| { |
| "paper_id": "2022", |
| "header": { |
| "generated_with": "S2ORC 1.0.0", |
| "date_generated": "2023-01-19T12:11:38.718088Z" |
| }, |
| "title": "Deconstructing destruction: A Cognitive Linguistics perspective on a computational analysis of diachronic change", |
| "authors": [ |
| { |
| "first": "Karlien", |
| "middle": [], |
| "last": "Franco", |
| "suffix": "", |
| "affiliation": { |
| "laboratory": "", |
| "institution": "KU Leuven & Dutch Language Institute (INT)", |
| "location": {} |
| }, |
| "email": "karlien.franco@kuleuven.be" |
| }, |
| { |
| "first": "Kris", |
| "middle": [], |
| "last": "Heylen", |
| "suffix": "", |
| "affiliation": { |
| "laboratory": "", |
| "institution": "KU Leuven & Dutch Language Institute (INT)", |
| "location": {} |
| }, |
| "email": "kris.heylen@ivdnt.org" |
| }, |
| { |
| "first": "Mariana", |
| "middle": [], |
| "last": "Montes", |
| "suffix": "", |
| "affiliation": { |
| "laboratory": "", |
| "institution": "KU Leuven & Dutch Language Institute (INT)", |
| "location": {} |
| }, |
| "email": "mariana.montes@kuleuven.be" |
| }, |
| { |
| "first": "K", |
| "middle": [ |
| "U" |
| ], |
| "last": "Leuven", |
| "suffix": "", |
| "affiliation": { |
| "laboratory": "", |
| "institution": "KU Leuven & Dutch Language Institute (INT)", |
| "location": {} |
| }, |
| "email": "" |
| } |
| ], |
| "year": "", |
| "venue": null, |
| "identifiers": {}, |
| "abstract": "In this paper, we aim to introduce a Cognitive Linguistics perspective into a computational analysis of near-synonyms. We focus on a single set of Dutch near-synonyms, vernielen and vernietigen, roughly translated as 'to destroy', replicating the analysis from Geeraerts (1997) with distributional models. Our analysis, which tracks the meaning of both words in a corpus of 16th-20th century prose data, shows that both lexical items have undergone semantic change, led by differences in their prototypical semantic core.", |
| "pdf_parse": { |
| "paper_id": "2022", |
| "_pdf_hash": "", |
| "abstract": [ |
| { |
| "text": "In this paper, we aim to introduce a Cognitive Linguistics perspective into a computational analysis of near-synonyms. We focus on a single set of Dutch near-synonyms, vernielen and vernietigen, roughly translated as 'to destroy', replicating the analysis from Geeraerts (1997) with distributional models. Our analysis, which tracks the meaning of both words in a corpus of 16th-20th century prose data, shows that both lexical items have undergone semantic change, led by differences in their prototypical semantic core.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Abstract", |
| "sec_num": null |
| } |
| ], |
| "body_text": [ |
| { |
| "text": "This paper aims to stimulate further convergence between Cognitive Linguistics approaches to language change and computational methods for semantic change. Cognitive Linguistics is a contemporary linguistic paradigm that assumes that linguistic knowledge is rooted in general cognitive capabilities, that language is shaped by usage and that meaning entails conceptualization (Dabrowska and Divjak, 2015) . It pays particular attention to the interaction between semasiological change, whereby a word's meaning changes over time, and onomasiological change, whereby the semantic configuration of a set of (near-)synonyms is reorganized over time. In this paper we focus on a single set of Dutch near-synonyms, vernielen and vernietigen, roughly translated as 'to destroy'. We replicate the analysis of these verbs from Geeraerts (1997) with distributional models. The analysis, which tracks the meaning of both words in a corpus of 16th-20th century prose data, shows that both lexical items have undergone semantic change led by differences in their prototypical semantic core, as predicted by Geeraerts' work. 2 Semasiology and onomasiology in Cognitive Linguistics", |
| "cite_spans": [ |
| { |
| "start": 376, |
| "end": 404, |
| "text": "(Dabrowska and Divjak, 2015)", |
| "ref_id": "BIBREF3" |
| }, |
| { |
| "start": 819, |
| "end": 835, |
| "text": "Geeraerts (1997)", |
| "ref_id": "BIBREF8" |
| }, |
| { |
| "start": 1095, |
| "end": 1111, |
| "text": "Geeraerts' work.", |
| "ref_id": null |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "Following from \"The cognitive commitment\", which entails that the description of human language should be congruent with what is known about cognition within and outside of linguistics (Lakoff, 1990) , Cognitive Linguistics aims to be a psychologically plausible model. Language is, thus, primarily studied as a means to communicate -a way to convey and process meaning. Furthermore, a maximalist, non-reductionist perspective on linguistic knowledge is assumed. Language systems are considered to be \"reflections of general conceptual organization, categorization principles, processing mechanisms, and experiential and environmental influences\" (Geeraerts and Cuyckens, 2007, 3) . The movement, therefore, places a large emphasis on meaning. Geeraerts et al. (1994) examine the structure of lexical variation in the use of clothing terminology in Dutch. Crucially, it is the first study to systematically emphasize the importance of two distinctions. On the one hand, it shows that in order to obtain a full picture of the structure of lexical variation, semasiological research should be complemented with an onomasiological approach. The semasiological perspective examines the range of applications of a particular expression. Semasiology is, for this reason, often defined as research into the meaning of a particular item: given a particular word or expression, what are the referents to which the word applies? In the case of the Dutch word monitor, for instance, a semasiological analysis would reveal that it can refer both to a SUPERVI-SOR, and to a COMPUTER SCREEN (see Heylen et al., 2015) . The onomasiological perspective investigates naming rather than meaning. An onomasiological approach, thus, starts from a particular (type of) referent or concept and determines which names exist or can be used to refer to the referent. For instance, an onomasiological analysis of the concept COMPUTER SCREEN in Dutch would reveal that both monitor and computerscherm can be used to express this concept.", |
| "cite_spans": [ |
| { |
| "start": 185, |
| "end": 199, |
| "text": "(Lakoff, 1990)", |
| "ref_id": "BIBREF14" |
| }, |
| { |
| "start": 647, |
| "end": 680, |
| "text": "(Geeraerts and Cuyckens, 2007, 3)", |
| "ref_id": null |
| }, |
| { |
| "start": 744, |
| "end": 767, |
| "text": "Geeraerts et al. (1994)", |
| "ref_id": "BIBREF10" |
| }, |
| { |
| "start": 1582, |
| "end": 1602, |
| "text": "Heylen et al., 2015)", |
| "ref_id": "BIBREF12" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "On the other hand, Geeraerts et al. (1994) was the first study to make the importance of the interaction between four different types of lexical variation for the structure of the lexicon explicit. First, it examines semasiological variation, the situation where a single lexical item can refer to more than one referent. For example, the lexical item pants can both be used to refer to a TWO-LEGGED TYPE OF OUTER GARMENT (IN GENERAL), but also to a more specific referent, viz. MEN'S UN-DERWEAR. The second and third types of lexical variation that are distinguished concern two varieties of onomasiological variation: conceptual onomasiological variation and formal onomasiological variation. Conceptual onomasiological variation concerns the situation where \"a referent or type of referent may be named by means of various conceptually distinct lexical categories\" (Geeraerts et al., 1994, 3-4) . For example, to refer to a pair of BLUE JEANS, a language user can either choose to select a lexical item belonging to the concept BLUE JEANS and use a word like jeans or blue jeans, or (s)he can conceptualize the referent as a type of PANTS, a superordinate concept, and call the denotatum trousers or pants. Formal onomasiological variation occurs when a choice has to be made between different synonymous expressions for a referent. In the blue jeans example, this would involve determining the relative frequency of the terms jeans versus blue jeans versus trousers versus pants. Finally, it shows how contextual variation can be at play both at the semasiological and onomasiological level. Contextual variation (also called speaker and situation related variation) is broadly defined: it includes both the relatively stable lectal properties of the interlocutors involved (like their gender or their nationality), but also transient situation-related features, like the register of the speech event (Geeraerts et al., 2010, 8) . For the (onomasiological) blue jeans example, for instance, contextual variation may take the form of determining whether older people are more likely to refer to the concept as blue jeans.", |
| "cite_spans": [ |
| { |
| "start": 19, |
| "end": 42, |
| "text": "Geeraerts et al. (1994)", |
| "ref_id": "BIBREF10" |
| }, |
| { |
| "start": 868, |
| "end": 897, |
| "text": "(Geeraerts et al., 1994, 3-4)", |
| "ref_id": null |
| }, |
| { |
| "start": 1905, |
| "end": 1932, |
| "text": "(Geeraerts et al., 2010, 8)", |
| "ref_id": null |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "The program laid out in Geeraerts et al. (1994) was applied to diachronic change in Geeraerts (1997) . An important finding of this work is that semasiological and onomasiological variation and change are not independent of each other: semasiological changes also affect the onomasiological structure of a language.", |
| "cite_spans": [ |
| { |
| "start": 24, |
| "end": 47, |
| "text": "Geeraerts et al. (1994)", |
| "ref_id": "BIBREF10" |
| }, |
| { |
| "start": 84, |
| "end": 100, |
| "text": "Geeraerts (1997)", |
| "ref_id": "BIBREF8" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "A classical analysis of the verbs meaning 'to destroy' in Dutch is presented in (Geeraerts, 1997 (Geeraerts, , 1985 (Geeraerts, , 1988 . In these papers, Geeraerts analyzes vernielen and vernietigen in 19th century data taken from the citations corpus for largest historical dictionary of Dutch, the Woordenboek der Nederlandsche taal 'Dictionary of the Dutch Language'. Etymologically, the near-synonyms do not have the same root. Vernielen is a verb formed with a verbalizing prefix ver-, and niel, an obsolete Dutch adjective that roughly translates to 'down to the ground'. The literal meaning of the verb vernielen is then 'to throw down to the ground, to tear down'. Vernietigen, in contrast, is based on a verbalizing prefix ver-, with the adjective nietig (which itself comes from niet 'not, nothing' + a suffix -ig). The meaning of vernietigen is then 'to annihilate, to bring to naught'.", |
| "cite_spans": [ |
| { |
| "start": 80, |
| "end": 96, |
| "text": "(Geeraerts, 1997", |
| "ref_id": "BIBREF8" |
| }, |
| { |
| "start": 97, |
| "end": 115, |
| "text": "(Geeraerts, , 1985", |
| "ref_id": "BIBREF6" |
| }, |
| { |
| "start": 116, |
| "end": 134, |
| "text": "(Geeraerts, , 1988", |
| "ref_id": "BIBREF7" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Destructive verbs in Dutch", |
| "sec_num": "3" |
| }, |
| { |
| "text": "Despite these divergent sources, Geeraerts shows that the near-synonyms can be used in similar contexts by the same author in the 19th century. For instance, in examples 1 and 2 (adapted from Geeraerts, 1988, 30-31) , vernielen and vernietigen occur in the context of a material artefact (a part of a building) being destroyed. He discusses many more examples that clearly show that the verbs are interchangeable in 19th century Dutch. Geeraerts (1997, 191-192) between the abstract (collective army) and personal (an individual soldier) contexts. While the verbs are found in similar contexts, a crucial point of Geeraerts' work is that the prototypical cores of the verbs differ. More specifically, vernielen prototypically occurs with concrete uses, such as destroying parts of buildings. Vernietigen prototypically occurs in abstract contexts such as the complete annihilation of existing situations or plans. Geeraerts also notes that, while both verbs can occur with instances of partial or complete destruction, vernielen is prototypically used in the partial destruction sense (e.g. when a building is destroyed by a fire, parts of the structure and ashes from the fire remain), whereas vernietigen often implies complete annihilation to naught (e.g. when a plan is destroyed, nothing remains). In addition, the difference between the concrete and abstract uses is also visible in the context of the 'destruction' of people: while vernielen occurs more with the more concrete sense of to kill someone, vernietigen is more often found in the more abstract contexts where someone's physical or mental health is affected.", |
| "cite_spans": [ |
| { |
| "start": 192, |
| "end": 215, |
| "text": "Geeraerts, 1988, 30-31)", |
| "ref_id": null |
| }, |
| { |
| "start": 436, |
| "end": 461, |
| "text": "Geeraerts (1997, 191-192)", |
| "ref_id": null |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Destructive verbs in Dutch", |
| "sec_num": "3" |
| }, |
| { |
| "text": "In Montes et al. (2021) , the near-synonyms were analyzed in synchronic contemporary newspaper data with distributional models. The analysis showed that, since the 19th century, the prototypical cores of vernielen and vernietigen have become even stronger and the verbs are no longer easily interchangeable in every context. A highly prototypical context for vernielen in the 21st century data is the destruction of (parts of) buildings by fire and vernietigen no longer occurs in this context. In contrast, for vernietigen, the cancellation of decisions or ideas by a governmental body makes up a large portion of the tokens in the corpus and vernielen is no longer possible there. Both variants seem to have retreated to their prototypical core. In the periphery of the semantics of the verbs, some new uses have come into existence (e.g. to destroy livestock, probably as a result of the industrialization of the food industry and the regulations installed by government to keep the industry safe for consumption).", |
| "cite_spans": [ |
| { |
| "start": 3, |
| "end": 23, |
| "text": "Montes et al. (2021)", |
| "ref_id": "BIBREF17" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Destructive verbs in Dutch", |
| "sec_num": "3" |
| }, |
| { |
| "text": "The aim of this paper is to track the diachronic change in the nearly synonymous pair vernielen and vernietigen throughout time in Dutch using distributional models. Based on the results in Geeraerts (1997 Geeraerts ( , 1985 Geeraerts ( , 1988 for 19th century dictionary attestations, and Montes et al. (2021) for 21st century corpus data, we expect to find in our study of continuous diachronic corpus data from the 16th to 20th century, that the overlap or interchangeability between the verbs reduces over time and that the verbs will retreat to their prototypical cores more and more over time. This finding would confirm that semasiological and onomasiological change interact and that these types of changes can be retrieved automatically from diachronic corpus data. Methodologically, we investigate the usefulness of distributional models for diachronic changes in a pair of near-synonyms.", |
| "cite_spans": [ |
| { |
| "start": 190, |
| "end": 205, |
| "text": "Geeraerts (1997", |
| "ref_id": "BIBREF8" |
| }, |
| { |
| "start": 206, |
| "end": 224, |
| "text": "Geeraerts ( , 1985", |
| "ref_id": "BIBREF6" |
| }, |
| { |
| "start": 225, |
| "end": 243, |
| "text": "Geeraerts ( , 1988", |
| "ref_id": "BIBREF7" |
| }, |
| { |
| "start": 290, |
| "end": 310, |
| "text": "Montes et al. (2021)", |
| "ref_id": "BIBREF17" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Destructive verbs in Dutch", |
| "sec_num": "3" |
| }, |
| { |
| "text": "In the analysis, we use a corpus of prose texts from DBNL, the Digitale Bibliotheek voor de Nederlandse Letteren 'digital library for Dutch languages and literature'. Some information about the corpus can be found in Depuydt and Brugman (2019) , though the corpus is not publicly available at this time. We specifically extracted all corpus texts tagged as prose in the metadata from the 16th, 17th, 18th, 19th and 20th century. Due to data sparseness, we combine the subcorpora for the 16th and 17th century in the analysis.", |
| "cite_spans": [ |
| { |
| "start": 217, |
| "end": 243, |
| "text": "Depuydt and Brugman (2019)", |
| "ref_id": "BIBREF5" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Data and methods", |
| "sec_num": "4" |
| }, |
| { |
| "text": "As no high-quality lemmatizers or PoS-taggers are as of yet available for historical Dutch, the only preprocessing we applied to the corpus was to transform the entire corpus to lower case and to automatically indicate sentence boundaries using the pretrained nltk sentence tokenizer (Bird and Loper, 2004) .", |
| "cite_spans": [ |
| { |
| "start": 284, |
| "end": 306, |
| "text": "(Bird and Loper, 2004)", |
| "ref_id": "BIBREF0" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Data and methods", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Next, we extracted all tokens for vernielen and vernietigen (including inflected forms and spelling variants) from the four subcorpora and took a random sample of N = 400 tokens for each subcorpus.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Data and methods", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Then, we constructed a single vector space model for the tokens in each subcorpus. The models that we built are based on the procedure outlined in Sch\u00fctze (1998) . More specifically, we construct a vector representation for each token in the corpus, using the words in the context of the tokens to construct the vectors (first order vectors). We supplement this information by also constructing a vector for each of the relevant context words (second order vectors). An association strength measure is used rather than raw co-occurrence frequencies, namely positive pointwise mutual information or PPMI (Church and Hanks, 1989; Bullinaria and Levy, 2007; Kiela and Clark, 2014) . This procedure has the advantage that both the context words for a token, as well as their semantic similarity with other context words, is taken into account (see De Pascale, 2019 and Montes, 2021) . This method represents an example of a context-counting distributional model, which we opt for here (rather than for a context-predicting method) because we want to keep the results as comparable as possible to the results obtained in Montes et al. (2021) on contemporary data. One line of future research is to replicate the results obtained here with diachronic contextualized word embeddings when they become available for Dutch.", |
| "cite_spans": [ |
| { |
| "start": 147, |
| "end": 161, |
| "text": "Sch\u00fctze (1998)", |
| "ref_id": "BIBREF20" |
| }, |
| { |
| "start": 603, |
| "end": 627, |
| "text": "(Church and Hanks, 1989;", |
| "ref_id": "BIBREF2" |
| }, |
| { |
| "start": 628, |
| "end": 654, |
| "text": "Bullinaria and Levy, 2007;", |
| "ref_id": "BIBREF1" |
| }, |
| { |
| "start": 655, |
| "end": 677, |
| "text": "Kiela and Clark, 2014)", |
| "ref_id": "BIBREF13" |
| }, |
| { |
| "start": 865, |
| "end": 878, |
| "text": "Montes, 2021)", |
| "ref_id": "BIBREF17" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Data and methods", |
| "sec_num": "4" |
| }, |
| { |
| "text": "The parameters that we used are largely based on the best model found in Montes et al. (2021) for the analysis of vernielen and vernietigen in 21st century newspaper data. However, since the diachronic corpus that we use is not lemmatized or PoS-tagged, the vectors represent word forms rather than lemmas and we do not use part-of-speech filters. Additionally, we decided to decrease the window size from 15 to 10 words to the left and right of the target token for the first-order context words because preliminary analyses revealed that in models with a broader window, too many irrelevant or noisy context features were included in the analysis (also due to the fact that PoS filters cannot be applied). The parameters settings that we used are the following:", |
| "cite_spans": [ |
| { |
| "start": 73, |
| "end": 93, |
| "text": "Montes et al. (2021)", |
| "ref_id": "BIBREF17" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Data and methods", |
| "sec_num": "4" |
| }, |
| { |
| "text": "\u2022 Bag-of-words model with a window size of 10 words to the left and right of each token.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Data and methods", |
| "sec_num": "4" |
| }, |
| { |
| "text": "\u2022 First-order context words: all wordforms [w+] with a frequency of at least 10 in the subcorpus. First-order contextwords are subsequently filtered by their PPMI value with the target token: only words with PPMI > 2 are considered.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Data and methods", |
| "sec_num": "4" |
| }, |
| { |
| "text": "\u2022 Second order context words: 5000 most frequent wordforms [w+] in the subcorpus, excluding the first 100 wordforms, as these are usually function words rather than content words and therefore do not contribute a lot of semantic information.", |
| "cite_spans": [ |
| { |
| "start": 59, |
| "end": 63, |
| "text": "[w+]", |
| "ref_id": null |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Data and methods", |
| "sec_num": "4" |
| }, |
| { |
| "text": "The models were constructed with the nephosem Python library (QLVL, 2021). The result of this procedure is a token-by-context matrix with 5000 dimensions, where the dimensions represent second order context vectors, i.e. the vectors of the context words around each token. Since for some tokens no relevant context words are found with our parameter settings, these tokens are excluded from the remainder of the analysis. For clustering and visualization, we transformed this matrix into a square distance matrix by computing the cosine vector of each pair of token-level vectors, without further dimensionality reduction beforehand. Thus, this final matrix describes the dissimilarity between the vector of each token and all other token vectors in the subcorpus. As a next step, we submitted each model to a clustering procedure in R (R Core Team, 2020). We used hierarchical clustering (Ward method), distinguishing four clusters, following the procedure in Montes et al. 2021for maximal comparability. Finally, we analyzed each cluster per subcorpus basing ourselves on a procedure outlined in Montes (2021) that is available in the Python library sema-sioFlow (Montes, 2022) . The procedure consists of a number of steps. First, the relevant context words, on which the token vectors are based, are extracted from the model data. Then, after each token is assigned to a cluster, we calculate how often a specific context word occurs within a particular cluster and outside of the cluster. Using this information, we can calculate which context words have an exceptionally high frequency in each cluster and therefore represent the semantics of each cluster well. In the analysis, we will only consider context words for which at least 50% of their occurrences are within the cluster of interest. Figure 1 (see Appendix A) shows the visualisations of the models, with one panel per subcorpus. Plot symbols show the variants (vernielen versus vernietigen) and colours indicate the clusters. The figure shows that over time, vernielen and vernietigen are distinguished more clearly by the models.", |
| "cite_spans": [ |
| { |
| "start": 1165, |
| "end": 1179, |
| "text": "(Montes, 2022)", |
| "ref_id": null |
| } |
| ], |
| "ref_spans": [ |
| { |
| "start": 1801, |
| "end": 1809, |
| "text": "Figure 1", |
| "ref_id": null |
| } |
| ], |
| "eq_spans": [], |
| "section": "Data and methods", |
| "sec_num": "4" |
| }, |
| { |
| "text": "While in the 16th/17th century, there is still quite some overlap between the variants, indicating that they are still interchangeable, in the 18th century vernielen mostly occurs at the bottom left of the plot and vernietigen at the top right. In the 19th century, vernielen is found in the left side of the plot and vernietigen mostly in the bottom right. By the 20th century, the variant vernielen had decreased dramatically in frequency and vernietigen takes up most of the figure. Only one cluster remains where vernielen is dominant: cluster 4 at the bottom right.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Results", |
| "sec_num": "5" |
| }, |
| { |
| "text": "Tables 2-5 show an overview of the most important context words per period and per cluster, obtained with the procedure outlined above. Only context words with a frequency of more than 2 are shown, to avoid that infrequent words get too much weight in the interpretation. The first column also shows an interpretation of each cluster. The final columns indicate the relative and absolute proportion of each variant in the cluster.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Results", |
| "sec_num": "5" |
| }, |
| { |
| "text": "In the first subcorpus (16th-17th century, Table 2), there are three clusters where vernielen clearly is the major variant (clusters 2, 3 and 4). It occurs in contexts related to killing persons, a small cluster with natural objects (no context words with frequency > 2) and concrete objects like ships and cities. The first and largest cluster (N = 182 tokens) is still quite diverse and both vernielen and vernietigen are possible. Thus, in the 16th and 17th century, vernielen and vernietigen are still mostly interchangeable, although there are already a few contexts where vernielen is preferred.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Results", |
| "sec_num": "5" |
| }, |
| { |
| "text": "In the 18th century subcorpus (Table 3) , the variants start receding to their prototypical core more. There are two clusters where vernielen is more frequent and two clusters where vernietigen takes over. Following the hypotheses outlined above, vernielen mostly occurs with concrete objects like buildings (cluster 4, consisting of tokens related to fires destroying parts of buildings). In addition, it seems to occur in passive tokens (with werden 'became, was') where persons are destroyed: cluster 2 contains some war-related lexemes like vijand 'enemy', troepen 'troups', leger 'army' and some lexemes related to people, such as hunne 'their' and elkaar 'each other'. In contrast, vernietigen occurs in tokens with abstract objects (cluster 3) and it is also the most frequent variant in the first cluster, which does not show a clear semantic picture. In most clusters, except for 4, both variants are still possible. The context words in cluster 4 show that vernielen has by now become the most preferred variant for the destruction of (parts of) buildings (often by fire).", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 30, |
| "end": 39, |
| "text": "(Table 3)", |
| "ref_id": "TABREF5" |
| } |
| ], |
| "eq_spans": [], |
| "section": "Results", |
| "sec_num": "5" |
| }, |
| { |
| "text": "In the 19th century (Table 4) , which coincides with the data analyzed in Geeraerts (1997 Geeraerts ( , 1985 Geeraerts ( , 1988 , there are three clusters where one variant takes over, but also one cluster where the variants are interchangeable. More specifically, vernielen remains the most frequent variant in contexts of the destruction of (parts of) buildings (by fire, cluster 1). In contrast with the 18th century subcorpus, vernietigen has by now taken over contexts related to the destruction of persons, including armies (cluster 2). In this cluster zichzelf 'hisself/herself/themselves' is the most frequent context word. This frequent use of the reflexive pronoun may indicate that the patient role for vernietigen in the 19th century is often the subject itself, or that it at least plays a major role. Finally, vernietigen also still occurs the most with abstract lexemes such as vrijheid 'freedom' (cluster 4). Cluster 3 only has one important context word, waan 'delusion', and both variants are possible in this cluster. The interpretation is not as clear as for the other clusters.", |
| "cite_spans": [ |
| { |
| "start": 74, |
| "end": 89, |
| "text": "Geeraerts (1997", |
| "ref_id": "BIBREF8" |
| }, |
| { |
| "start": 90, |
| "end": 108, |
| "text": "Geeraerts ( , 1985", |
| "ref_id": "BIBREF6" |
| }, |
| { |
| "start": 109, |
| "end": 127, |
| "text": "Geeraerts ( , 1988", |
| "ref_id": "BIBREF7" |
| } |
| ], |
| "ref_spans": [ |
| { |
| "start": 20, |
| "end": 29, |
| "text": "(Table 4)", |
| "ref_id": "TABREF6" |
| } |
| ], |
| "eq_spans": [], |
| "section": "Results", |
| "sec_num": "5" |
| }, |
| { |
| "text": "Finally, in the subcorpus for the 20th century (Table 5) , vernietigen is much more frequent than vernielen. Only 108 tokens for vernielen occur in the complete 20th century subcorpus, but 446 are available for vernietigen. This may indicate that vernielen is on its way out, or that it is retreating to very specific contexts. The cluster analysis shows that there are still some clear contexts in which vernietigen is the preferred variant, but that in the 20th century data, not all clusters represent clear semantic differences. This may be partly related to the fact that vernielen has become very infrequent: most of the tokens that are modelled are for vernietigen and it is possible that the model distinguishes syntactic constructions rather than semantic contexts in which vernietigen can occur.", |
| "cite_spans": [], |
| "ref_spans": [ |
| { |
| "start": 47, |
| "end": 56, |
| "text": "(Table 5)", |
| "ref_id": "TABREF8" |
| } |
| ], |
| "eq_spans": [], |
| "section": "Results", |
| "sec_num": "5" |
| }, |
| { |
| "text": "First, vernietigen is the most frequent variant in cluster 1, which is a diverse cluster, with the most frequent context word related to complete destruction (geheel 'completely'), but also consisting of other types of lexical items such as abstract concepts. In cluster 2, vernietigen is the most frequent variant as well. This is a semantic cluster with many war-related lexical items, although it also contains other concrete objects. The context words in cluster 3, where vernietigen is also the most frequent variant, are mostly function words, such as adverbs and reflexive pronouns. This cluster is not determined by semantic similarity between the tokens, but rather by the type of construction the tokens occur in. The context words in cluster 4, where both vernielen and vernietigen are possible, are mostly related to (parts of) buildings. This is a clear change compared to the earlier data, where the destruction of parts of building correlated strongly with the use of vernielen. However, the context words in this cluster have quite a low frequency so likely not all tokens are related to the destruction of (parts of) buildings: perhaps vernielen has become so infrequent that even this prototypical use is not frequent enough anymore to be distinguished by the model and clustering procedure.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Results", |
| "sec_num": "5" |
| }, |
| { |
| "text": "The models for the four subcorpora show how the relationship between the near-synonyms vernielen and vernietigen has changed over time. Semasiologically, vernielen was the major variant in the 16th and 17th century, occurring in tokens related to the death of persons and concrete, natural objects. Over the course of the 18th century, it developed its prototypical meaning related to the destruction of (parts of) buildings, often by fire, and this meaning remained its core usage in the 19th century. By the 20th century, the verb had decreased in frequency and its prototypical core was no longer distinguishable from the data. Vernietigen, in contrast, was the less frequent variant in the 16th and 17th century and at that time, there were no clear contexts yet where the verb occurred. It was mostly found in a semantically diverse cluster where its nearsynonym vernielen was possible as well. From the 18th century onwards, the verb started to increase in frequency and it developed its prototypical sense of being used with abstract objects. In the 19th century, it also started to invade contexts where vernielen was preferred before (specifically related to the death of persons and to war). In the 20th century data, we also found a syntactic cluster, consisting of function words that often occur in the context of vernietigen.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Discussion", |
| "sec_num": "6" |
| }, |
| { |
| "text": "Onomasiologically, the analysis showcased how the nuances in the concept 'to destroy' evolve over time and have become more outspoken. For instance, the clusters related to the destruction of parts of buildings are not yet visible in the oldest data but they are important clusters in the more recent datasets. Similarly, the cluster with abstract objects is not yet distinguished by the analysis for the 16th and 17th century, but these objects form a cluster on their own in the 18th and 19th century data. Moreover, the analysis also showed how these particular nuances of meaning are typically expressed by a particular verb. In the visualization, for instance, there is clearly less overlap (or interchangeability) between the verbs in the later periods (except in the 20th century data, where vernielen is infrequent).", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Discussion", |
| "sec_num": "6" |
| }, |
| { |
| "text": "Thus, this case-study showcases an example of how formal onomasiological variation and conceptual onomasiological variation can interact. On the one hand, vernielen and vernietigen serve as formal alternatives in the largest cluster from the 16th and 17th century data. However, from the 18th century onwards, each verb increasingly retreats to its prototypical core. Arguably, they should therefore be considered conceptually distinct, prototypically referring to different nuances of meaning, even though they remain nearly synonymous.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Discussion", |
| "sec_num": "6" |
| }, |
| { |
| "text": "Methodologically, our usage of distributional models combined with a cluster analysis and the method, developed in Montes (2021) , to analyze the context words that are good representatives for the clusters, allowed us to show how both verbs changed semantically over time. The procedure employed was quite straightforward, using a single set of parameter settings to model tokens from four diachronic subcorpora. With this procedure, we extended the analyses in Montes et al. (2021) and Geeraerts (1997) to a much longer time span. Despite the fact that we used a completely different dataset (a continuous diachronic corpus rather than dictionary citations from the 19th century only) and analysis method (an automatic procedure rather than a manual linguistic analysis), the hypotheses outlined in Geeraerts (1997) were mostly confirmed. Further, this method allowed us to track semasiological change and to investigate how this interacts with onomasiological variation over time.", |
| "cite_spans": [ |
| { |
| "start": 115, |
| "end": 128, |
| "text": "Montes (2021)", |
| "ref_id": "BIBREF17" |
| }, |
| { |
| "start": 463, |
| "end": 483, |
| "text": "Montes et al. (2021)", |
| "ref_id": "BIBREF17" |
| }, |
| { |
| "start": 488, |
| "end": 504, |
| "text": "Geeraerts (1997)", |
| "ref_id": "BIBREF8" |
| }, |
| { |
| "start": 801, |
| "end": 817, |
| "text": "Geeraerts (1997)", |
| "ref_id": "BIBREF8" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Discussion", |
| "sec_num": "6" |
| }, |
| { |
| "text": "One shortcoming of the approach is that the ideal settings for the parameters need not be the same for other near-synonyms or for a comparable linguistic alternation in other languages. In fact, this is one of the major findings of Montes (2021) , who showed that there is no direct link between a choice of parameters and the linguistic phenomena that are revealed by a model constructed with the method proposed by Sch\u00fctze (1998) . Therefore, while in ', geheel 'completely'; 4: natuur 'nature', geluk 'luck', duizend 'thousand', veranderingen 'changes', zonder 'without', werden 'became (pl.)', schulden 'debts', werd 'became (sg.)'; 3: gramschap 'wrath', beeld 'statue, picture', vorsten 'monarchs', oogenblik 'moment', kunt 'can', word 'become (sg.) 10: dese 'this '; 5: doot 'death'; 4: desen 'this', t 'it'; 3: wet 'law', sulcke 'this', selve 'self', vyanden 'enemies', Christi '(of) Christ', dooden 'to kill', omme 'in order to', macht 'power', verlaten 12: werden 'become (pl.)'; 4: hunne 'their'; 3: elka\u00e2r 'each other ', vijand 'enemy', troepen 'troups', leger 'army', gebroken 'broken', slag 'battle', vloot 'fleet', oogst 'harvest' vernielen: 0.61 (60), vernietigen: 0.39 (39) 3", |
| "cite_spans": [ |
| { |
| "start": 232, |
| "end": 245, |
| "text": "Montes (2021)", |
| "ref_id": "BIBREF17" |
| }, |
| { |
| "start": 417, |
| "end": 431, |
| "text": "Sch\u00fctze (1998)", |
| "ref_id": "BIBREF20" |
| }, |
| { |
| "start": 454, |
| "end": 754, |
| "text": "', geheel 'completely'; 4: natuur 'nature', geluk 'luck', duizend 'thousand', veranderingen 'changes', zonder 'without', werden 'became (pl.)', schulden 'debts', werd 'became (sg.)'; 3: gramschap 'wrath', beeld 'statue, picture', vorsten 'monarchs', oogenblik 'moment', kunt 'can', word 'become (sg.)", |
| "ref_id": null |
| }, |
| { |
| "start": 770, |
| "end": 772, |
| "text": "';", |
| "ref_id": null |
| }, |
| { |
| "start": 773, |
| "end": 789, |
| "text": "5: doot 'death';", |
| "ref_id": null |
| }, |
| { |
| "start": 790, |
| "end": 806, |
| "text": "4: desen 'this',", |
| "ref_id": null |
| }, |
| { |
| "start": 807, |
| "end": 814, |
| "text": "t 'it';", |
| "ref_id": null |
| }, |
| { |
| "start": 815, |
| "end": 828, |
| "text": "3: wet 'law',", |
| "ref_id": null |
| }, |
| { |
| "start": 829, |
| "end": 843, |
| "text": "sulcke 'this',", |
| "ref_id": null |
| }, |
| { |
| "start": 844, |
| "end": 857, |
| "text": "selve 'self',", |
| "ref_id": null |
| }, |
| { |
| "start": 858, |
| "end": 876, |
| "text": "vyanden 'enemies',", |
| "ref_id": null |
| }, |
| { |
| "start": 877, |
| "end": 899, |
| "text": "Christi '(of) Christ',", |
| "ref_id": null |
| }, |
| { |
| "start": 900, |
| "end": 917, |
| "text": "dooden 'to kill',", |
| "ref_id": null |
| }, |
| { |
| "start": 918, |
| "end": 937, |
| "text": "omme 'in order to',", |
| "ref_id": null |
| }, |
| { |
| "start": 938, |
| "end": 952, |
| "text": "macht 'power',", |
| "ref_id": null |
| }, |
| { |
| "start": 953, |
| "end": 961, |
| "text": "verlaten", |
| "ref_id": null |
| }, |
| { |
| "start": 1029, |
| "end": 1189, |
| "text": "', vijand 'enemy', troepen 'troups', leger 'army', gebroken 'broken', slag 'battle', vloot 'fleet', oogst 'harvest' vernielen: 0.61 (60), vernietigen: 0.39 (39)", |
| "ref_id": null |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Discussion", |
| "sec_num": "6" |
| }, |
| { |
| "text": "(AB-STRACT OBJECTS)", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Discussion", |
| "sec_num": "6" |
| }, |
| { |
| "text": "5: invloed 'influence '; 4: zedelijk 'virtuous', kracht 'strength', macht 'power', revolutie 'revolution', bestaan 'existence, to exist', vrijheid 'freedom' vernielen: 0.20 (14) , vernietigen: 0.80 (56) 4 ((PARTS OF) BUILD-INGS (FIRE)) 6: brand 'fire'; 4: stad 'city'; 3: kerken 'churches', huizen 'houses', steden 'cities' vernielen: 0.97 (35), vernietigen: 0.03 (1) 6: zichzelf 'hisself/herself/themselves', volkomen 'completely', steden 'cities'; 5: werden 'became (pl.) ', leger 'army', vloot 'fleet', schepen 'ships'; 3: zorgvuldig 'carefully', gedeeltelijk 'partly', volledig 'completely', willen 'to want', brieven 'letters' vernielen: 0.36 (35) , vernietigen: 0.64 (63) 3(AB- 8: oorlog 'war'; 5: werden 'became (pl.)', nadat 'after', hele 'whole', gehele 'whole', oplage 'edition'; 4: documenten 'documents', moesten 'had to', volk 'people'; 3: recht 'right', kaart 'map', zouden 'would', joodse 'jewish', steden 'cities', exemplaren 'samples', geworden 'become (participle)', goden 'gods', zestig 'sixty', wereldoorlog 'world war', europese 'european' vernielen: 0.16 (19), vernietigen: 0.84 (97) 3(FUNC -TION WORDS) 14: alles 'everything'; 10: zelfs 'even'; 9: zichzelf 'hisself/herself/itself'; 7: niets 'nothing'; 6: uiteindelijk 'eventually'; 4: mens 'human'; 3: waarna 'after which', god 'god', erbij 'near it', erop 'on it', definitief 'definitive', jezelf 'yourself', onmogeliijk 'impossible' 4: huis 'house', aarde 'earth'; 3: muren 'walls', stenen 'stones' vernielen: 0.52 (17), vernietigen: 0.48 16this contribution we focus on a single set of parameters settings that were shown to be useful in analyses of the same linguistic example in another century, an alternative approach, that has been successfully employed in Montes (2021) , is to consider a broader number of parameter settings to analyze linguistic phenomena.", |
| "cite_spans": [ |
| { |
| "start": 22, |
| "end": 177, |
| "text": "'; 4: zedelijk 'virtuous', kracht 'strength', macht 'power', revolutie 'revolution', bestaan 'existence, to exist', vrijheid 'freedom' vernielen: 0.20 (14)", |
| "ref_id": null |
| }, |
| { |
| "start": 474, |
| "end": 652, |
| "text": "', leger 'army', vloot 'fleet', schepen 'ships'; 3: zorgvuldig 'carefully', gedeeltelijk 'partly', volledig 'completely', willen 'to want', brieven 'letters' vernielen: 0.36 (35)", |
| "ref_id": null |
| }, |
| { |
| "start": 1741, |
| "end": 1754, |
| "text": "Montes (2021)", |
| "ref_id": "BIBREF17" |
| } |
| ], |
| "ref_spans": [ |
| { |
| "start": 1113, |
| "end": 1127, |
| "text": "-TION WORDS)", |
| "ref_id": null |
| } |
| ], |
| "eq_spans": [], |
| "section": "Discussion", |
| "sec_num": "6" |
| } |
| ], |
| "back_matter": [ |
| { |
| "text": "Figure 1: Data in four subcorpora", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "A Appendix", |
| "sec_num": null |
| } |
| ], |
| "bib_entries": { |
| "BIBREF0": { |
| "ref_id": "b0", |
| "title": "NLTK: The natural language toolkit", |
| "authors": [ |
| { |
| "first": "Steven", |
| "middle": [], |
| "last": "Bird", |
| "suffix": "" |
| }, |
| { |
| "first": "Edward", |
| "middle": [], |
| "last": "Loper", |
| "suffix": "" |
| } |
| ], |
| "year": 2004, |
| "venue": "Proceedings of the ACL Interactive Poster and Demonstration Sessions", |
| "volume": "", |
| "issue": "", |
| "pages": "214--217", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Steven Bird and Edward Loper. 2004. NLTK: The natu- ral language toolkit. In Proceedings of the ACL In- teractive Poster and Demonstration Sessions, pages 214-217, Barcelona, Spain. Association for Compu- tational Linguistics.", |
| "links": null |
| }, |
| "BIBREF1": { |
| "ref_id": "b1", |
| "title": "Extracting semantic representations from word co-occurrence statistics: A computational study", |
| "authors": [ |
| { |
| "first": "John", |
| "middle": [ |
| "A" |
| ], |
| "last": "Bullinaria", |
| "suffix": "" |
| }, |
| { |
| "first": "Joseph", |
| "middle": [ |
| "P" |
| ], |
| "last": "Levy", |
| "suffix": "" |
| } |
| ], |
| "year": 2007, |
| "venue": "Behavior Research Methods", |
| "volume": "39", |
| "issue": "3", |
| "pages": "510--526", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "John A. Bullinaria and Joseph P. Levy. 2007. Extracting semantic representations from word co-occurrence statistics: A computational study. Behavior Research Methods, 39(3):510-526.", |
| "links": null |
| }, |
| "BIBREF2": { |
| "ref_id": "b2", |
| "title": "Word association norms, mutual information, and lexicography", |
| "authors": [ |
| { |
| "first": "Kenneth", |
| "middle": [ |
| "Ward" |
| ], |
| "last": "Church", |
| "suffix": "" |
| }, |
| { |
| "first": "Patrick", |
| "middle": [], |
| "last": "Hanks", |
| "suffix": "" |
| } |
| ], |
| "year": 1989, |
| "venue": "ACL '89: Proceedings of the 27th Annual Meeting on Association for Computational Linguistic", |
| "volume": "", |
| "issue": "", |
| "pages": "76--83", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Kenneth Ward Church and Patrick Hanks. 1989. Word association norms, mutual information, and lexicog- raphy. In ACL '89: Proceedings of the 27th Annual Meeting on Association for Computational Linguistic, pages 76-83.", |
| "links": null |
| }, |
| "BIBREF3": { |
| "ref_id": "b3", |
| "title": "Introduction", |
| "authors": [ |
| { |
| "first": "Ewa", |
| "middle": [], |
| "last": "Dabrowska", |
| "suffix": "" |
| }, |
| { |
| "first": "Dagmar", |
| "middle": [], |
| "last": "Divjak", |
| "suffix": "" |
| } |
| ], |
| "year": 2015, |
| "venue": "Handbook of Cognitive Linguistics. De Gruyter Mouton", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Ewa Dabrowska and Dagmar Divjak. 2015. Introduc- tion. In Ewa Dabrowska and Dagmar Divjak, editors, Handbook of Cognitive Linguistics. De Gruyter Mou- ton, Berlin.", |
| "links": null |
| }, |
| "BIBREF4": { |
| "ref_id": "b4", |
| "title": "Token-based vector space models as semantic control in lexical sociolectometry", |
| "authors": [ |
| { |
| "first": "Pascale", |
| "middle": [], |
| "last": "Stefano De", |
| "suffix": "" |
| } |
| ], |
| "year": 2019, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Stefano De Pascale. 2019. Token-based vector space models as semantic control in lexical sociolectometry. Doctoral dissertation, Leuven: KU Leuven.", |
| "links": null |
| }, |
| "BIBREF5": { |
| "ref_id": "b5", |
| "title": "Turning Digitised Material into a Diachronic Corpus: Metadata Challenges in the Nederlab Project", |
| "authors": [ |
| { |
| "first": "Katrien", |
| "middle": [], |
| "last": "Depuydt", |
| "suffix": "" |
| }, |
| { |
| "first": "Hennie", |
| "middle": [], |
| "last": "Brugman", |
| "suffix": "" |
| } |
| ], |
| "year": 2019, |
| "venue": "DAT-eCH2019", |
| "volume": "", |
| "issue": "", |
| "pages": "169--173", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Katrien Depuydt and Hennie Brugman. 2019. Turning Digitised Material into a Diachronic Corpus: Meta- data Challenges in the Nederlab Project. In DAT- eCH2019, pages 169-173, New York.", |
| "links": null |
| }, |
| "BIBREF6": { |
| "ref_id": "b6", |
| "title": "Preponderantieverschillen bij bijnasynoniemen. De Nieuwe Taalgids", |
| "authors": [ |
| { |
| "first": "Dirk", |
| "middle": [], |
| "last": "Geeraerts", |
| "suffix": "" |
| } |
| ], |
| "year": 1985, |
| "venue": "", |
| "volume": "78", |
| "issue": "", |
| "pages": "18--27", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Dirk Geeraerts. 1985. Preponderantieverschillen bij bijnasynoniemen. De Nieuwe Taalgids, 78:18-27.", |
| "links": null |
| }, |
| "BIBREF7": { |
| "ref_id": "b7", |
| "title": "Where does prototypicality come from?", |
| "authors": [ |
| { |
| "first": "Dirk", |
| "middle": [], |
| "last": "Geeraerts", |
| "suffix": "" |
| } |
| ], |
| "year": 1988, |
| "venue": "Topics in Cognitive Linguistics", |
| "volume": "", |
| "issue": "", |
| "pages": "207--229", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Dirk Geeraerts. 1988. Where does prototypicality come from? In Brygida Rudzka-Ostyn, editor, Topics in Cognitive Linguistics, pages 207-229. John Ben- jamins, Amsterdam/Philadelphia.", |
| "links": null |
| }, |
| "BIBREF8": { |
| "ref_id": "b8", |
| "title": "Diachronic prototype semantics: a contribution to historical lexicology", |
| "authors": [ |
| { |
| "first": "Dirk", |
| "middle": [], |
| "last": "Geeraerts", |
| "suffix": "" |
| } |
| ], |
| "year": 1997, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Dirk Geeraerts. 1997. Diachronic prototype semantics: a contribution to historical lexicology. Oxford Uni- versity Press, Oxford.", |
| "links": null |
| }, |
| "BIBREF9": { |
| "ref_id": "b9", |
| "title": "Introducing Cognitive Linguistics", |
| "authors": [ |
| { |
| "first": "Dirk", |
| "middle": [], |
| "last": "Geeraerts", |
| "suffix": "" |
| }, |
| { |
| "first": "Hubert", |
| "middle": [], |
| "last": "Cuyckens", |
| "suffix": "" |
| } |
| ], |
| "year": 2007, |
| "venue": "The Oxford Handbook of Cognitive Linguistics", |
| "volume": "", |
| "issue": "", |
| "pages": "3--21", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Dirk Geeraerts and Hubert Cuyckens. 2007. Introduc- ing Cognitive Linguistics. In Dirk Geeraerts and Hubert Cuyckens, editors, The Oxford Handbook of Cognitive Linguistics, pages 3-21. Oxford University Press, Oxford.", |
| "links": null |
| }, |
| "BIBREF10": { |
| "ref_id": "b10", |
| "title": "The Structure of Lexical Variation: Meaning, Naming, and Context", |
| "authors": [ |
| { |
| "first": "Dirk", |
| "middle": [], |
| "last": "Geeraerts", |
| "suffix": "" |
| }, |
| { |
| "first": "Stefan", |
| "middle": [], |
| "last": "Grondelaers", |
| "suffix": "" |
| }, |
| { |
| "first": "Peter", |
| "middle": [], |
| "last": "Bakema", |
| "suffix": "" |
| } |
| ], |
| "year": 1994, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": { |
| "DOI": [ |
| "10.1515/9783110873061" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "Dirk Geeraerts, Stefan Grondelaers, and Peter Bakema. 1994. The Structure of Lexical Variation: Meaning, Naming, and Context. De Gruyter Mouton, Berlin, New York.", |
| "links": null |
| }, |
| "BIBREF11": { |
| "ref_id": "b11", |
| "title": "Advances in cognitive sociolinguistics", |
| "authors": [ |
| { |
| "first": "Dirk", |
| "middle": [], |
| "last": "Geeraerts", |
| "suffix": "" |
| }, |
| { |
| "first": "Gitte", |
| "middle": [], |
| "last": "Kristiansen", |
| "suffix": "" |
| }, |
| { |
| "first": "Yves", |
| "middle": [], |
| "last": "Peirsman", |
| "suffix": "" |
| } |
| ], |
| "year": 2010, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Dirk Geeraerts, Gitte Kristiansen, and Yves Peirsman, editors. 2010. Advances in cognitive sociolinguistics. Mouton de Gruyter, New York, N.Y.", |
| "links": null |
| }, |
| "BIBREF12": { |
| "ref_id": "b12", |
| "title": "Monitoring polysemy: Word space models as a tool for large-scale lexical semantic analysis", |
| "authors": [ |
| { |
| "first": "K", |
| "middle": [], |
| "last": "Heylen", |
| "suffix": "" |
| }, |
| { |
| "first": "T", |
| "middle": [], |
| "last": "Wielfaert", |
| "suffix": "" |
| }, |
| { |
| "first": "D", |
| "middle": [], |
| "last": "Speelman", |
| "suffix": "" |
| }, |
| { |
| "first": "D", |
| "middle": [], |
| "last": "Geeraerts", |
| "suffix": "" |
| } |
| ], |
| "year": 2015, |
| "venue": "Lingua", |
| "volume": "157", |
| "issue": "", |
| "pages": "153--172", |
| "other_ids": { |
| "DOI": [ |
| "10.1016/j.lingua.2014.12.001" |
| ] |
| }, |
| "num": null, |
| "urls": [], |
| "raw_text": "K. Heylen, T. Wielfaert, D. Speelman, and D. Geeraerts. 2015. Monitoring polysemy: Word space models as a tool for large-scale lexical semantic analysis. Lingua, 157:153-172.", |
| "links": null |
| }, |
| "BIBREF13": { |
| "ref_id": "b13", |
| "title": "A Systematic Study of Semantic Vector Space Model Parameters", |
| "authors": [ |
| { |
| "first": "Douwe", |
| "middle": [], |
| "last": "Kiela", |
| "suffix": "" |
| }, |
| { |
| "first": "Stephen", |
| "middle": [], |
| "last": "Clark", |
| "suffix": "" |
| } |
| ], |
| "year": 2014, |
| "venue": "Proceedings of the 2nd Workshop on Continuous Vector Space Models and Their Compositionality", |
| "volume": "", |
| "issue": "", |
| "pages": "21--30", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Douwe Kiela and Stephen Clark. 2014. A Systematic Study of Semantic Vector Space Model Parameters. In Proceedings of the 2nd Workshop on Continuous Vector Space Models and Their Compositionality, pages 21-30, Gothenburg. ACL.", |
| "links": null |
| }, |
| "BIBREF14": { |
| "ref_id": "b14", |
| "title": "The invariance hypothesis", |
| "authors": [ |
| { |
| "first": "George", |
| "middle": [], |
| "last": "Lakoff", |
| "suffix": "" |
| } |
| ], |
| "year": 1990, |
| "venue": "Cognitive Linguistics", |
| "volume": "1", |
| "issue": "1", |
| "pages": "39--74", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "George Lakoff. 1990. The invariance hypothesis. Cog- nitive Linguistics, 1(1):39-74.", |
| "links": null |
| }, |
| "BIBREF15": { |
| "ref_id": "b15", |
| "title": "Mariana Montes. 2021. Cloudspotting. Visual analytics for distributional semantics. Doctoral dissertation", |
| "authors": [], |
| "year": null, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Mariana Montes. 2021. Cloudspotting. Visual analytics for distributional semantics. Doctoral dissertation, Leuven: KU Leuven.", |
| "links": null |
| }, |
| "BIBREF17": { |
| "ref_id": "b17", |
| "title": "Indestructible insights. A case study in distributional prototype semantics", |
| "authors": [ |
| { |
| "first": "Mariana", |
| "middle": [], |
| "last": "Montes", |
| "suffix": "" |
| }, |
| { |
| "first": "Karlien", |
| "middle": [], |
| "last": "Franco", |
| "suffix": "" |
| }, |
| { |
| "first": "Kris", |
| "middle": [], |
| "last": "Heylen", |
| "suffix": "" |
| } |
| ], |
| "year": 2021, |
| "venue": "Cognitive Sociolinguistics Revisited", |
| "volume": "", |
| "issue": "", |
| "pages": "251--264", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Mariana Montes, Karlien Franco, and Kris Heylen. 2021. Indestructible insights. A case study in dis- tributional prototype semantics. In Gitte Kristiansen, Karlien Franco, Stefano De Pascale, Laura Rosseel, and Weiwei Zhang, editors, Cognitive Sociolinguis- tics Revisited, pages 251-264. De Gruyter Mouton, Berlin/Boston.", |
| "links": null |
| }, |
| "BIBREF18": { |
| "ref_id": "b18", |
| "title": "Python module for type-and token-level distributional models", |
| "authors": [], |
| "year": null, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "QLVL. 2021. nephosem. Python module for type-and token-level distributional models.", |
| "links": null |
| }, |
| "BIBREF19": { |
| "ref_id": "b19", |
| "title": "R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing", |
| "authors": [ |
| { |
| "first": "", |
| "middle": [], |
| "last": "R Core Team", |
| "suffix": "" |
| } |
| ], |
| "year": 2020, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "R Core Team. 2020. R: A Language and Environment for Statistical Computing. R Foundation for Statisti- cal Computing, Vienna, Austria.", |
| "links": null |
| }, |
| "BIBREF20": { |
| "ref_id": "b20", |
| "title": "Automatic word sense discrimination", |
| "authors": [ |
| { |
| "first": "Hinrich", |
| "middle": [], |
| "last": "Sch\u00fctze", |
| "suffix": "" |
| } |
| ], |
| "year": 1998, |
| "venue": "Computational Linguistics", |
| "volume": "24", |
| "issue": "1", |
| "pages": "97--123", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Hinrich Sch\u00fctze. 1998. Automatic word sense discrimi- nation. Computational Linguistics, 24(1):97-123.", |
| "links": null |
| } |
| }, |
| "ref_entries": { |
| "TABREF1": { |
| "content": "<table/>", |
| "type_str": "table", |
| "num": null, |
| "html": null, |
| "text": "Uses of vernielen and vernietigen. Adapted from" |
| }, |
| "TABREF2": { |
| "content": "<table><tr><td>Cluster</td><td>Context words</td></tr></table>", |
| "type_str": "table", |
| "num": null, |
| "html": null, |
| "text": "16th & 17th century" |
| }, |
| "TABREF5": { |
| "content": "<table><tr><td>Cluster</td><td>Context words</td></tr></table>", |
| "type_str": "table", |
| "num": null, |
| "html": null, |
| "text": "18th century" |
| }, |
| "TABREF6": { |
| "content": "<table><tr><td>Cluster</td><td>Context words</td></tr></table>", |
| "type_str": "table", |
| "num": null, |
| "html": null, |
| "text": "19th century" |
| }, |
| "TABREF8": { |
| "content": "<table><tr><td>Cluster</td><td>Context words</td></tr></table>", |
| "type_str": "table", |
| "num": null, |
| "html": null, |
| "text": "20th century" |
| } |
| } |
| } |
| } |
