Frontiers in Language Sciences

Metaphors are a distinctive feature of literary language, yet they remain less studied experimentally than everyday metaphors. Moreover, previous psycholinguistic and computational approaches overlooked the temporal dimension, although many literary metaphors were coined centuries apart from contemporary readers. This study innovatively applies tools from diachronic distributional semantics to assess whether the processing costs of literary metaphors varied over time and genre. Specifically, we used the Word2Vec algorithm and period-specific literary and nonliterary corpora, for a total of 124 million tokens, to generate word embeddings. Using these embeddings, we modeled changes in the semantic similarity between topics and vehicles of 515 19th-century literary metaphors, taking this measure as a proxy for metaphor processing demands. Overall, semantic similarity, and hence metaphor processing demands, remained stable over time. However, genre played a key role: metaphors appeared more difficult (i.e., lower topic-vehicle similarity) in modern literary contexts than in 19th-century literature, but easier (i.e., higher topic-vehicle similarity) in today's nonliterary language (e.g., the web) than in 19th-century nonliterary texts. This pattern was further shaped by semantic features of metaphors' individual terms, such as vector coherence and semantic neighborhood density. Collectively, these findings align with broader linguistic changes in Italian, such as the stylistic simplification of modern literature, which may have increased metaphor processing demands, and the high creativity of the web's language, which seems to render metaphor more accessible.

استعاره‌ها ویژگی متمایز زبان ادبی هستند، اما به‌صورت تجربی کمتر از استعاره‌های روزمره مورد مطالعه قرار گرفته‌اند. افزون بر این، رویکردهای پیشین روان‌زبان‌شناسی و محاسباتی، بُعد زمانی را نادیده گرفته‌اند، در حالی که بسیاری از استعاره‌های ادبی قرن‌ها پیش از خوانندگان معاصر ابداع شده‌اند. این مطالعه به‌طور نوآورانه از ابزارهای معناشناسی توزیعی تاریخی برای ارزیابی این‌که آیا هزینه‌های پردازش استعاره‌های ادبی در طول زمان و ژانر متفاوت بوده است یا خیر، بهره می‌گیرد. به‌طور مشخص، الگوریتم Word2Vec و پیکره‌های ادبی و غیرادبی مختص دوره‌های زمانی مختلف، با مجموع ۱۲۴ میلیون توکن، برای تولید تعبیه‌های واژه‌ای به‌کار گرفته شد. با استفاده از این تعبیه‌ها، تغییرات در شباهت معنایی میان موضوع و وسیله‌ی ۵۱۵ استعاره ادبی قرن نوزدهم مدل‌سازی شد و این سنجه به‌عنوان نماینده‌ای برای سنجش دشواری پردازش استعاره در نظر گرفته شد. به‌طور کلی، شباهت معنایی و در نتیجه دشواری پردازش استعاره در طول زمان پایدار باقی ماند. با این حال، ژانر نقش کلیدی ایفا کرد: استعاره‌ها در زمینه‌های ادبی معاصر دشوارتر (یعنی شباهت کمتر میان موضوع و وسیله) از ادبیات قرن نوزدهم ظاهر شدند، اما در زبان غیرادبی امروزی (مانند وب) آسان‌تر (یعنی شباهت بیشتر میان موضوع و وسیله) از متون غیرادبی قرن نوزدهم بودند. این الگو همچنین تحت تأثیر ویژگی‌های معنایی واژگان منفرد استعاره‌ها، مانند انسجام برداری و تراکم همسایگی معنایی، قرار گرفت. در مجموع، این یافته‌ها با تغییرات زبانی گسترده‌تر در زبان ایتالیایی، مانند ساده‌سازی سبکی ادبیات معاصر که ممکن است دشواری پردازش استعاره را افزایش داده باشد و خلاقیت بالای زبان وب که به نظر می‌رسد استعاره را دسترس‌پذیرتر می‌سازد، هم‌راستا است.

Introduction Event nominalizations are a central feature of academic language: they compress complex events into nominal structures, increasing informational density while reducing the transparency of thematic-role assignment. The present eye-tracking study examined whether nominalizations impose greater processing demands than corresponding active and passive constructions in developing readers, and whether these effects vary across readers with different linguistic and literacy profiles. Methods Seventy-nine fifth- and sixth-grade children participated in the study: 28 typically developing monolingual Italian children, 21 monolingual Italian children with developmental dyslexia, and 30 typically developing bilingual children with migrant background and Italian as the societal language. Participants silently read short passages in which the same target event was expressed either as a nominalization, a passive construction, or an active construction, and then answered comprehension questions while their eye movements were recorded. Accuracy, standard eye-movement measures, and scanpaths were analyzed to capture comprehension outcomes and online reading strategies. Results Nominalizations did not significantly reduce comprehension accuracy, but they affected online processing during both initial reading and the question-answering phase. Scanpath dissimilarity was higher for nominalizations than for both verbal constructions during initial reading, and than for passives during question-answering in correctly answered trials. In these trials, nominalizations were also associated with a higher probability of partial rescanning and a lower probability of minimal rescanning relative to active texts, indicating greater reliance on rereading when answering questions correctly. Moreover, children with dyslexia showed a generalized increase in processing effort, whereas bilingual children displayed an intermediate profile, with increased reliance on rereading during question-answering. Discussion These findings indicate that nominalizations significantly influence online reading processes even when questions are answered correctly, suggesting additional demands in retrieving information about event participants. Overall processing patterns differed across reader groups, and children with dyslexia showed the greatest effort in both online reading and question-answering.