[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-85349-en":3,"doc-seo-85349-105":29,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},85349,13056703020460,"Valentina","https://ap-avatar.wpscdn.com/avatar/be000253dac470eee5d?_k=1778207105932848923",8,"Research & Report","Losing My Composure: Predicting Compositionality Over Time","Explores semantic change in German and English noun compounds by modeling gradual shifts in meaning and compositionality across past decades. Introduces the Compositionality Trend Prediction task, evaluated on a new dataset of in-context compositionality ratings sampled from diachronic corpora, covering 23 German and 26 English target compounds with per-decade scores. Uses temporal-slice training of semantic vector representations, yielding ~100 models per representation type and slice length (1–5 decades). Finds only a small negative compositionality trend over time and shows narrow-slice training and static models can outperform a half-century window approach.","arXiv :2607 . 1 1667v 1 [ cs .CL] 13 Jul 2026  \nLong Paper  \nLosing My Composure: Predicting Compositionality Over Time  \nChris Jenkins∗, Emma Raimundo Schulz, Filip Mileti´c, Sabine Schulte im Walde  \nUniversity of Stuttgart [christopher.jenkins@ims.uni-stuttgart.de](christopher.jenkins@ims.uni-stuttgart.de)  \nWe explore the phenomenon of semantic change of German and English noun compounds, with the objective of investigating and modeling gradual changes of meanings and degrees of compositionality in the past and over time. To do so, we introduce the Compositionality Trend Prediction task, which is evaluated against a novel dataset of in-context compositionality ratings sampled across several decades of diachronic corpora for 23 German and 26 English target compounds, uniquely providing per-decade ratings and corresponding trends over time. These per-decade compositionality ratings allow us to investigate empirically untested hypotheses of generalized trends in compositionality over time, such as the idea that compounds should become less compositional (less transparent) over time. Beyond our empirical observations from the diachronic compositionality annotations, we perform experiments with semantic vector representations of varying complexity, as well as several temporal granularities for training these representations on diachronic data, resulting in about 100 models of each representation type, each covering a different 1–5 decade slice of a diachronic corpus. Contrary to the decisive tendency posited in the literature, we find only a small negative trend in compositionality overtime in our target compounds. In our computational experiments, we find that using models trained on narrow time slices of diachronic data (single decades, or incrementally expanding temporal windows) align better with the per-decade compositionality ratings than those trained on an entire half-century window, the latter setting being an analog for the prevalent modeling approach of training representations on an entire half of a corpus’ data. Additionally, we find static representations to be competitive with contextual representations in the Compositionality Trend Prediction task.  \n1. Motivation  \nEvery word is, so to speak, on its own journey through the ages, and compounds have some company along the way: Kaffeehaus (coffee house) can be analyzed in terms of its use as a whole (e.g. whether it is synonymous with the French borrowing café or as a consequence of how its constituents are used (e.g. German Kaffee as a plant, a drink, or  \n∗ Corresponding Author  \nThis is a pre-print under review by Computational Linguistics  \n© 2026 Association for Computational Linguistics  \nComputational Linguistics Volume vv, Number nn  \na meal) . Ice water may have been used in differing contexts before, during, and after the era of global trade in naturally-occurring ice (≈19th century), prior to the invention of electrical refrigeration, while uses of ice or water may not have changed as much. Bosom friend (a close friend) is less compositional with respect to its modifier bosom than it is with respect to its head friend. If it was ever used in a more compositional way (e.g. a friend that one is constantly hugging), such a meaning is not extant in the diachronic corpus that we examined. Investigating the semantic change of noun compounds thus takes advantage of the ready-made comparison that each compound invites, as the overall meaning of a compound noun can be interpreted as a single –possibly idiomatic– unit, or with respect to the meanings of its constituent nouns. In this vein, we consider noun compounds as a specific instance of multiword expressions while investigating the tension between the potential for semantic change at any point in time, and the stability of many meanings over time (Blank 1999; Koch 2016) .  \nWhile some changes have occurred dramatically at a known time, like the signification of atomic after 19451, later divergent uses as evi","cbCainbReuz5UPzk","https://ap.wps.com/l/cbCainbReuz5UPzk","pdf",996682,1,58,"English","en",105,"# Motivation\n## Semantic change and multiword expressions\n## Temporal scales in language change\n## Beyond two-era comparisons and decade granularity","[{\"question\":\"What problem does the paper address about noun compounds?\",\"answer\":\"It investigates how noun-compound meanings and degrees of compositionality change over time in German and English.\"},{\"question\":\"What is the Compositionality Trend Prediction task?\",\"answer\":\"A task introduced to predict compositionality trends using per-decade in-context ratings derived from diachronic corpora for specified German and English target compounds.\"},{\"question\":\"What conclusions does the paper draw about compositionality trends over time?\",\"answer\":\"The experiments show only a small negative trend in compositionality over time, and models trained on narrower time slices align better with per-decade ratings than models trained on broader half-century windows.\"}]",1784202703,146,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":27},"losing-my-composure-predicting-compositionality-over-time","",{"@graph":35,"@context":85},[36,53,68],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/losing-my-composure-predicting-compositionality-over-time/85349/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-24","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does the paper address about noun compounds?","Question",{"text":75,"@type":76},"It investigates how noun-compound meanings and degrees of compositionality change over time in German and English.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What is the Compositionality Trend Prediction task?",{"text":80,"@type":76},"A task introduced to predict compositionality trends using per-decade in-context ratings derived from diachronic corpora for specified German and English target compounds.",{"name":82,"@type":73,"acceptedAnswer":83},"What conclusions does the paper draw about compositionality trends over time?",{"text":84,"@type":76},"The experiments show only a small negative trend in compositionality over time, and models trained on narrower time slices align better with per-decade ratings than models trained on broader half-century windows.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":45,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":45,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":45,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":45,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":45,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":45,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":45,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]