between higher creativity and lower order. To this end, Jackson et al. generate a linguistic index of cultural tightness based on the Google Books Ngram corpus and use this index to show that American norms loosened between 1800 and 2000. While we remain agnostic toward a potential loosening of American culture and a statistical association with creativity/order, we show here that the methods used by Jackson et al. are neither suitable for testing the validity of the index nor for establishing possible relationships with creativity/order.
In a previous study, Aceves and Evans present a large-scale quantitative information-theoretic analysis of parallel corpus data in ~1,000 languages to show that there are apparently strong associations between the way languages encode information into words and patterns of communication, e.g. the configuration of semantic information. During the peer review process, one reviewer raised the question of the extent to which the presented results depend on different corpus sizes (see the Peer Review File). This is a very important question given that most, if not all, of the quantities associated with word frequency distributions vary systematically with corpus size. While Aceves and Evans claim that corpus size does not affect the results presented, I challenge this view by presenting reanalyses of the data that clearly suggest that it does.
We start by trying to answer a question that has already been asked by de Schryver et al. (2006): Do dictionary users (frequently) look up words that are frequent in a corpus. Contrary to their results, our results that are based on the analysis of log files from two different online dictionaries indicate that users indeed look up frequent words frequently. When combining frequency information from the Mannheim German Reference Corpus and information about the number of visits in the Digital Dictionary of the German Language as well as the German language edition of Wiktionary, a clear connection between corpus and look-up frequencies can be observed. In a follow-up study, we show that another important factor for the look-up frequency of a word is its temporal social relevance. To make this effect visible, we propose a de-trending method where we control both frequency effects and overall look-up trends.
A central goal of linguistics is to understand the diverse ways in which human language can be organized (Gibson et al. 2019; Lupyan/Dale 2016). In our contribution, we present results of a large scale cross-linguistic analysis of the statistical structure of written language (Koplenig/Wolfer/Meyer 2023) we approach this question from an information-theoretic perspective. To this end, we conduct a large scale quantitative cross-linguistic analysis of written language by training a language model on more than 6,500 different documents as represented in 41 multilingual text collections, so-called corpora, consisting of ~3.5 billion words or ~9.0 billion characters and covering 2,069 different languages that are spoken as a native language by more than 90% of the world population. We statistically infer the entropy of each language model as an index of un. To this end, we have trained a language model on more than 6,500 different documents as represented in 41 parallel/multilingual corpora consisting of ~3.5 billion words or ~9.0 billion characters and covering 2,069 different languages that are spoken as a native language by more than 90% of the world population or ~46% of all languages that have a standardized written representation. Figure 1 shows that our database covers a large variety of different text types, e.g. religious texts, legalese texts, subtitles for various movies and talks, newspaper texts, web crawls, Wikipedia articles, or translated example sentences from a free collaborative online database. Furthermore, we use word frequency information from the Crúbadán project that aims at creating text corpora for a large number of (especially under-resourced) languages (Scannell 2007). We statistically infer the entropy rate of each language model as an information-theoretic index of (un)predictability/complexity (Schürmann/Grassberger 1996; Takahira/Tanaka-Ishii/Dębowski 2016). Equipped with this database and information-theoretic estimation framework, we first evaluate the so-called ‘equi-complexity hypothesis’, the idea that all languages are equally complex (Sampson 2009). We compare complexity rankings across corpora and show that a language that tends to be more complex than another language in one corpus also tends to be more complex in another corpus. This constitutes evidence against the equi-complexity hypothesis from an information-theoretic perspective. We then present, discuss and evaluate evidence for a complexity-efficiency trade-off that unexpectedly emerged when we analysed our database: high-entropy languages tend to need fewer symbols to encode messages and vice versa. Given that, from an information theoretic point of view, the message length quantifies efficiency – the shorter the encoded message the higher the efficiency (Gibson et al. 2019) – this indicates that human languages trade off efficiency against complexity. More explicitly, a higher average amount of choice/uncertainty per produced/received symbol is compensated by a shorter average message length. Finally, we present results that could point toward the idea that the absolute amount of information in parallel texts is invariant across different languages.
We introduce DeReKoGram, a novel frequency dataset containing lemma and part-of-speech (POS) information for 1-, 2-, and 3-grams from the German Reference Corpus. The dataset contains information based on a corpus of 43.2 billion tokens and is divided into 16 parts based on 16 corpus folds. We describe how the dataset was created and structured. By evaluating the distribution over the 16 folds, we show that it is possible to work with a subset of the folds in many use cases (e.g., to save computational resources). In a case study, we investigate the growth of vocabulary (as well as the number of hapax legomena) as an increasing number of folds are included in the analysis. We cross-combine this with the various cleaning stages of the dataset. We also give some guidance in the form of Python, R, and Stata markdown scripts on how to work with the resource.
Large-scale empirical evidence indicates a fascinating statistical relationship between the estimated number of language users and its linguistic and statistical structure. In this context, the linguistic niche hypothesis argues that this relationship reflects a negative selection against morphological paradigms that are hard to learn for adults, because languages with a large number of speakers are assumed to be typically spoken and learned by greater proportions of adults. In this paper, this conjecture is tested empirically for more than 2000 languages. The results question the idea of the impact of non-native speakers on the grammatical and statistical structure of languages, as it is demonstrated that the relative proportion of non-native speakers does not significantly correlate with either morphological or information-theoretic complexity. While it thus seems that large numbers of adult learners/speakers do not affect the (grammatical or statistical) structure of a language, the results suggest that there is indeed a relationship between the number of speakers and (especially) information-theoretic complexity, i.e. entropy rates. A potential explanation for the observed relationship is discussed.
Computational language models (LMs), most notably exemplified by the widespread success of OpenAI's ChatGPT chatbot, show impressive performance on a wide range of linguistic tasks, thus providing cognitive science and linguistics with a computational working model to empirically study different aspects of human language. Here, we use LMs to test the hypothesis that languages with more speakers tend to be easier to learn. In two experiments, we train several LMs—ranging from very simple n-gram models to state-of-the-art deep neural networks—on written cross-linguistic corpus data covering 1293 different languages and statistically estimate learning difficulty. Using a variety of quantitative methods and machine learning techniques to account for phylogenetic relatedness and geographical proximity of languages, we show that there is robust evidence for a relationship between learning difficulty and speaker population size. However, contrary to expectations derived from previous research, our results suggest that languages with more speakers tend to be harder to learn.
Less than one percent of words would be affected by gender-inclusive language in German press texts
(2024)
Research on gender and language is tightly knitted to social debates on gender equality and non-discriminatory language use. Psycholinguistic scholars have made significant contributions in this field. However, corpus-based studies that investigate these matters within the context of language use are still rare. In our study, we address the question of how much textual material would actually have to be changed if non-gender-inclusive texts were rewritten to be gender-inclusive. This quantitative measure is an important empirical insight, as a recurring argument against the use of gender-inclusive German is that it supposedly makes written texts too long and complicated. It is also argued that gender-inclusive language has negative effects on language learners. However, such effects are only likely if gender-inclusive texts are very different from those that are not gender-inclusive. In our corpus-linguistic study, we manually annotated German press texts to identify the parts that would have to be changed. Our results show that, on average, less than 1% of all tokens would be affected by gender-inclusive language. This small proportion calls into question whether gender-inclusive German presents a substantial barrier to understanding and learning the language, particularly when we take into account the potential complexities of interpreting masculine generics.