TY - JOUR TI - Language Dataset Reuse in Methodological Research on Down-Sampling Corpus Data AU - Sönning, Lukas PY - 2026 SP - 1-10 PB - Otto-Friedrich-Universität CY - Bamberg UR - https://fis.uni-bamberg.de/handle/uniba/116551 LA - eng AB - This paper looks at opportunities and challenges of language dataset reuse in methodological research, using the development of down-sampling designs for corpus-linguistic work as a case study. This research is concerned with the construction and evaluation of strategies for down-sizing data that are too large given the resource constraints confronting a study. Work in this line of inquiry is heavily dependent on real datasets to field-test and compare techniques, which makes it a prime example of how open data can serve as a catalyst for methodological progress. In the reuse case discussed here, a methodological study made use of an informally published dataset. The original data were revised and modified, and the derived dataset archived in the Tromsø Repository of Language and Linguistics (TROLLing). The current paper draws attention to a number of questions that may arise when repurposing language data, including issues related to licensing and the question of how to determine authorship credit. It also demonstrates how depositors and reusers of data can profit from a careful review process, which is ideally assisted by trained data curators. The case study points to TROLLing as an exemplary repository that offers quality control and substantive feedback on a host of aspects including dataset documentation as well as legal and research-ethical matters. KW - methodological research KW - down-sampling KW - corpus linguistics KW - data curation KW - licensing KW - TROLLing ER - TY - JOUR TI - Language Dataset Reuse in Methodological Research on Down-Sampling Corpus Data AU - Sönning, Lukas PY - 2026 JO - Journal of open humanities data : JOHD VL - 12 IS - 78 SP - 1-10 PB - Ubiquity Press CY - London SN - 2059-481X DO - 10.5334/johd.531 UR - https://fis.uni-bamberg.de/handle/uniba/116550 LA - eng AB - This paper looks at opportunities and challenges of language dataset reuse in methodological research, using the development of down-sampling designs for corpus-linguistic work as a case study. This research is concerned with the construction and evaluation of strategies for down-sizing data that are too large given the resource constraints confronting a study. Work in this line of inquiry is heavily dependent on real datasets to field-test and compare techniques, which makes it a prime example of how open data can serve as a catalyst for methodological progress. In the reuse case discussed here, a methodological study made use of an informally published dataset. The original data were revised and modified, and the derived dataset archived in the Tromsø Repository of Language and Linguistics (TROLLing). The current paper draws attention to a number of questions that may arise when repurposing language data, including issues related to licensing and the question of how to determine authorship credit. It also demonstrates how depositors and reusers of data can profit from a careful review process, which is ideally assisted by trained data curators. The case study points to TROLLing as an exemplary repository that offers quality control and substantive feedback on a host of aspects including dataset documentation as well as legal and research-ethical matters. KW - methodological research KW - down-sampling KW - corpus linguistics KW - data curation KW - licensing KW - TROLLing ER - TY - BOOK TI - Juggling Intra- and Extralinguistic Factors in the Study of Variation and Change ED - Werner, Valentin ED - Schützler, Ole ED - Schlüter, Julia ED - Knappe, Gabriele ED - Sönning, Lukas ED - Vetter, Fabian PY - 2026 SP - 460 PB - de Gruyter Mouton CY - Berlin T3 - Topics in English Linguistics SN - 978-3-11-914658-6 DO - 10.1515/9783112208281 UR - https://fis.uni-bamberg.de/handle/uniba/116279 LA - eng ER - TY - CHAP TI - Integrating the study of intra- and extralinguistic factors in variation and change in English AU - Schlüter, Julia AU - Werner, Valentin AU - Schützler, Ole AU - Knappe, Gabriele AU - Vetter, Fabian AU - Sönning, Lukas PY - 2026 T2 - Juggling Intra- and Extralinguistic Factors in the Study of Variation and Change SP - 3-23 PB - de Gruyter Mouton CY - Berlin SN - 978-3-11-914658-6 DO - 10.1515/9783112208281-001 UR - https://fis.uni-bamberg.de/handle/uniba/116281 LA - eng ER - TY - JOUR TI - Ordinal response scales : Psychometric grounding for design and analysis AU - Sönning, Lukas PY - 2025 SP - 1-18 PB - Otto-Friedrich-Universität CY - Bamberg UR - https://fis.uni-bamberg.de/handle/uniba/107123 LA - eng AB - Ordinal response scales are commonly used in applied linguistics. To summarize the distribution of ratings or judgments provided by informants, these are usually converted into numbers and then averaged or analyzed with ordinary regression models. This approach has been criticized in the literature; one caveat (among others) is the assumption that distances between categories are known. The present paper illustrates how empirical insights into the perception of response labels may inform the design and analysis stage of a study. We start with a review of how ordinal scales are used in linguistic research. Our survey offers insights into typical scale layouts and analysis strategies, and it allows us to identify three commonly used rating dimensions (agreement, intensity, and frequency). We take stock of the experimental literature on the perception of relevant scale point labels and then demonstrate how psychometric insights may direct scale design and data analysis. This includes a careful consideration of measurement-theoretic and statistical issues surrounding the numeric-conversion approach to ordinal data. We focus on the consequences of these drawbacks for the interpretation of empirical findings, which will enable researchers to make informed decisions and avoid drawing false conclusions from their data. We present a case study on yous(e) in two varieties of English, which shows that reliance on psychometric scale values can alter statistical conclusions, while also giving due consideration to the key limitations of the numeric-conversion approach to ordinal data analysis. KW - Ordinal data KW - Rating scales KW - Acceptability judgments KW - Likert scale KW - Judgment task KW - Measurement KW - Psychological scaling ER - TY - JOUR TI - Down-sampling from hierarchically structured corpus data AU - Sönning, Lukas PY - 2024 JO - International journal of corpus linguistics : IJCL VL - 29 IS - 4 SP - 507-533 PB - Benjamins CY - Amsterdam [u.a.] SN - 1569-9811 DO - 10.1075/ijcl.23079.son UR - https://fis.uni-bamberg.de/handle/uniba/113986 LA - eng ER - TY - JOUR TI - Ordinal response scales : Psychometric grounding for design and analysis AU - Sönning, Lukas PY - 2024 JO - Research methods in applied linguistics : RMAL VL - 3 IS - 3, 100156 SP - 1-18 PB - Elsevier ScienceDirect CY - Amsterdam SN - 2772-7661 DO - 10.1016/j.rmal.2024.100156 UR - https://fis.uni-bamberg.de/handle/uniba/107122 LA - eng AB - Ordinal response scales are commonly used in applied linguistics. To summarize the distribution of ratings or judgments provided by informants, these are usually converted into numbers and then averaged or analyzed with ordinary regression models. This approach has been criticized in the literature; one caveat (among others) is the assumption that distances between categories are known. The present paper illustrates how empirical insights into the perception of response labels may inform the design and analysis stage of a study. We start with a review of how ordinal scales are used in linguistic research. Our survey offers insights into typical scale layouts and analysis strategies, and it allows us to identify three commonly used rating dimensions (agreement, intensity, and frequency). We take stock of the experimental literature on the perception of relevant scale point labels and then demonstrate how psychometric insights may direct scale design and data analysis. This includes a careful consideration of measurement-theoretic and statistical issues surrounding the numeric-conversion approach to ordinal data. We focus on the consequences of these drawbacks for the interpretation of empirical findings, which will enable researchers to make informed decisions and avoid drawing false conclusions from their data. We present a case study on yous(e) in two varieties of English, which shows that reliance on psychometric scale values can alter statistical conclusions, while also giving due consideration to the key limitations of the numeric-conversion approach to ordinal data analysis. KW - Ordinal data KW - Rating scales KW - Acceptability judgments KW - Likert scale KW - Judgment task KW - Measurement KW - Psychological scaling ER - TY - JOUR TI - Seeing the wood for the trees : Predictive margins for random forests AU - Sönning, Lukas AU - Grafmiller, Jason PY - 2024 SP - 1-26 PB - Otto-Friedrich-Universität CY - Bamberg UR - https://fis.uni-bamberg.de/handle/uniba/96274 LA - eng AB - Classification trees and random forests offer a number of attractive features to corpus data analysts. However, the way in which these models are typically reported – a decision tree and/or set of variable importance scores – offers insufficient information if interest centers on the (form of) relationship between (multiple) predictors and the outcome. This paper develops predictive margins as an interpretative approach to ensemble techniques such as random forests. These are model summaries in the form of adjusted predictions, which provide a clearer picture of patterns in the data and allow us to query a model on potential non-linear associations and interactions among predictor variables. The present paper outlines the general strategy for forming predictive margins and addresses methodological issues from an explicitly (corpus) linguistic perspective. For illustration, we use data on the English genitive alternation and provide an R package and code for their implementation. KW - average predictive comparisons KW - classification trees KW - interpretable machine learning KW - predictive modeling KW - random forests ER - TY - JOUR TI - Evaluation of keyness metrics : Performance and reliability AU - Sönning, Lukas PY - 2024 SP - 1-24 PB - Otto-Friedrich-Universität CY - Bamberg UR - https://fis.uni-bamberg.de/handle/uniba/96272 LA - eng AB - The methodological debates surrounding keyword analysis have given rise to a wide range of keyness metrics. The present paper delineates four dimensions of keyness, which distinguish between frequency-and dispersion-related perspectives. Existing measures are then organized according to these dimensions and evaluated with regard to their performance on a specific keyword analysis task: The identification of key verbs in academic writing. To this end, the rankings produced by 32 different metrics are evaluated against an established academic word list. Further, the reliability of measures is assessed, to determine whether they produce stable rankings across repeated studies on the same pair of text varieties. We observe notable differences among metrics with regard to these criteria. Our findings provide further support for the superiority of the Wilcoxon rank sum test and textdispersion–based measures, and allow us to identify, within each dimension of keyness, metrics that may be given preference in applied work. KW - keyness KW - keyword analysis KW - frequency KW - dispersion KW - methodology ER - TY - GEN TI - Latent-variable modeling of ordinal outcomes in language data analysis AU - Sönning, Lukas AU - Krug, Manfred AU - Vetter, Fabian AU - Schmid, Timo AU - Leucht, Anne AU - Messer, Paul PY - 2024 SP - 1-22 PB - Otto-Friedrich-Universität CY - Bamberg UR - https://fis.uni-bamberg.de/handle/uniba/95382 LA - eng AB - In empirical work, ordinal variables are typically analyzed using means based on numeric scores assigned to categories. While this strategy has met with justified criticism in the methodological literature, it also generates simple and informative data summaries, a standard often not met by statistically more adequate procedures. Motivated by a survey of how ordered variables are dealt with in language research, we draw attention to an un(der)used latent-variable approach to ordinal data modeling, an alternative perspective on the most widely used form of ordered regression, the cumulative model. Since the latent-variable approach is not mentioned in statistical textbooks for linguists and does not feature in any of the studies in our survey, we believe it is worthwhile to promote its benefits. To this end, we draw on questionnaire-based preference ratings by speakers of Maltese English, who indicated on a 5-point scale which of two synonymous expressions (e.g. package-parcel) they (tend to) use. We demonstrate that a latent-variable analysis affords nuanced and interpretable data summaries that can be visualized effectively, while at the same time avoiding limitations inherent in mean response models (e.g. distortions induced by floor and ceiling effects). The online supplementary materials include a tutorial for its implementation in R. KW - Latent-variable modeling ER - TY - JOUR TI - Latent-Variable Modelling of Ordinal Outcomes in Language Data Analysis AU - Sönning, Lukas AU - Krug, Manfred AU - Vetter, Fabian AU - Schmid, Timo AU - Leucht, Anne AU - Messer, Paul PY - 2024 JO - Journal of quantitative linguistics : official journal of the International Quantitative Linguistics Association (IQLA) VL - 31 IS - 2 SP - 77-106 PB - Informa UK Limited CY - London SN - 1744-5035 DO - 10.1080/09296174.2024.2329448 UR - https://fis.uni-bamberg.de/handle/uniba/95412 LA - eng ER - TY - GEN TI - Latent-variable modeling of ordinal outcomes in language data analysis AU - Sönning, Lukas AU - Krug, Manfred AU - Vetter, Fabian AU - Schmid, Timo AU - Leucht, Anne AU - Messer, Paul PY - 2024 SP - 1-22 PB - Center for Open Science DO - 10.31219/osf.io/jhv6b UR - https://fis.uni-bamberg.de/handle/uniba/93359 LA - eng AB - In empirical work, ordinal variables are typically analyzed using means based on numeric scores assigned to categories. While this strategy has met with justified criticism in the methodological literature, it also generates simple and informative data summaries, a standard often not met by statistically more adequate procedures. Motivated by a survey of how ordered variables are dealt with in language research, we draw attention to an un(der)used latent-variable approach to ordinal data modeling, an alternative perspective on the most widely used form of ordered regression, the cumulative model. Since the latent-variable approach is not mentioned in statistical textbooks for linguists and does not feature in any of the studies in our survey, we believe it is worthwhile to promote its benefits. To this end, we draw on questionnaire-based preference ratings by speakers of Maltese English, who indicated on a 5-point scale which of two synonymous expressions (e.g. package-parcel) they (tend to) use. We demonstrate that a latent-variable analysis affords nuanced and interpretable data summaries that can be visualized effectively, while at the same time avoiding limitations inherent in mean response models (e.g. distortions induced by floor and ceiling effects). The online supplementary materials include a tutorial for its implementation in R. KW - Latent-variable modeling ER - TY - CHAP TI - (Re-)viewing the Acquisition of Rhythm in the Light of L2 Phonological Theories AU - Sönning, Lukas PY - 2023 SP - 123-157 PB - Otto-Friedrich-Universität CY - Bamberg UR - https://fis.uni-bamberg.de/handle/uniba/107456 LA - eng AB - Previous work on non-native speech rhythm has often drawn on L2 phonological theory for the interpretation of findings. The explicit confrontation of theory-derived hypotheses with data remains scarce, however. This paper illustrates how a hypothetico-deductive approach can contribute to our understanding of L2 speech rhythm. We consider cross-sectional data on prominence alternations in German learner speech from the viewpoint of two dynamic frameworks: The Ontogeny Phylogeny Model (OPM) and the Linguistic Theory of L2 Phonological Development (LTD). While both theories deal with L1-independent, universal forces in L2 acquisition, the OPM further considers the role of L1 transfer, similarity, and markedness. The predictions we formulate based on the two models lead us to pursue distinct methodological strategies. While our reading of the OPM prompts us to measure speech rhythm as a single, global category of speech, the LTD suggests a more nuanced, componential approach to L2 rhythm. Our application of the OPM confronts us squarely with the limited utility of rhythm metrics for L2 speech research and points to a number of issues at the theory-data interface. Overall, the LTD generates more informative predictions and provides a richer framework for the empirical study of prominence grading in L2 speech. KW - L2 Phonological Theories ER - TY - CHAP TI - Drawing on principles of perception : the line plot AU - Sönning, Lukas PY - 2023 PB - Otto-Friedrich-Universität CY - Bamberg UR - https://fis.uni-bamberg.de/handle/uniba/92358 LA - eng AB - This paper draws attention to an underused display type for corpus data visualization: the line plot. While this graph type is commonly associated with time series data, its true potential arguably unfolds in the application to multifactorial data sets involving discrete (categorical) variables. Data layouts of this kind are typical of corpus-based work and the preferred vehicle for their visualization is currently the bar chart. It is sometimes argued that line plots should only be used when the horizontal axis represents a continuous trait. However, once we allow for the levels of binary and categorical variables to be connected by lines, we recognize that this form offers several distinct advantages over bar charts. This is especially true for visualization tasks involving multiple predictor variables. The paper starts out by providing some theoretical background for the comparative evaluation of graph types, with a focus on quantitative comparisons and perceptual processing. Drawing on empirical insights into visual perception, evidence-based recommendations for the design of line plots are given. These include the choice of line types and plotting symbols, the use of direct labeling, and the arrangement of variables in the display. Following this, key advantages of line plots are illustrated. These include pictorial minimalism, the availability of extended encoding strategies and the scaffolding provided by perceptual grouping laws. The paper closes by emphasizing limitations of this display type. These concern the depiction of non-continuous x-variables and the asymmetric perception of interactions among predictor variables. While ample attention should be paid to these issues, we argue for a (more) routine use of line plots in corpus data visualization. KW - line plot ER - TY - BOOK TI - Data visualization in corpus linguistics : Reflections and future directions ED - Sönning, Lukas ED - Schützler, Ole PY - 2023 PB - VARIENG CY - Helsinki T3 - Studies in Variation, Contacts and Change in English UR - https://fis.uni-bamberg.de/handle/uniba/92326 LA - eng ER - TY - CHAP TI - Drawing on principles of perception : the line plot AU - Sönning, Lukas PY - 2023 T2 - Data visualization in corpus linguistics : Reflections and future directions PB - VARIENG CY - Helsinki UR - https://fis.uni-bamberg.de/handle/uniba/92327 LA - eng AB - This paper draws attention to an underused display type for corpus data visualization: the line plot. While this graph type is commonly associated with time series data, its true potential arguably unfolds in the application to multifactorial data sets involving discrete (categorical) variables. Data layouts of this kind are typical of corpus-based work and the preferred vehicle for their visualization is currently the bar chart. It is sometimes argued that line plots should only be used when the horizontal axis represents a continuous trait. However, once we allow for the levels of binary and categorical variables to be connected by lines, we recognize that this form offers several distinct advantages over bar charts. This is especially true for visualization tasks involving multiple predictor variables. The paper starts out by providing some theoretical background for the comparative evaluation of graph types, with a focus on quantitative comparisons and perceptual processing. Drawing on empirical insights into visual perception, evidence-based recommendations for the design of line plots are given. These include the choice of line types and plotting symbols, the use of direct labeling, and the arrangement of variables in the display. Following this, key advantages of line plots are illustrated. These include pictorial minimalism, the availability of extended encoding strategies and the scaffolding provided by perceptual grouping laws. The paper closes by emphasizing limitations of this display type. These concern the depiction of non-continuous x-variables and the asymmetric perception of interactions among predictor variables. While ample attention should be paid to these issues, we argue for a (more) routine use of line plots in corpus data visualization. KW - line plot ER - TY - CHAP TI - (Re-)viewing the Acquisition of Rhythm in the Light of L2 Phonological Theories AU - Sönning, Lukas PY - 2023 T2 - Speech rhythm in learner and second language varieties of English SP - 123-157 PB - Springer CY - Singapore SN - 978-981-19-8939-1 DO - 10.1007/978-981-19-8940-7_6 UR - https://fis.uni-bamberg.de/handle/uniba/90969 LA - eng AB - Previous work on non-native speech rhythm has often drawn on L2 phonological theory for the interpretation of findings. The explicit confrontation of theory-derived hypotheses with data remains scarce, however. This paper illustrates how a hypothetico-deductive approach can contribute to our understanding of L2 speech rhythm. We consider cross-sectional data on prominence alternations in German learner speech from the viewpoint of two dynamic frameworks: The Ontogeny Phylogeny Model (OPM) and the Linguistic Theory of L2 Phonological Development (LTD). While both theories deal with L1-independent, universal forces in L2 acquisition, the OPM further considers the role of L1 transfer, similarity, and markedness. The predictions we formulate based on the two models lead us to pursue distinct methodological strategies. While our reading of the OPM prompts us to measure speech rhythm as a single, global category of speech, the LTD suggests a more nuanced, componential approach to L2 rhythm. Our application of the OPM confronts us squarely with the limited utility of rhythm metrics for L2 speech research and points to a number of issues at the theory-data interface. Overall, the LTD generates more informative predictions and provides a richer framework for the empirical study of prominence grading in L2 speech. KW - L2 Phonological Theories ER - TY - CHAP TI - Comparing Study Designs and Down-Sampling Strategies in Corpus Analysis : The Importance of Speaker Metadata in the BNCs of 1994 and 2014 AU - Sönning, Lukas AU - Krug, Manfred PY - 2023 SP - 127-160 PB - Otto-Friedrich-Universität CY - Bamberg UR - https://fis.uni-bamberg.de/handle/uniba/59831 LA - eng AB - This chapter throws into relief the importance of the link between corpus hits and their sources (i.e. texts or speakers) by comparing study designs uninformed and informed by such metadata. The authors’ argument draws attention to the consequences of uneven distributions of observations across the basic text units of a corpus. In the case study (a distributional analysis of the use of actually in the BNCs of 1994 and 2014), it is demonstrated that disproportionate word counts contributed by individual sources (in this case, speakers) will distort estimates for relevant subsections of corpus data (in this case, demographic groups defined by age and gender) if the analysis assigns the same weight to every observation. The proposed solution is to factor in the text (or speaker) level, but this hinges on the availability of the relevant metadata. Moreover, insights into the hierarchical structure of corpus data are shown to benefit the design stage of a study. Thus, if manual post-processing steps preclude an exhaustive analysis of corpus hits, insights into the organization of data points can direct down-sampling strategies to generate a statistically efficient subset of tokens. KW - Corpus Analysis ER - TY - JOUR TI - Vowel reduction in German Learner English : An instrumental study AU - Sönning, Lukas PY - 2023 SP - 163-173 PB - Otto-Friedrich-Universität CY - Bamberg UR - https://fis.uni-bamberg.de/handle/uniba/59607 LA - eng AB - This study investigates the production of vowels in unstressed syllables by advanced German learners of English in comparison with native speakers of Standard Southern British English. Two acoustic properties were measured: duration and formant structure. The results indicate that duration of unstressed vowels is similar in the two groups, though there is some variation depending on the phonetic context. In terms of formant structure, learners produce slightly higher F1 and considerably lower F2, the difference in F2 being statistically significant for each learner. Formant values varied as a function of context and orthographic representation of the vowel. KW - vowel reduction KW - acoustic analysis KW - SLA KW - L1 German KW - L2 English ER - TY - CHAP TI - Comparing Standard Reference Corpora and Google Books Ngrams : Strengths, Limitations and Synergies in the Contrastive Study of Variable h- in British and American English AU - Sönning, Lukas AU - Schlüter, Julia PY - 2023 SP - 17-45 PB - Otto-Friedrich-Universität CY - Bamberg UR - https://fis.uni-bamberg.de/handle/uniba/59543 LA - eng AB - This chapter is based on two standard reference corpora, the British National Corpus and the Corpus of Contemporary American English, as opposed to the multi-billion-word database of Google Books Ngrams, which has, despite its allure, not been used in many systematic linguistic studies so far. Focusing on indefinite article allomorphy (a vs an) as an orthographic cue to the phonological strength of ‹h›-onsets in British and American English, the size advantage of the Ngrams database expectedly plays out in larger type and token counts, more stable estimates and fewer distortions due to data sparsity. However, as metadata are extremely limited (to year and variety), a fully accountable analysis is not feasible. The case study illustrates how richly annotated corpora can shed light on potential disturbances arising from two sources: genre differences and between-author variability. A sensitivity analysis offers some degree of reassurance when extending the analysis to the Ngrams database. In this way, the authors demonstrate that the strengths and limitations of corpora and big data resources can, with due caution, be counterbalanced to answer questions of linguistic interest. KW - Google Books Ngrams KW - big data KW - metadata KW - type frequency KW - token frequency KW - hierarchical data structure KW - corpus comparability KW - data quality ER - TY - JOUR TI - Evaluation of keyness metrics : Performance and reliability AU - Sönning, Lukas PY - 2023 JO - Corpus Linguistics and Linguistic Theory VL - 20 IS - 2 SP - 263-288 PB - de Gruyter CY - Berlin SN - 1613-7035 DO - 10.1515/cllt-2022-0116 UR - https://fis.uni-bamberg.de/handle/uniba/59524 LA - eng AB - The methodological debates surrounding keyword analysis have given rise to a wide range of keyness metrics. The present paper delineates four dimensions of keyness, which distinguish between frequency-and dispersion-related perspectives. Existing measures are then organized according to these dimensions and evaluated with regard to their performance on a specific keyword analysis task: The identification of key verbs in academic writing. To this end, the rankings produced by 32 different metrics are evaluated against an established academic word list. Further, the reliability of measures is assessed, to determine whether they produce stable rankings across repeated studies on the same pair of text varieties. We observe notable differences among metrics with regard to these criteria. Our findings provide further support for the superiority of the Wilcoxon rank sum test and textdispersion–based measures, and allow us to identify, within each dimension of keyness, metrics that may be given preference in applied work. KW - keyness KW - keyword analysis KW - frequency KW - dispersion KW - methodology ER - TY - JOUR TI - Seeing the wood for the trees : Predictive margins for random forests AU - Sönning, Lukas AU - Grafmiller, Jason PY - 2023 JO - Corpus Linguistics and Linguistic Theory VL - 20 IS - 1 SP - 153-181 PB - de Gruyter CY - Berlin SN - 1613-7035 DO - 10.1515/cllt-2022-0083 UR - https://fis.uni-bamberg.de/handle/uniba/59523 LA - eng AB - Classification trees and random forests offer a number of attractive features to corpus data analysts. However, the way in which these models are typically reported – a decision tree and/or set of variable importance scores – offers insufficient information if interest centers on the (form of) relationship between (multiple) predictors and the outcome. This paper develops predictive margins as an interpretative approach to ensemble techniques such as random forests. These are model summaries in the form of adjusted predictions, which provide a clearer picture of patterns in the data and allow us to query a model on potential non-linear associations and interactions among predictor variables. The present paper outlines the general strategy for forming predictive margins and addresses methodological issues from an explicitly (corpus) linguistic perspective. For illustration, we use data on the English genitive alternation and provide an R package and code for their implementation. KW - average predictive comparisons KW - classification trees KW - interpretable machine learning KW - predictive modeling KW - random forests ER - TY - CHAP TI - Language change In Maltese English : the influence of age and parental languages AU - Krug, Manfred AU - Sönning, Lukas PY - 2022 SP - 247-270 PB - Otto-Friedrich-Universität CY - Bamberg UR - https://fis.uni-bamberg.de/handle/uniba/55199 LA - eng AB - In this study, which is based on questionnaire data collected in 2013 from 430 Maltese informants, we ninvestigate ongoing language change in Maltese English. We concentrate on 63 pairs of lexical variants that are known to differ in usage between British English and American English (e.g. vacation vs. holiday). Overall, informants clearly tend towards BrE usage. Regardless of the statistical approach we adopt, our studies show consistently apparent-time trends towards a less exclusively British English usage in Malta, converging on a more globalized usage of lexical items, in particular among the youngest cohorts. This confirms trends reported for older Maltese English data (collected in 2008; see Krug 2015). While Age emerges as the most important factor in our data, lexical choices are also sensitive to the native languages of the informants’ parents. When the mother’s native language(s) includes English, the informants’ lexical choices are biased in the expected direction, figuring in an increase in Britishness of the informants. Informants whose parents’ L1 is neither English nor Maltese show the highest degree of linguistic globalization. Overall, the native language(s) of the mother appeared to be more influential than that of the father. KW - Maltese English KW - language change KW - questionnaire data ER - TY - CHAP TI - Comparing Standard Reference Corpora and Google Books Ngrams : Strengths, Limitations and Synergies in the Contrastive Study of Variable h- in British and American English AU - Sönning, Lukas AU - Schlüter, Julia PY - 2022 T2 - Data and Methods in Corpus Linguistics : Comparative Approaches SP - 17-45 PB - Cambridge University Press CY - Cambridge ; New York SN - 9781108499644 DO - 10.1017/9781108589314.002 UR - https://fis.uni-bamberg.de/handle/uniba/54064 LA - eng AB - This chapter is based on two standard reference corpora, the British National Corpus and the Corpus of Contemporary American English, as opposed to the multi-billion-word database of Google Books Ngrams, which has, despite its allure, not been used in many systematic linguistic studies so far. Focusing on indefinite article allomorphy (a vs an) as an orthographic cue to the phonological strength of ‹h›-onsets in British and American English, the size advantage of the Ngrams database expectedly plays out in larger type and token counts, more stable estimates and fewer distortions due to data sparsity. However, as metadata are extremely limited (to year and variety), a fully accountable analysis is not feasible. The case study illustrates how richly annotated corpora can shed light on potential disturbances arising from two sources: genre differences and between-author variability. A sensitivity analysis offers some degree of reassurance when extending the analysis to the Ngrams database. In this way, the authors demonstrate that the strengths and limitations of corpora and big data resources can, with due caution, be counterbalanced to answer questions of linguistic interest. KW - Google Books Ngrams KW - big data KW - metadata KW - type frequency KW - token frequency KW - hierarchical data structure KW - corpus comparability KW - data quality ER - TY - JOUR TI - Comparing Study Designs and Down-Sampling Strategies in Corpus Analysis : The Importance of Speaker Metadata in the BNCs of 1994 and 2014 AU - Sönning, Lukas AU - Krug, Manfred PY - 2022 SP - 127-160 PB - Cambridge University Press CY - Cambridge ; New York DO - 10.1017/9781108589314.006 UR - https://fis.uni-bamberg.de/handle/uniba/54062 LA - eng AB - This chapter throws into relief the importance of the link between corpus hits and their sources (i.e. texts or speakers) by comparing study designs uninformed and informed by such metadata. The authors’ argument draws attention to the consequences of uneven distributions of observations across the basic text units of a corpus. In the case study (a distributional analysis of the use of actually in the BNCs of 1994 and 2014), it is demonstrated that disproportionate word counts contributed by individual sources (in this case, speakers) will distort estimates for relevant subsections of corpus data (in this case, demographic groups defined by age and gender) if the analysis assigns the same weight to every observation. The proposed solution is to factor in the text (or speaker) level, but this hinges on the availability of the relevant metadata. Moreover, insights into the hierarchical structure of corpus data are shown to benefit the design stage of a study. Thus, if manual post-processing steps preclude an exhaustive analysis of corpus hits, insights into the organization of data points can direct down-sampling strategies to generate a statistically efficient subset of tokens. KW - Corpus Analysis ER - TY - JOUR TI - The replication crisis, scientific revolutions, and linguistics AU - Sönning, Lukas AU - Werner, Valentin PY - 2021 SP - 1179-1206 PB - Otto-Friedrich-Universität CY - Bamberg DO - 10.20378/irb-51570 UR - https://fis.uni-bamberg.de/handle/uniba/51570 LA - eng AB - This introductory article sets the scene for the special issue of Linguistics entitled The replication crisis: implications for linguistics. We start out by sketching key issues surrounding the replication crisis, a state of methodological unrest that has caused turbulence across quantitative disciplines since the early 2000s. Given the trend throughout the language sciences towards a greater reliance on empirical methods, we argue that the linguistic community should be perceptive to this broader discourse and foster an open and lively discussion culture. To this end, the contributions to this special issue engage with methodological aspects that are at the core of this debate, taking an explicitly linguistic point of view. After an overview of the individual articles, we offer a perspective of the replication crisis from an angle inspired by the philosophy of science. We turn to Kuhn ([1962] 1996) to understand today’s methodological controversies in the context of his cyclical model of scientific progress. This allows us to interpret the replication crisis as a transitory stage, with revolutionary forces propagating a paradigm shift. We attempt to capture and contrast essential features of the status quo and what may eventually emerge as the new methodological paradigm in quantitative linguistics. KW - methodology KW - paradigm shift KW - philosophy of science KW - quantitative linguistics KW - replication crisis KW - scientific revolution ER - TY - JOUR TI - The replication crisis, scientific revolutions, and linguistics AU - Sönning, Lukas AU - Werner, Valentin PY - 2021 JO - Linguistics : an interdisciplinary journal of the language sciences VL - 59 IS - 5 SP - 1179-1206 PB - Mouton de Gruyter CY - Berlin SN - 0024-3949 DO - 10.1515/ling-2019-0045 UR - https://fis.uni-bamberg.de/handle/uniba/51673 LA - eng AB - This introductory article sets the scene for the special issue of Linguistics entitled The replication crisis: implications for linguistics. We start out by sketching key issues surrounding the replication crisis, a state of methodological unrest that has caused turbulence across quantitative disciplines since the early 2000s. Given the trend throughout the language sciences towards a greater reliance on empirical methods, we argue that the linguistic community should be perceptive to this broader discourse and foster an open and lively discussion culture. To this end, the contributions to this special issue engage with methodological aspects that are at the core of this debate, taking an explicitly linguistic point of view. After an overview of the individual articles, we offer a perspective of the replication crisis from an angle inspired by the philosophy of science. We turn to Kuhn ([1962] 1996) to understand today’s methodological controversies in the context of his cyclical model of scientific progress. This allows us to interpret the replication crisis as a transitory stage, with revolutionary forces propagating a paradigm shift. We attempt to capture and contrast essential features of the status quo and what may eventually emerge as the new methodological paradigm in quantitative linguistics. KW - methodology KW - paradigm shift KW - philosophy of science KW - quantitative linguistics KW - replication crisis KW - scientific revolution ER - TY - GEN TI - Linguistics : an interdisciplinary journal of the language sciences ED - Sönning, Lukas ED - Werner, Valentin PY - 2021 VL - 59 IS - 5, Special Issue: The replication crisis: Implications for linguistics SP - 1179-1358 PB - Mouton de Gruyter CY - Berlin SN - 0024-3949 UR - https://fis.uni-bamberg.de/handle/uniba/51571 LA - eng ER - TY - THES TI - Phonological variation in German Learner English AU - Sönning, Lukas PY - 2020 PB - Otto-Friedrich-Universität CY - Bamberg DO - 10.20378/irb-49135 UR - https://fis.uni-bamberg.de/handle/uniba/49135 LA - eng AB - This study is concerned with linguistic constraints underlying the acquisition of a second language (L2) phonology. It pursues two aims: The first is to offer a survey of theoretical work on L2 phonology with a focus on the scope and predictive adequacy of individual frameworks. A set of 20 contributions is evaluated in terms of their domain of application and the specificity of their predictions. The second aim is to make empirical contributions to the study of phonological variation in German Learner English by exploring theoretically motivated constraints across a diverse range of phonological structures and proficiency levels. Theory-derived hypotheses are confronted with data from a cross-sectional study on various segmental features in the pronunciation of 62 German learners of English (including 27 native speakers as a baseline of comparison). The study relies on acoustic and auditory analyses to shed light on several notorious structures, including final obstruent (de)voicing, dental fricatives, the labiodental fricative /v/, the labio-velar glide /w/, English /r/, ‘clear’ and ‘dark’ laterals, and the TRAP-DRESS contrast. KW - L2 acquisition KW - Phonetics and phonology KW - German Learner English KW - L1 German KW - Acoustic analysis KW - Auditory analysis KW - TRAP vowel KW - DRESS vowel KW - Laterals KW - Clear and dark /l/ KW - Second language acquisition KW - Rhoticity KW - English /r/ KW - Labiovelar glide /w/ KW - Labiodental fricative /v/ KW - Hypercorrection KW - Over-correction KW - Dental fricatives KW - Final voiced obstruents KW - Final devoicing KW - Preceding vowel duration KW - Vowel length variation ER - TY - CHAP TI - Language Change In Maltese English : the Influence Of Age And Parental Languages AU - Krug, Manfred AU - Sönning, Lukas PY - 2018 T2 - The languages of Malta SP - 247-270 PB - Language Science Press CY - Berlin SN - 978-3-96110-070-5 DO - 10.5281/zenodo.1181801 UR - https://fis.uni-bamberg.de/handle/uniba/55261 LA - eng AB - In this study, which is based on questionnaire data collected in 2013 from 430 Maltese informants, we ninvestigate ongoing language change in Maltese English. We concentrate on 63 pairs of lexical variants that are known to differ in usage between British English and American English (e.g. vacation vs. holiday). Overall, informants clearly tend towards BrE usage. Regardless of the statistical approach we adopt, our studies show consistently apparent-time trends towards a less exclusively British English usage in Malta, converging on a more globalized usage of lexical items, in particular among the youngest cohorts. This confirms trends reported for older Maltese English data (collected in 2008; see Krug 2015). While Age emerges as the most important factor in our data, lexical choices are also sensitive to the native languages of the informants’ parents. When the mother’s native language(s) includes English, the informants’ lexical choices are biased in the expected direction, figuring in an increase in Britishness of the informants. Informants whose parents’ L1 is neither English nor Maltese show the highest degree of linguistic globalization. Overall, the native language(s) of the mother appeared to be more influential than that of the father. KW - Maltese English KW - language change KW - questionnaire data ER - TY - CHAP TI - The dot plot : A graphical tool for data analysis and presentation AU - Sönning, Lukas PY - 2018 PB - opus CY - Bamberg DO - 10.20378/irbo-51101 UR - https://fis.uni-bamberg.de/handle/uniba/43180 LA - eng ER - TY - CHAP TI - Introduction AU - Christ, Hanna AU - Klenovšak, Daniel AU - Sönning, Lukas AU - Werner, Valentin PY - 2018 PB - opus CY - Bamberg DO - 10.20378/irbo-51100 UR - https://fis.uni-bamberg.de/handle/uniba/43172 LA - eng ER - TY - BOOK TI - A blend of MaLT : Selected contributions from the Methods and Linguistic Theories Symposium 2015 AU - Christ, Hanna AU - Klenovšak, Daniel AU - Sönning, Lukas AU - Werner, Valentin AU - Ziem, Alexander AU - Staicov, Adina AU - Rüdiger, Sofia AU - Häussler, Jana AU - Juzek, Tom AU - Naccarato, Chiara AU - Zaretsky, Eugen AU - Lange, Benjamin P. ED - Christ, Hanna ED - Klenovšak, Daniel ED - Sönning, Lukas ED - Werner, Valentin PY - 2016 PB - University of Bamberg Press CY - Bamberg T3 - Bamberger Beiträge zur Linguistik SN - 978-3-86309-457-7 DO - 10.20378/irb-41360 UR - https://fis.uni-bamberg.de/handle/uniba/41360 LA - eng AB - Over the past few decades, linguistic theorizing has benefited from an increasing trend towards empirical methodologies across all disciplines. Methodological know-how – both productive and receptive – has thus become one of the key qualifications for researchers. The empirical turn in linguistics has gone hand in hand with a considerable diversification of research methods. This diversity, which has come to be seen as a strength of linguistics as a field, has also benefited linguistic theory building. The present volume contains selected contributions from the 2015 Methods and Linguistic Theories (MaLT) symposium that address the aforementioned issues from an empirical and/or theoretical perspective. They can be seen as the essence of what MaLT was about, and illustrate the range of topics covered as well as the various concerns and approaches that featured during the event. KW - Linguistic methods KW - Linguistic theory KW - Empirical research KW - Methods and theories symposium KW - Approaches in linguistics ER - TY - CPAPER TI - Developmental patterns in the reduction of unstressed vowel by German learners of English AU - Sönning, Lukas ED - Veselovská, Ludmila ED - Janebová, Markéta PY - 2014 T2 - Complex visibles out there : proceedings of the Olomouc Linguistics Colloquium 2014: Language Use and Linguistic Structure SP - 765-778 PB - Palacký University CY - Olomouc T3 - Olomuc modern language series SN - 978-80-244-4384-3 UR - https://fis.uni-bamberg.de/handle/uniba/41262 LA - eng AB - This study investigates vowel reduction in unstressed syllables in German Learner English, comparing learners at three different proficiency levels with native speakers of Standard Southern British English. Two acoustic properties of unstressed vowels were measured: duration and formant structure. Based on Major’s (2001) Ontogeny Phylogeny Model of second language acquisition, a U-shaped developmental pattern was predicted. The results follow the predicted pattern and seem to support the chronological corollary of Major’s model: Initial transfer of L1 stress-timing properties appears to facilitate vowel reduction. The subsequent influence of universal structures surfaces in an overarticulation of unstressed vowels. The final increase of target language structures leads to a higher degree of vowel reduction. KW - vowel reduction; acoustic study; SLA; L1 German; L2 English ER - TY - JOUR TI - Unstressed Vowels in German Learner English: An Instrumental Study AU - Sönning, Lukas PY - 2014 JO - Research in language : RL VL - 12 IS - 2 SP - 163-173 PB - Lodz University Press CY - Łódź SN - 2083-4616 DO - 10.2478/rela-2014-0001 UR - https://fis.uni-bamberg.de/handle/uniba/41263 LA - eng AB - This study investigates the production of vowels in unstressed syllables by advanced German learners of English in comparison with native speakers of Standard Southern British English. Two acoustic properties were measured: duration and formant structure. The results indicate that duration of unstressed vowels is similar in the two groups, though there is some variation depending on the phonetic context. In terms of formant structure, learners produce slightly higher F1 and considerably lower F2, the difference in F2 being statistically significant for each learner. Formant values varied as a function of context and orthographic representation of the vowel. KW - vowel reduction KW - acoustic analysis KW - SLA KW - L1 German KW - L2 English ER - TY - CPAPER TI - Scrabble yourself to success: Methods in teaching transcription AU - Sönning, Lukas PY - 2013 T2 - Proceedings of the Phonetics Teaching and Learning Conference ; UCL, London, 8–10 August 2013 SP - 83-86 CY - London SN - 978-0-9926394-0-2 UR - https://fis.uni-bamberg.de/handle/uniba/54083 LA - eng AB - This paper summarizes the results of a study that tested and evaluated two methods for teaching phonemic transcription to German EFL students at university level. The research design included a control group and two training groups receiving treatment with different methods: one based on the transcription of auditive stimuli, the other based on a phonemic adaption of the board game Scrabble®. The different training methods had significant effects on the improvement of transcription skills. A comparison of the training groups did not yield significant differences – the descriptive statistics, however, suggest that phoneme scrabble cannot be recommended as a uniform teaching method. KW - phonemic transcription KW - evaluation of teaching methods KW - experimental study ER - TY - CPAPER TI - Scrabble yourself to success: Methods in teaching transcription AU - Sönning, Lukas PY - 2013 SP - 83-86 PB - Otto-Friedrich-Universität CY - Bamberg UR - https://fis.uni-bamberg.de/handle/uniba/2422 LA - eng AB - This paper summarizes the results of a study that tested and evaluated two methods for teaching phonemic transcription to German EFL students at university level. The research design included a control group and two training groups receiving treatment with different methods: one based on the transcription of auditive stimuli, the other based on a phonemic adaption of the board game Scrabble®. The different training methods had significant effects on the improvement of transcription skills. A comparison of the training groups did not yield significant differences – the descriptive statistics, however, suggest that phoneme scrabble cannot be recommended as a uniform teaching method. KW - phonemic transcription KW - evaluation of teaching methods KW - experimental study ER -