Skip to main content
Article

Urban Transportation in Latvian Novels or Why Do You Use a 19th-Century Horse-drawn Cab When You Have a 20th-Century Taxi?


Abstract

The article explores the depiction of urban transportation in Latvian novels from the late 19th century to the mid-20th century, aiming to understand how these images reflect broader societal modernization and technological and social changes in an increasingly urbanized environment. The first part of the study explores the frequency and dissemination of the mentions of urban vehicles in the novels. Two methods were used to identify relevant transport terms – Word2Vec and the Gemini 1.5 language model, comparing the results of both approaches. In the second part of the study, particular attention is given to the horse-drawn cab and the taxi, which illustrate economic development, societal modernization, and the growing disparity between different social strata. The study emphasizes that transportation is a practical means of mobility and a significant cultural and social symbol. The study uses the data set Corpus of Latvian Early Novels, which includes novels published between 1879 and 1940.

Keywords: Latvian novels, urban transportation, Word2Vec, large language models, horse-drawn cab, taxi

How to Cite:

Eglāja-Kristsone, E., Baklāne, A. & Saulespurēns, V., (2026) “Urban Transportation in Latvian Novels or Why Do You Use a 19th-Century Horse-drawn Cab When You Have a 20th-Century Taxi?”, Journal of Computational Literary Studies 5(1). doi: https://doi.org/10.48694/jcls.4222

141 Views

22 Downloads

Published on
2026-05-07

Peer Reviewed

1. Introduction

Literature as a source for the history of everyday life has become a vast field of research, offering profound insights into people’s life experiences, emotions, and social processes, which are often absent from traditional historical sources. The link between everyday life and modernization processes is particularly noteworthy in prose fiction: technological progress, urbanization, and changes in social structure are documented in the literature both directly and through symbols, revealing not only the external world but also changes in the psychology and culture of society.

Urban transportation is not only a functional infrastructure; it is also a powerful cultural symbol. As cities modernize, modes of transport shape and reflect changing experiences of space, time, identity, and technology. In literature, vehicles such as horse-drawn cabs1 (ormanis, fūrmanis, važonis) and taxis (taksis, taksītis, taksometrs) often serve as more than a means of travel; they become narrative devices that express social aspiration, class tension, emotional crisis, or ideological transition. This article explores how the evolution of urban transportation – from horse-drawn carriages to motorized taxis – is reflected in Latvian fiction across the late 19th and early 20th centuries. The core research question is: How does the transition from horse-drawn cab to taxi reflect shifting urban subjectivities and sociotechnical imaginaries in Latvian literature?

To approach this question, we adopt a hybrid methodology that combines computational analysis of literary corpora with close reading of selected narrative episodes. Our approach begins with the analysis of transport-related vocabulary using word embeddings and a large language model for keyword extraction across the LatSenRom corpus, which comprises Latvian novels published between 1876 and 1940. To contextualise the prevalence of the concepts of the horse-drawn cab and taxi among other vehicles, we also examine the frequency of mentions of other land vehicles, including both mechanised vehicles modern for their time (auto, ‘car’; automobilis, ‘automobile’; dzelzceļš, ‘railway’; tramvajs, ‘tram’, etc.) and horse-drawn transport (pajūgs, ‘carriage’; vezums, ‘wagon’; ore, ‘farm wagon’, etc.). We then turn to interpretative analysis, examining how various modes of transport function symbolically and narratively within individual texts, particularly in relation to representations of modernity, class, gender, and psychological experience (Kohlrausch and Behrends 2014).

This dual approach is situated within the broader field of computational literary studies, which has increasingly emphasized the value of combining distant reading with interpretive frameworks. The methods employed in the article engage and contribute to the vast tradition of computational and digital literary studies, as well as digital history studies in particular sharing the ambition to analyze bodies of text that span long periods to capture changes in language, style, genre, as well as culture and society that occur over time (Moretti 2005, Moretti 2013, Jockers 2013, Underwood 2019, Piper 2018, Graham et al. 2016, Fridlund et al. 2023). In approaching the study of cities, urbanization, and modernization, it is notable that the representation of urban and rural spaces has long occupied a central place in scholarly research. Franco Moretti, in his pioneering work Atlas of the European Novel 1800–1900, examined not only the geographical distribution of literary forms but also the spatial dynamics depicted within the novels themselves (Moretti 1998). Building on this foundation, other scholars have expanded the exploration of how rural and urban spaces are modeled in literature. Dennis Yi Tenen, for example, integrated narratological concepts such as diegetic density and clutter distance to capture the complexity of spatial representation in literary texts (Tenen 2018). Federica Bologna, meanwhile, investigated the lexical presence of urban-related terms in twentieth-century English science fiction (Bologna 2020).2 Moving beyond the mere cataloging of place names, current research on urban spaces increasingly foregrounds the vocabulary of urban material culture, highlighting its significance for modeling and identifying various types of spaces. Our study builds on these foundations by applying such methods to a small-language literary tradition, where digital tools remain underutilized. This paper, however, does not yet attempt to model broader semantic domains, opting instead to concentrate on the more narrowly scoped domain of vehicles.

While 21st-century scholars have increasingly embraced complex methodologies, particularly those leveraging machine learning, to move beyond the traditional corpus analysis paradigm of word search and frequency analysis, the rise of new-generation language models has sparked a renewed interest in earlier approaches. Word embeddings and new-generation language models offer unprecedented opportunities to identify terms that align with the specific semantic domains a researcher aims to explore.

In this study, we explore a mixed-methods approach, integrating digital tools, simple document frequency and word frequency counts, and qualitative interpretative methods. We begin by employing the Word2Vec machine learning algorithm (Mikolov et al. 2013) along with the Gemini 1.5 language model (Gemini Team 2024) to identify transport-related concepts. This is followed by computer-assisted frequency counts to trace the occurrence of these concepts across a corpus of novels. Next, we use the concordance and word frequency list features of a corpus analysis platform3 to investigate linguistic patterns in greater depth. Finally, we conduct a close interpretative analysis of representations of the horse-drawn cab and taxi in the Corpus of Latvian Early Novels (1879–1940), drawing on perspectives from modern and urban material culture studies within Latvian as well as broader literary and cultural histories.4

The most technologically complex aspect of the study lies in the methodology used for identifying vehicle-related terms within the text. In the past, researchers have relied on manual or semiautomated methods to discover and annotate concepts of interest. In recent years, efforts have increasingly scaled up with the adoption of word embedding-based techniques (Mikolov et al. 2013), followed more recently by zero-shot large language models (LLMs) (Karjus 2025, Fan et al. 2023, Törnberg 2024, Ziems et al. 2023). Methods based on embeddings have proven particularly effective in automating the identification of semantic similarity and difference in terms or larger discourse segments (Rodman 2020, Rodriguez et al. 2023).

The dataset used in the study is the Corpus of Latvian Early Novels (1879–1940) / Latviešu senāko romānu korpuss (hereafter LatSenRom), which includes novels written in Latvian and published in book form between 1879 and 1940. As a corpus, LatSenRom exemplifies historical digitized datasets that span several decades and reflect a paradigm shift in typeface usage and orthographic norms. It highlights persistent challenges that humanities researchers face, such as optical character recognition errors, evolving writing conventions, and polysemy. Moreover, there is a significant disparity in the number of works published before and after the 1920s. Working with such data underscores the importance of addressing data heterogeneity and quality issues. It advocates for digital analysis not as a ‘quick fix’ but as a relatively slow, iterative process allowing backtracking to correct errors or refine approaches. Without the ability to contextualize the analysis results within the text, or when datasets are too large for manual verification, the reliability of those results may be compromised. These considerations emphasize the need for meticulous methodologies and critical engagement when working with digitized historical data.

The case study was inspired by the question posed in Kārlis Lapiņš’s novel Students in the Farm (Studenti fermā): “Why do you drive a 19th-century carriage when you have a 20th-century taxi at hand?” (see Lapiņš 1934b, 129). The question succinctly describes the tension between tradition and modernity, symbolically embodied by the transition from the carriage to the taxi, which examines the impact of this transition in the broader context of urban transport development in Latvian literature.

The introduction of motorized taxis was a significant shift in urban mobility, offering greater efficiency and a new form of urban anonymity that contrasted sharply with the more personalized and slower journeys provided by horse-drawn carriages. Viewed through this lens, the transition from horse-drawn carriages to motorized taxis not only reshaped urban transport but also left a lasting mark on the cultural and literary landscape of the era. From a narrative analysis perspective, the depiction of urban transport reflects the characters’ emotional state, social status, and the broader cultural environment. The increasing presence of urban transportation – horse-drawn carriages, trains, trams, buses (or omnibuses), and cars – within literature appears as a significant element that serves both as a backdrop for the narrative’s action and as a symbol of the broader social changes of the era.

The insights gained from the LatSenRom research into the prevalence of vehicle concepts in late 19th- and early 20th-century Latvian long prose are based on an examination of the largest Latvian fiction dataset to date. The results provide information on the mention of the horse-drawn cab, the taxi, and other vehicles in the data studied and, thanks to the juxtaposition with the facts of transport history and the insights of transport history researchers, can also contribute to studying material history and cultural history in general.

2. Dataset and the Methodology of Term Extraction

2.1 Dataset

The LatSenRom dataset was developed at the National Library of Latvia in consultation with researchers at the Institute of Literature, Folklore and Art at the University of Latvia. The corpus version used for this article contains 463 works (novels and parts of books, trilogies, tetralogies) by 190 authors. The dataset contains about 36.3 million tokens, including about 28.8 million words.5

The creation of the LatSenRom began as part of an international project – the COST Action Distant Reading for European Literary History (CA16204), which took place from 2017 to 2022. The project’s main objective was to establish the European Literary Text Collection (ELTeC) (Schöch et al. 2021). ELTeC is a collection of datasets comprising corpora of novels in various languages, compiled according to unified principles. Each language corpus was designed to include 100 works published in European countries between 1840 and 1920. These works were carefully selected from the broader range of publications based on balancing criteria: a balanced number of works from each decade, author gender categories, work length categories, and the canonical status of the works. Additionally, in line with the selection algorithm, no more than three works by any single author were included in the corpus.

The strict selection criteria served well in creating a representative corpus for major European languages, where it was possible to choose from a vast selection of thousands of works. At the same time, the distinctiveness of smaller languages became strikingly evident – in several countries, the novelistic literary tradition developed later, involved fewer women, and had a relatively small number of works considered canonical. This experience prompted the creators of the Latvian novel corpus to adopt a different approach for its further development: instead of forming a representative selection, a decision was made to include in the dataset all original novels published in book form in the Latvian language in the present-day territory of Latvia or abroad.

Figure 1 illustrates the frequency of new novel publishing over time. It is important to note that the dataset excludes novels published solely in periodicals during the period under consideration. Based on book publishing data, the novel as a genre remained relatively unfamiliar in Latvia at the end of the 19th and the beginning of the 20th centuries. Until 1899, only 12 novels were published, with an additional 38 appearing between 1900 and 1918 (year of the founding of the independent Republic of Latvia). However, following Latvia’s independence, publishing of books surged, with more than 400 new novels released between 1919 and 1940.

Figure 1: LatSenRom first editions from 1879 to 1940.

Data processing and analysis were carried out using a normalized, morphologically tagged version of the LatSenRom corpus. The normalization process involved the automatic conversion of texts originally printed in an older orthographic tradition that employed the Fraktur script into the contemporary orthography using the Antiqua-based alphabet. These texts comprise approximately 16 percent of the corpus. Morphological tagging allows all word forms to be linked to their base forms (lemmas), enabling comprehensive retrieval across the corpus. Working with lemmas also improves processing efficiency and supports the generation of Word2Vec word embeddings.

2.2 Methodology

The first phase of the study focused on identifying vehicle terms that appear in LatSenRom. Various methodologies could be applied to detect transport-related concepts, such as subjectively hypothesizing vehicle names, compiling vocabularies from dictionaries or research sources dedicated to a particular topic (see the dictionary-based selection method used in Bologna 2020) or manually annotating them within the texts. However, compiling vocabularies subjectively or from external sources risks incompleteness, as it lacks systematic verification, while the manual annotation is highly labor-intensive. To maximize the identification of transport-related terms, the study initially employed a natural language processing technique based on Word2Vec to analyze the texts. These results were later cross-referenced with those generated by the Gemini 1.5 language model, ensuring a more comprehensive and robust analysis.

Although large language models (LLMs) in all likelihood outperform Word2Vec, this is not yet consistently the case for the Latvian language at the time of writing this paper. The performance of LLMs remains notably weaker for low-resource languages and the methodologies employed still require further validation at this stage of model development. A comprehensive comparison of multiple LLMs was beyond the scope of this study, as the primary objective was not the evaluation of methods but the accurate identification of as many transportation-related terms as possible. Preliminary testing, however, identified several state-of-the-art LLMs that demonstrated acceptable performance for Latvian; among them, Gemini 1.5 was selected for the exploration of LatSenRom.6 While non-commercial models would be preferable for research purposes, fully open-source alternatives currently fail to provide satisfactory results for texts in Latvian. Additionally, LLMs in general remain prone to various bias, anomalies, and hallucinations (Törnberg 2024). To mitigate errors, a hybrid approach should be implemented for cross-verifying results. In this study, Word2Vec outputs and human verification were used to cross-examine the findings generated by Gemini 1.5.

The Word2Vec method, widely known since a 2013 study by Google scientists (Mikolov et al. 2013), uses neural networks to create word embeddings. Compared to other early language model-based methods such as GPT and BERT, Word2Vec is directly optimized for finding similarities between concepts. Although Word2Vec’s capabilities are limited compared even to BERT and early GPT, Word2Vec is more accessible for everyday use in research practice as a LatSenRom-sized model using a pre-trained list of lemmas can be trained in a day on a standard personal computer, without the need for a supercomputer or cloud computing solutions. The Python programming language external Gensim library (Řehůřek and Sojka 2010) with built-in Word2Vec support was used in the study.

Two datasets were processed to generate Word2Vec embeddings. The first model was based on the LatSenRom corpus, consisting of 463 documents. After text cleaning and optimization, embeddings for 105,677 lemmas were derived from the raw dataset of over 36 million tokens. However, empirical analysis of the model’s performance indicated that its size and lexical coverage were insufficient to identify a comprehensive list of transport-related concepts. 7 To achieve a broader representation of transport terminology, a second, larger model was developed using Latvian periodicals published between 1920 and 1940. This periodical-based model utilized 172,240 documents (articles) as the primary data source. Unlike the LatSenRom model, the training corpus for this model was constructed by selecting only documents containing the verb ‘to drive’ (braukt) in various conjugate forms, rather than incorporating all available articles from the period. In terms of temporal coverage, the dataset included materials that overlapped with the publication years of LatSenRom, ensuring consistency. The raw data volume of the periodical corpus amounted to approximately 140 million text units, yielding vectors for 565,623 lemmas (the amount of lemmas is exaggerated due to optical recognition errors).

To obtain a comprehensive inventory of vehicle terms, an initial subjective list was compiled,8 and queries containing these keywords were used to extract broader sets of related concepts from the Word2Vec LatSenRom and Periodicals models. For each queried term, a list of the most similar words was generated, applying a similarity score threshold of 0.6. These lists were then manually evaluated. This process identified dozens of land vehicles, including various horse-drawn carriages and mechanized vehicles. Different spelling variants were also discovered, which is particularly important when working with changing orthography and noisy data. In the subsequent phase of the study, the terms identified in the periodicals were employed to locate references to vehicles in LatSenRom. For example, only Word2Vec Periodicals model yielded names of specific car brands, some of which were later found in LatSenRom as well.

To identify vehicles in LatSenRom using Gemini 1.5, only the LatSenRom corpus was utilized. Several prompts were tested to inquire the corpus with Gemini 1.5, including instructions that included the list of keywords that were used for querying Word2Vec models as few-shot examples. The selected prompt prioritized precision and produced the highest number of valid results following data cleaning. The prompt aimed to give clear, structured instructions; it defines the form of input and output data, stipulates the expertise in the Latvian language and transportation and emphasizes that no interpretation and transformation of results is expected; no examples of vehicles are provided:9

You are an expert on Latvian language and transportation. Please extract a comprehensive list of all mentioned land transportation vehicles from the given texts. Show all specific land transportation types found in the text as a list. No other information is needed, just the transportation terms as they appear in the text. Provide the terms only in Latvian, preserving their original transcription variants as they appear in the text. Latvian Text follows:

The prompt proved to be highly effective for identifying vehicles; it could be further adjusted to diminish the number of false positives or terms that do not correspond with the researcher’s hypothetical definition of vehicles.

The initial extraction of terms from the Word2Vec LatSenRom model after cleaning yielded only 31 valid unique terms (in addition to initial keywords);10 Word2Vec results form the Periodicals supplied 127 unique terms (76 later found also in LatSenRom); Gemini 1.5 model results after cleaning yielded 159 unique terms (see Table 1). Notably, the Word2Vec training data due to the nature of the pre-processing generated only single-word terms and compound words. In contrast, the Gemini model identified 57 multi-word expressions that incorporate a base term already present in the vocabulary, typically representing a specific subtype of a vehicle (e.g., car – sports car; railway – electric railway). Therefore, when considering only single-word terms, the difference between combined Word2Vec and Gemini 1.5 outputs is less pronounced. The data cleaning process proved to be similar across both approaches, each exhibiting comparable levels of redundancy. Only 13.8 percent of the Word2Vec LatSenRom results within the similarity threshold (score ≥ 0.6) represented valid terms, compared to 48.2 percent in the Word2Vec Periodicals model, and 11.4 percent in the Gemini 1.5 results.11 While Gemini 1.5 generated a higher number of useful terms, the prevalence of false positives rendered this approach unsuitable for automated annotation without significant methodological improvements. In contrast, the comparatively larger Word2Vec Periodicals model produced the fewest false positives within the given similarity score range.

Table 1: Number and percentage of valid terms per method.

Measure Word2Vec LatSenRom Word2Vec Periodicals Gemini 1.5
Number of valid terms 31 127 / 76 159
Percentage of valid terms 13.8 48.2 11.4

3. Analysis of the Mentions of Vehicles in LatSenRom

3.1 Overview of Quantitative Findings

As shown in the previous section, approximately 160 unique terms and multi-word expressions referring to vehicles were identified in the general scan of the corpus. Regarding the proportion of motorized and horse-drawn transport, approximately 50 percent of the combined Word2Vec findings in LatSenRom referred to horse-drawn vehicles and bicycles, with the remainder being motorized vehicles. In the Gemini results, close to 49 percent of the identified terms referred to motorized vehicles. This suggests that, by the end of the 1930s, the vocabulary related to modern vehicles was nearly as extensive as that associated with traditional horse-drawn vehicles, despite modern transportation still being a relatively recent phenomenon. Nevertheless, references to modern modes of transportation remain comparatively less frequent than those to horse-drawn vehicles in the novels (see Figure 2).

Figure 2: Average relative document frequency of terms referring to motorized and horse-drawn vehicles (all terms).

Mapping mentions of individual concepts is challenging due to the wide variety of forms used to designate the same vehicle: this includes OCR errors, spelling variations, longer and abbreviated forms, as well as literary and colloquial expressions. A number of terms have several meanings, prohibiting simple counting operations. Although the vehicle lists derived from the models were thoroughly reviewed and validated, it is still likely that some terms and variations remain unaccounted for.

When assessing the increase in the frequency of references to specific concepts, it is worth noting that the lack of chronological balance in the corpus results in volatile scores for the relative frequencies in the early years, followed by a substantial increase in absolute counts of mentions in the later years. The absolute document frequency increases significantly over time because of the considerably larger number of editions published: the more novels, the more times all types of vehicles are mentioned (see Figure 3). However, when analyzing a comprehensive corpus that contains all published novels, the apparent increase in mentions should not be dismissed merely as a distortion caused by chronological imbalance. Instead, the absolute document frequency may partly serve as an indicator of the extent to which a concept entered the cultural mainstream, assuming the novel, as a genre, functioned as one of the key vehicles shaping public imagination. Speculatively, it could be hypothesized that, in contrast to lyric poetry and drama, the novel is a medium that captures and documents the everyday life and material culture of an era with particular sensitivity.

Figure 3: Distribution of selected land vehicles across documents. The range of number of novels for each vehicle is provided next to each term.

A detailed analysis of references to urban transport concepts reveals that the automobile and the bus appear in novels around ten years after their introduction to Riga. It remains to be discovered whether these innovations were documented earlier in prose works published in periodicals. At the same time, the corpus of novels reveals that depictions of the material culture in futuristic novels sometimes precedes real-world developments – for example, electric cars have been running on the streets of Riga since 1930s (Paulockis 1938a, Ģirupnieks 1939).

Figure 2 shows aggregated (average) relative document frequencies of terms pertaining to horse-drawn and motorized means of transportation. While horse-drawn vehicles remain dominant in novels even into the 1930s, it is noteworthy that references to both types of transportation at times tend to follow similar patterns of increase and decrease over time. This parallel suggests that some works may simply feature more travel overall. Supporting this hypothesis, examples from the corpus show that horse-drawn cabs and motorized taxis are sometimes mentioned in the same context – typically when characters are deciding or suggesting which mode of transport to use.

Figure 3 shows the absolute document frequency of selected prominent vehicles. It supports the observation that even though motorized vehicles were increasingly appearing in the literary vocabulary in the 1920s and 1930s the horse-drawn vehicles were still mentioned in a large number of works. The increase of references to motorized vehicles starting from the 1920s highlights both the growing prominence of the novel as a literary genre and the emergence of modern life in the newly established Republic of Latvia. Among the most frequently used terms are horse-drawn vehicles such as rig (pajūgs), single-horse carriage (vienjūgs), pair carriage (divjūgs), three-horse carriage (trijjūgs), four-horse carriage (četrjūgs), droshky (droška), line-droshky (līnijdroška), farm wagon (ore), horse-drawn cab (ormanis), wagon (vezums), covered wagon (kulba), spring wagon (federrati), coach (kariete), sleigh (kamanas), sled (ragavas), wain (vāģi). Among motorized vehicles frequently mentioned are railway (dzelzceļš) and train (vilciens), car (auto), automobile (automobile), motorcar (automašīna), machine (mašīna, used colloquially for ‘car’), limousine (limuzīns), bus (autobuss), tram (tramvajs), trolleybus (trolejbuss), taxi or teximeter cab (taksis, taksītis, taksometrs, taksomotors), motorcycle and motorbike (motocikls, motociklets). The modern lifestyle also introduces the use of bicycles (velosipēds, divritenis).

The Latvian word mašīna (machine), now used colloquially to refer to a car, appeared as early as in the first novels, but it was not initially used to mean ‘automobile’. Machines first entered rural households as agricultural machinery, e.g., threshing machines. Since the early 20th century, in the works of many authors, machines referred to factory equipment; as early as in 1899 machines also appear as household appliances – coffee machines and sewing machines (Deglavs 1897–1899). Already in one of the earliest novels the concept of ‘machine’ appears as an abstract notion or metaphor: the novel’s protagonist utters that other people were machines released into the world for the sake of the nobles (Māters 1879).

The first mentions of the word automobilis (automobile) can be found in novels published in 1912 and 1913 (Skuja 1912, Kaija 1913, Upi̇tis 1913). In total, there are 818 mentions of the word automobilis across 157 works by 96 authors. Starting from the 1920s, the shortened version of the word – auto (car) – became increasingly popular. The first mentions of this variant appear in 1921 (Akuraters 1921, Upi̇ts 1921). By 1940, there were 1,637 mentions across 35 novels by 111 authors. The word automašīna (motorcar) was used comparatively less frequently, appearing only from the 1930s onwards, with a total of 40 mentions across 16 works by 11 authors.

When focusing on modern public and hired land transport vehicles, the railway and train undoubtedly play a central role. The railway (dzelzceļš) is mentioned as early as in the first Latvian novel (M. Kaudzi̇te and R. Kaudzi̇te 1879), while the first mention of a train (vilciens) can be found in 1891 (Deglavs 1891). 12

Mentions of trams (tramvajs) appear in novels starting from 1895 (Poruks 1895); in total, the LatSenRom corpus contains 611 mentions of the word tramvajs across 155 works by 89 authors. The word autobuss (bus) appears from 1924 (Skuja 1924). In total, there are 250 mentions across 79 works by 49 authors.

Terms related to horse-drawn cab (ormanis) remained relevant up until 1940. Mentions of it can be found in novels starting from 1892 (Purapuķe 1892). In total, there are 1,000 mentions across 183 works by 95 authors.

The search for mentions of taxi is complicated by the various forms in which the word is written. The earliest form, chronologically, might be taksomotrs (Skuja 1924). This term appears a total of 19 times across nine works by five authors. Alternative term taksometrs (also spelled taksametrs) has been mentioned since 1925 (Gulbis 1925). Overall, there are 98 mentions across 47 works by 35 authors. Forms like taksis and taksītis are also encountered. Similar to the case with vilciens (train), tracking mentions of these terms is more challenging due to polysemy. The first mention of taksis appears in an anonymous author’s work (Anonymous 1926); the word is used a total of 66 times across 33 works by 27 authors. The term taksītis first appears in 1927 (Erss 1927); it is used a total of 57 times across 25 works by 19 authors (excluding uses with non-relevant meanings).

The bicycle is also increasingly mentioned as cycling was a cheaper than the horse or the car mean of personal transport and a popular leisure activity. In the journalism and fiction of the period, it was also given a role as a symbol of the New Woman and the abolition of fin-de-siècle gender restrictions. References to the bicycle date back to 1899, in August Deglavs’s novel The New World (Deglavs 1897–1899).

Combined Word2Vec and Gemini 1.5 results revealed also less frequently used words, including specific car brands and types: Ford, Mercedes, Chevrolet, Rolls-Royce, Oldsmobile, Alfa Romeo, Buick Essex, sports car, roadster etc. which indicates the consolidation of the role of modern mechanized transport in culture in the late 1920s and 1930s.

An analysis of references to vehicles reveals two key insights. Firstly, as the 1940s approach, the vocabulary of transportation extracted from Latvian novels is nearly evenly divided between terms for horse-drawn and motorized vehicles. Secondly, the ranking of the frequency of mentions of transport includes novels and novelists whose names are less well known in the history of Latvian literature or who have been marginal in the genre of the novel, providing an opportunity to get to know several of them anew, such as Miķelis Paulockis, Kārlis Lapiņš or Ansis Gulbis. This analysis of the large corpus of novels and full-text data is thus not based on canonical texts; on the contrary, it offers a democratic and unencumbered-by artistic quality criteria approach to analyzing a particular phenomenon in a large corpus of texts.

Figure 4 shows the authors whose works mention popular vehicles the most (the height of the bars represents the total number of mentions of a given name, summed over all the authors’ works).13 Although dozens of authors have mentioned various vehicles, most references are brief, and only for a few does a vehicle become an essential part of the message. For example, ‘car’ plays a prominent role in the works of Miķelis Paulockis, Jānis Sārts, Vilis Lācis, Kārlis Miltiņš, and Arvīds Vanags, while ‘automobile’ is most often found in the novels of Ansis Gulbis, Aīda Niedra, Andrejs Upītis, and Kārlis Ieviņš. Modern vehicles such as the bus and the motorcycle have not become the ‘main characters’ in any of the works of the LatSenRom corpus, i.e. they are represented with a much smaller number of mentions, but it is interesting to observe that they have been important for several authors who have also written about cars and automobiles, such as Vili Lācis, Miķelis Paulockis, Jānis Sārts, Andrejs Upītis.

Figure 4: Word frequency of selected vehicles: top five authors with the highest mentions for each term.

3.2 Towards Mechanized Transport: Riga as a Baltic Metropolis

Observations on vehicles entering literature at the end of the nineteenth century point to phenomena that each warrants further study and analysis that exceed the capacity of an article – only the taxi and horse-drawn cab were selected to illustrate the rich semantic load embedded in the literary representation of transportation.

As Steven A. Mansbach has observed, Riga was “the only true metropolis in the Baltic region,” where “the forces of history and modernism reached an accommodation for northeastern Europe” (Mansbach 2014, p. 261). This metropolitan status was not only symbolic or architectural, but also rooted in Riga’s unique role as the industrial and technological hub of the region. At the beginning of the 20th century, Riga was home to several key players in the automotive industry, including Alexander Leutner & Co., the Rossiya factory, and the Russo-Baltic Wagon Factory, which produced the first automobiles in the Russian Empire. During the interwar period, Riga further solidified its position as a regional automotive center through the Ford-Vairogs plant, which operated under license from the Ford Motor Company. The presence of these factories meant that automobiles – particularly taxis – were more readily available in Riga than elsewhere in the Baltics, which explains their early literary visibility as markers of urban modernity and social mobility.

In parallel with this industrial infrastructure, the establishment of the Supreme Board of Roads and Structures of Latvia in 1919 introduced a regulatory framework that fundamentally altered the conditions of mobility. This institution oversaw road development, enforced traffic regulations, and issued licenses, thereby transforming the automobile from a luxury item into a managed element of everyday urban life. The increasing presence of legal norms, signage, and speed control in the public space paralleled the narrative tension visible in Latvian fiction, where taxis often appear as symbols of not only acceleration and freedom, but also of risk, collision, and bureaucratic control.

With the development of modern urban spaces, in Riga and other European and Russian metropolises, the landscapes, sounds, scents, and transport flow became an integral part of life, even for those who did not use this transport daily. This experience was present both day and night. It became a significant element in literature that has increasingly turned its focus to the urban environment, encompassing and reflecting the impact of transport on the urban landscape and its inhabitants. Old habits persisted into the transitional decades around the turn of the century, just as horses continued to trot through the streets. Yet, new mobility habits and transportation systems also emerged.14 Street traffic is precisely depicted in Jānis Veselis’ novel Dienas krusts (1931) through the character Mežaks riding a bicycle through Riga:

The evening twilight, bustling with shop assistants coming home, shimmered and sparkled in the glow of illuminated windows in the rushing lights of trams and buses. Where shadows darted, met, and parted at every street corner, Mežaks rode his bicycle to the workers’ sports ball. He deftly weaved between the large buses, speedy taxis, and slow pedestrians, sometimes whistling, ringing a bell, braking, and others speeding up. With varying speeds, resembling a perilous collection of beasts, he maintained harmony. He felt like a bird in the air: the true residents of the street were his friends, and the high curve of the street was all alike, from the ten house owners, deputies, and ministers to the poor courier, shoemaker, seamstress, and dockworker. Only those who referred to a bicycle or car as their own or those using it had special rights: they could. (Veselis 1931)

Anchoring the literary data in this industrial and administrative context allows for a richer interpretation of how novels mediate technological change as both experience and imagination. The material culture depicted in novels partly reflects developments in the author’s immediate surroundings, but it can also represent advancements that have not yet reached their homeland or even prefigure future changes. The first automobile appeared on the streets of Riga in 1904; in the novel, it was first mentioned in 1912, in an episode set in Moscow (Skuja 1912); further references to automobiles from books published in 1913 are situated in Riga. In Paris, automobiles appeared only slightly earlier – in 1896 – underscoring that Riga has always been a modern city and everything new in Europe arrived in Riga (then under the rule of the Russian Empire) with only a slight delay (Stirna 2024).

Omnibuses, trams, and buses represent large-scale public transportation, typically carrying sizable crowds of passengers, which was a new development in the context of inner-city mobility (Biedriņš 2021). The predecessors of trams and buses – horse-drawn omnibuses – first appeared in the streets of Riga in 1852 (Budiloviča 2024) and were also mentioned in the first Latvian novel (M. Kaudzi̇te and R. Kaudzi̇te 1879). In the following decades, the term appeared in only seven works, with its last mention in Pāvils Rozītis’s novel in 1936, where it is portrayed as a crude, dirty, and slow vehicle – so much so that a girl steps out of it to avoid staining her dress (Rozi̇tis 1936). When considering the modes of transport that carried citizens through the streets of Riga, much more attention in novels is given to the tram. The inception of tram services dates back to 1882, when the first horse-drawn trams appeared on the streets of Riga. The first electric tram line was launched in 1899 in Liepāja; in 1900, the Riga Trams Joint Stock Company was established, and in 1901, Riga’s electric tram services were launched (Budiloviča 2024). The tram has been mentioned in Latvian novels since 1895, for the first time in the narration that takes place in Dresden (Poruks 1895). Further references, starting from 1900, are likely situated in Liepāja and Riga. In addition to trams, it is worth noting the arrival of the bus in Riga in 191315 – though it is mentioned in novels only after World War I.

In rural settings, the dominant modes of transport until the First World War were on foot, horseback, or carriage, and these modes are often used to emphasize the deep connection between the characters and the landscape and to open up space for personal reflection and a deeper connection with the environment. Rural modes of transport emphasize a closer connection to the land and a slower pace of life. Horse-drawn carriages are mainly used for agricultural purposes and transporting goods, showing the hard work and traditional way of life of rural life, which is increasingly threatened by the encroachment of modernization; for example, trains often play a different role in rural settings than in urban narratives. While in cities, they symbolize connectivity between distant places and even countries and progress, in rural areas, trains can symbolize the encroachment of the industrial world. They can create both opportunities and significant disruptions to traditional rural communities. The arrival of railways in remote regions is a technological achievement and a catalyst for social change, affecting the social fabric and physical landscape of the rural areas where these stories are set.

Just to briefly touch upon other means of transportation, water transport on rivers and the sea in novels has always symbolized discovery, travel, personal transformation, and confrontation with the forces of nature.16 The depiction of air transport in literature appeared later and rarely (until the turn of the 20th century). After the First World War, airplanes symbolized the new technological challenge. It is only natural that most novels written in Latvian refer to airplanes and aero-planes as military means of warfare. Still, also small sports planes are mentioned in The Silver Sun Leaps by Skuju Frīdis (Skuja 1924), and one striking mention of the airplane as a private vehicle is in a feminist utopian novel by Amanda Kloto’s The Victory of Woman (Kloto 1927), where all the women of the world fly to an extraordinary women’s congress in airplanes or, as the novel has it, aero-planes. Another example, notable for its time, is Miķelis Paulockis’s science-fiction novel The Saviour of the World (Paulockis 1938b). It anticipated technological developments, including constructing a subway in Riga in 1986, which never happened.

3.3 Symbolism and Social Context of Transport

Transport in European and world literature is an essential narrative element and a capacious symbol of the age’s social, cultural, and technological changes. Transport scholars have used literary and historical methodologies to explore horse-drawn carriage travel, the waiting associated with rail travel, and the impact of the train window on passengers’ perception of the landscape as ’another world (Kellerman 2019, Livesey 2016).

In addition, works on transport history and literary mobility assess the role of the horse in the city in the 19th century and the changes in perception brought about by rail travel (Gavin 2015, McShane and Tarr 2003).

Horses were vital to industry, commerce, agriculture, and employment, serving other transport systems, including rail and shipping. In addition to the omnibus, taxi, carriage, cart, and coal wagon functions, horses were used for deliveries to stores, postal services, public holidays, refuse removal, emergency services, wedding parties, and funeral ceremonies (Gavin 2015). With trams, which were initially horse-drawn and then electricity from the 1880s onwards, and rails appeared on city streets, trams were a space where people from different social classes could physically meet and appreciate each other (Finch 2023).

Barbara Schmucki has shown that urban dwellers in Britain and Germany had different views and relationships with the horse tram and its electrified descendants, sometimes viewing newcomers with distaste for the associated rails and electric wires. However, “by the 1950s, trams were fully integrated into everyday life” (Schmucki 2012).

The appearance of trains and railways in the literature often marks the development and mobility of society and symbolizes progress. Trams and buses capture the rhythm of everyday urban life, highlighting the interaction between different social classes and the everyday life of urban dwellers. This sense of mobility was particularly characteristic of the turn-of-the-century youth who came to the metropolises: “They are brought to and transported around the capital by the new mobility offered by the rail network and trams, and it is their dreams of upward social mobility that drive them forward.” (Ameel 2014) In his study of urban experiences in Helsinki (1890–1940), Finnish comparative literature scholar Lieven Ameel points out, “In many important Finnish literary texts dealing with urban experiences in this period, modes of transport and images of mobility acquire more than a symbolic status. They are central to the development of provincial characters in the city.” (Ameel 2014)

An analysis of specific period novels, such as those by Virginia Woolf, James Joyce, and F. Scott Fitzgerald, reveals how modern transport and urban mobility function not only as realistic markers of modernity but also as narrative devices and socio-cultural signifiers. For example, the role of the omnibus in Virginia Woolf’s novels is explored on several levels: as a realistic indicator of technological progress, as a cultural commentary on the British class system, and as a device of Woolf’s narrative technique. Having frequently traveled in both horse-drawn and later motorized omnibuses, Woolf knew the colors, routes, and fares of the omnibuses that made up London’s route network (McNees 2010). The different modes of transport offer glimpses into a public sphere where anonymity and observation intertwine, as in Virginia Woolf’s depiction of London, where buses allow characters to traverse and observe the city’s diverse social landscapes.

The period under review witnessed unprecedented advances in various forms of mechanical transport and related technologies, each of which introduced new dimensions to the plot and thematic setting of the novels, which can also be observed in the analysis of the LatSenRom corpus, for example, the technological development is accurately described in Jānis Plaudis’ novel Gymnasium students:

If a hundred years ago, one was still dreaming of the revolving balloon and relieving pain through surgical operations, now one flies like a bird in an aeroplane and begins to think about the questions of eternal youth. (Plaudis 1935)

As the hero of the novel by Vilis Aizstrauts predicts:

Technology is increasingly making its mark in life. The time is not far off when humankind will do almost all its work with electricity. Electric plows will drill in the soil, and electric airplanes will roar in the air; ships and trains will be powered by electric batteries instead of coal. In short, electricity will rule the world. The question is how to get the power we need most cheaply and conveniently. Along these shores, millions of electrical energy go unused. (Aizstrauts 1933)

Alongside the excitement about the new possibilities, there was also criticism and nostalgia, for example, a quote from Augusts Mežsēts’ novel The Enchanted City (1929), which reflects a critique of technological progress and a fear of being overwhelmed by the possibilities of the new technologies:

It is true that in our days when people are jolting around the earth in cars and aeroplanes are lifting them into the air, human thought doesn’t need to fly. It is enough for the human body to fly. Young lovers wander in armchairs on boulevards, gaze into electric light bulbs, or sit in cinemas and marvel at the enchanting love that fades on the screen in an hour. (Mežsēts 1929)

4. The Public Transport System in Riga. Horse-drawn Cab vs. Taxi

As highlighted in the introduction, exploring urban transport in the literature provides a fruitful perspective for understanding sociocultural dynamics. This aspect allows for the revelation of not only the technological achievements of its time and the transformations of the urban environment but also a deeper insight into the daily lives of individuals and the social and psychological landscape that emerged during these changes, transitioning from the 19th to the 20th century.

Compared to trams and buses, horse-drawn cabs and taxis represent more private forms of public transport, typically serving just one or a few individuals at a time. An important aspect is that taxis gradually replaced the legendary horse-drawn cabs, but a particularly notable period was the 1920s when both types of vehicles coexisted. Figure 5 presents the relative document frequency of references to horse-drawn and motorized cabs, with the various terms used for each type consolidated into two respective lines. While the yellow line prevails, it is gradually declining, whereas the blue line rises steadily, signaling the arrival of a new era. Therefore, examining the transition from the once-popular horse-drawn cab to the mechanized taxi in early 20th-century Latvian novels is particularly fascinating.

Figure 5: The proportion of novels in which the terms ‘horse-drawn cab’ (including terms ormanis, fūrmanis, važonis) and ‘taxi’ (including terms taksis, taksītis, taksometrs, taksometris, taksomotors) appear in different years.

‘Taximeter’ is the key term for this transition, as its original meaning is ‘mechanical meter’, which was attached to the horse-drawn carriage to count the distance traveled and thus calculate the fare.17 It was introduced due to detected fraudulent practices by drivers to obtain larger payments. The introduction of the taximeter as a fare-measuring device at the end of the 19th century brought a new element to private and public transport, ensuring fair pricing based on distance traveled or time. This innovation was crucial for building trust and standardizing tariffs in the growing urban centers. The first taxi cars appeared in Riga in 1913; however, their numbers surged after the First World War, particularly during Latvia’s period of independence, when they also began to appear frequently in novels (for counts, see subsection 3.1 in this paper).

As mentioned in the introduction, the case study was inspired by the question asked in Kārlis Lapiņš’s novel Students in the Farm: “Why do you drive a 19th-century carriage when you have a 20th-century taxi at hand?” As cities expanded and modernized, the transition from horse-drawn carriages to motorized taxis marked a significant development in public transport and a new relationship with speed, 18 reflecting changes in urban lifestyles and technological advancement. Therefore, a brief insight into how this is revealed in the corpus of Latvian novels.

4.1 The Legacy of the Horse-drawn Cab in Riga

The horse-drawn cab (ormanis), as historian Andris Caune notes, “is the oldest form of public transport in Riga,” (Caune 2020) is mentioned in the historical sources of Riga since the 13th century. The horse-drawn cab as a romanticized symbol of Riga is vividly portrayed in Latvian poetry, notably by Aleksandrs Čaks with several poems as Pathetic Quartets and Poem about the Horse-drawn Cab (Čaks 1930). And also in novels, for example, in Jānis Jaunsudrabiņš’s novel Aija, the character Jānis, while in the countryside, occasionally feels a sentiment for the urban identity and recalls “the beautiful streets of Riga on winter evenings. The bells of the horse-drawn cabs tinkled. Illuminated by electric lights, people came and went.” (Jaunsudrabiņš 1911) Thus the horse-drawn cab is a striking figure associated with the city (see Table 2).

Table 2: Comparison of narrative and symbolic functions of ormanis and taxi.

Feature Ormanis (horse-drawn cab) Taxi
Narrative Function Human-centered, emotional depth Dynamic, anonymous, fast-paced scenes
Symbolism   Tradition, nostalgia, rootedness   Speed, modernity, societal transition
Social Class Reflection Often tied to working or poorer class Emerging middle class, upward mobility
Temporal Feel     Slow, contemplative, historic     Fast, fragmented, urban modern
Driver Role     Named, empathetic characters     Anonymous, functional presence
Gendered Dynamics Personal, respectful, often passive Can involve erotic tension, autonomy
Weather Vulnerability     Open to cold, wet, discomfort     Protected, modern convenience

Often, the most significant role in novels belongs not to the transport itself but to the driver of the horse-drawn carriage, as seen in Andrejs Upīts’s novel The Last Latvian (Upi̇tis 1913), which highlights the human characteristics of the carriage driver through various aspects. In Upīts’s novel, the carriage driver is more than just a driver; he observes and reacts to Kalve’s emotional state, becoming a silent witness to his passenger’s inner turmoil. The episode of the novel vividly depicts the bustling and somewhat chaotic urban environment of that time, highlighting different societal attitudes and behaviors through the characters’ interrelations, particularly around the figures of the carriage driver and Alberts Kalve. The novel’s action takes place in a traffic jam caused by a tram and two timber-carrying carts, a common occurrence on the busy streets of a growing city. This obstacle provokes a series of reactions from Kalve, revealing not only his impatience but also a more profound commentary on the rhythm and priorities of urban life. Kalve’s impatience and the subsequent interaction with the tram driver and timber carriers reflect the differences and expectations of social classes regarding service. In this scenario, the ‘carriage driver’ acts not only as a service provider but also as a participant in the broader dynamics of society. His initial inaction and then hasty reaction after Kalve’s prompting with a cane emphasize the power dynamics and the expectations placed on individuals based on their roles in society. However, Kalve also perceives something unacceptable in the general attitude towards carriage drivers, namely that a carriage driver must wear a number on his back at the collar, even though the number is visible on the back of the cart, at the front by the pole, and on the sides. This numbering system is part of a strict control and identification mechanism. Kalve, describing this situation with its mechanical and depersonalized approach, ironically suggests that a person should wear a number plate of their home address on their back to conform to this absurd system fully.

Conversely, the episode in which the carriage driver is urged to pick up the pace with a hit of the cane (or another object) is quite typical in accounts of carriage drivers, for example, in Andrejs Upīts’s novel Gold, where the young woman Made is particularly impressed by the journey with the carriage driver alongside her brother, who has come into wealth. Both how Sveilis, with his cane, touches the carriage driver to halt the carriage and how he exclaims: “You, man, wait: I don’t have small change!” (Upi̇ts 1921)

This expression seems astonishing to Made, as it sounds so gentlemanly, contrasting sharply with her previous poverty and experiences in impoverished conditions. Such moments of superiority are particularly striking in the portrayal of the carriage driver, which is not typical in accounts involving taxi drivers. It is noteworthy that carriage drivers are also much more affected by weather conditions, as, for the most part, the driver’s seat is not equipped with a roof or any other form of protection against rain, cold, or heat, which impacts the comfort of the journey. For instance, in Ivande Kaija’s novel In the Yoke, the carriage driver humbly asks a passenger: “Madam, are we going to roam around the city for long? I am starting to feel cold; the horse also needs a rest.” (Kaija 1919)

Modern vehicles, which replaced horse-drawn transport, were considered by some as peculiar and indifferent machines that threatened to change human character as well. A significant example of generational clashes is in Kārlis Lapiņš’s novel The Degenerate (Lapiņš 1934a). One morning on Barona Street, a distressing event occurs. The father, Pumpītis Senior, moves slowly along the street with a horse-drawn cab when suddenly a taxi appears from behind and swiftly overtakes them. Pumpītis looks suspiciously at the taxi, in which a familiar wide-brimmed hat seems to flash. He instructs the horse driver to urge the horse to try to catch up with the taxi, but it proves unsuccessful, and the car disappears around the corner, continuously emitting loud signals. Pumpītis hisses, dissatisfied with what happened, as he suspects the taxi passenger could be his son. When they finally arrive home, the driver stops beside the taxi. Seeing his father, the son greets him with a casual “Good morning, father,” but Pumpītis does not respond. He feels uncomfortable and angry, realizing that this is neither the place nor the time to discuss what has transpired as both father and son return from a night out.

4.2 The Rise of the Taxi and Urban Modernization

The appearance of taxi in literature is negligible before 1925, which aligns with the historical introduction of automobiles in Riga. From the mid-1920s onward, however, its presence grows steadily. By the early 1930s, taxi achieves narrative parity with ormanis, and in the years leading up to the Second World War, it slightly surpasses it. This pattern corresponds to the accelerating pace of modern urban life and the shifting aesthetic of literature toward speed, anonymity, and urban dynamism (see Table 2).

In Pāvils Rozītis’ novel Ceplis (Rozi̇tis 1928), the life of Riga in the spring of the 1920s is vividly depicted, highlighting how the changing weather conditions influenced the use of transport. Snow was often followed by windy rain, making the streets muddy. The drivers, who tied their horses to sleighs in the mornings, were forced to switch to carts by noon, only to replace them with sleighs again in the evening. This uncertainty and frequent transition from one mode of transport to another caused inconvenience for both the drivers and their clients. The horses appeared dirty and drenched, reflecting the prevailing gloomy atmosphere of the city. Meanwhile, the few taxis moved with a particular enthusiasm, splashing pedestrians, building walls, and shop windows.

Riga’s residents mainly used taxis when intoxicated, when money was no object, and life was wanted to be experienced at a faster pace than the cab’s leisurely journey. Only the wealthiest or those poisoned by the desire for life’s accelerating tempo would ride taxis sober, relishing the carefree speed that allowed them to race past the sleepy gray façades of buildings and dash across the entire city in mere minutes. However, there were not many of these people in Riga, just as the wealthy were still in the process.

This portrayal frames the taxi as a modern, dynamic, and somewhat extravagant choice, contrasting with traditional and slow horse-drawn transport, and offers insight into urban life and how technological change reshaped habits and lifestyles. The taxi symbolizes the changing pace of urban environments and the restructuring of society. “Limousines and taxis race along Freedom Boulevard; it is the aorta of the city” (Prūsa 1936) – this is how Emily Prūsa describes the new urban landscape in her novel The Temptation of Distance (1936). The taxi provides a notion of the car for those who have not yet traveled in one; this joy of the ride is described in Jānis Veselis’s novel The Uprising of People, where Meiklis invites guests for a ride in his new car:

The car purred charmingly and pleasantly; the wheels began to roll across the paved yard and rolled out onto the street. Meiklis, who had rarely driven in a taxi, and Lija, who had rarely enjoyed the delights of such a journey, both felt a rocking pleasure as the rubbery wheels swiftly whisked them through the already cobbled streets, racing dreamily past houses, lampposts, people, and autumnal smoky flames, as if all of that were a life to be left behind irrevocably. (Veselis 1934)

The city description seen through the window of a moving car tends to emphasize the experience of speed and the fragmentary and ephemeral nature of observation. Meanwhile, the taxi driver’s interaction with clients can reveal urban dynamics and themes of anonymity. The cultural significance of the taxi extends beyond literary representation and is embedded within the broader mediascape of the early twentieth century. In addition to novels, the motorcar emerged as a prominent visual and narrative motif in early cinema. It appeared frequently in dramatic scenes, including chases, romantic encounters, and depictions of urban glamour or danger. These cinematic representations amplified the symbolic associations of the taxi with speed, risk, and emotional intensity.

Like horse-drawn carriages, novels mention taxis in scenes involving personal reflections, character interactions, and dynamic movement. The taxi serves not only as a means of transport but also as a place for character and plot development. Taxis appear most frequently in Miķelis Paulocki’s novels. In Paulocki’s works, taxis fulfill various narrative functions, from the movement of characters to significant interactions or places for contemplation; they symbolize aspects of modernity, urban life, or transitions, reflecting the characters’ experiences and broader societal changes of that era. In Paulocki’s novel The Secret of the Old Lighthouse, it is notably revealed that the new speed provided by mechanized transport, such as the taxi, does not always fulfill the characters’ wishes. The protagonist, Krauze, needs to reach Zasulauks pier; therefore, he begins his journey with a horse-drawn cab. However, although he asks to go faster, the horse can only tread slowly in the small cart, causing Krauze’s dissatisfaction. Feeling the pressure of time, Krauze decides to switch to a taxi to speed up his journey. This transition symbolizes not only a physical movement from one vehicle to another but also a shift from the slow pace of the old world to the new – faster – tempo of urban life. Krauze, who despised taxis, was nonetheless compelled to use this new mode of transport. The taxi ride at the corner of Kalnciems Street suddenly turns into a catastrophe as a car collision occurs. Krauze is severely injured and understands that his hope to escape his previous life and the problems related to drinking and loss of property has been shattered. His attempt to leave the city and find a new beginning in the countryside has stalled because the city’s asphalt, symbolically speaking, “held him tightly in its firm grip.” (Paulockis 1936)

As part of the new urban aesthetics expressed in the literature of this period, a new central urban experience appears – the night metropolis, seen through the window of a moving car. In several novels and stories, driving at night is depicted as a liminal space that is partially private and partially public. It serves as a space where boundaries are transgressed with a clear sensual and sexual undertone. In such scenes, the enclosed car, much like the night city as a whole, acquires a sexual ambiance; for instance, in Ansis Gulbis’ novel The Legacy of the Fantasts, a taxi ride is used as a place for expressing intimacy and passion. Romina and Jeremejevs, sitting in the taxi, are physically close to one another, creating an intimate atmosphere. Jeremejevs touches Romina’s leg, generating a subtle erotic tension (Gulbis 1925). This moment, occurring within the confines of the automobile contrasts with the anonymity and openness of the outside world, where they are not hidden from view, highlighting the car as a location where private and public boundaries blur, becoming a place for sensual and emotional expressions.

Meanwhile, in an episode from Miķelis Paulockis’ novel Professor Sūna’s Wonderful Elixir, the city and taxi serve as a backdrop and symbol for Herta’s inner tension and yearning for life. The refreshing breeze of a late autumn day contrasts with Herta’s feverishly hot face, revealing her restlessness and desire to enjoy life. Herta’s internal dialogue, where she acknowledges her “mad cravings” and her wish to live and revel in her beauty, directly references her freedom and self-confidence. The taxi she hails becomes a symbol of her hurried pace of life and inner anxiety. The driver, who looks at Herta with surprise and desire, offers her a self-affirming acknowledgment of her beauty and attractiveness. Herta’s rush and desire to escape the routine of everyday life and return to the enjoyment of life are reflected in the motion of the taxi:

The taxi sped madly down the boulevard, crossed paths with horse riders in front of its snouts, and splashed pedestrians with mud from puddles still collected from the recently fallen rain. It seemed as though Herta’s anxiety had overtaken the lifeless machine, rumbling in its cylinders, hissing as the tires rubbed against the asphalt. Faster! Faster! (Paulockis 1938b)

The episode reveals the intensity of urban life and Herta’s crisis, using the taxi ride as a symbol of her inner turmoil and yearning for life.

As seen from the example of the novel, while the horse-drawn carriages evoke nostalgia for a slower, gentler era, taxis symbolize the dynamic, constantly changing nature of the modern city at that moment. These vehicles and their drivers can be either creators or resolvers of conflict, as their interactions with other characters in the novel and cars can lead to significant plot twists, revealing the characters’ traits and motivations. Furthermore, they serve as a metaphor for the irreversible transition from tradition to progress, accurately reflecting the complexity of the relationship between nature and humanity in modernization. The shift from the horse as a living being to the engine as a lifeless technology illustrates technological progress and more profound changes in human relationships with nature and technology.

5. Conclusion

The LatSenRom corpus expands opportunities for literature researchers by providing access to vast datasets and facilitating more effective text exploration. This resource supports both quantitative and qualitative research, enabling comparisons across authors and tracing the evolution of literary motifs and themes over time. Additionally, digitized texts make it possible to uncover lesser-known authors and works while fostering interdisciplinary research that blends literary studies with history and sociology. This integrated approach enriches our understanding of cultural and societal dynamics.

The use of Word2Vec and Gemini 1.5 language models led to the discovery of over 160 unique terms related to land vehicles, with approximately half referring to modern transportation and the other half to horse-drawn transport. The Gemini 1.5 model proved to be the most effective in identifying terms within texts compared to other techniques. However, even the results from the most productive prompt contained approximately 88 percent false positives. Although approximately half of the transportation-related vocabulary refers to motorized vehicles, the average relative document frequency remains higher for horse-drawn vehicles throughout the entire period studied. An examination of terms referring to horse-drawn cabs and taxis reveals a clear trend: mentions of horse-drawn transportation steadily declined throughout the 1920s and 1930s, while references to motorized taxis increased. The analysis of vehicle term frequencies provides a valuable evidential basis for conjectures about the popularity of various modes of transportation and their distribution across authors.

The portrayal of horse-drawn cabs and taxis in literature during the first decades of the 20th century offers a nuanced insight into the evolution of urban public transport and its impact on social interaction and urban experience. These vehicles facilitate physical movement and promote a deeper exploration of themes such as anonymity, modernity, and the individual’s place within the urban landscape. The cabs symbolize mobility, bridging distances and seemingly connecting parts of the narrative, often serving as physical and metaphorical means of transition that facilitate characters’ personal transformation and development. During the interwar years, both transport forms coexist in literature, mirroring urban reality. This overlap period is particularly rich for symbolic readings: it is the moment when characters must choose between the familiar ormanis and the modern taxi, often with ideological undertones.

In several interwar novels, similar to the modern poetry of that time, the theme of speed and technology stands out, serving as a metaphor for the challenges of human existence and the changes in the pace of life during modernity. This observation indicates the role of technology as a response to existential questions, making it an important aspect of analysis that provides additional interpretive possibilities in literature. In Latvian literature, like in Europe, transport is closely associated with mobility, freedom, and change themes.

In sum, the quantitative patterns of term usage in Latvian fiction correlate closely with the qualitative shifts in narrative representation. The ormanis and the taxi serve not only as transport modalities but as literary devices charged with ideological and emotional meaning. Their respective presences in novels mark the changing pace of life, class dynamics, gender roles, and psychological atmospheres of the city. The transition from horse-drawn carriages to taxis thus reflects a deeper literary modernization, in which vehicles become symbols of cultural acceleration, disruption, and the reconfiguration of the self in the urban milieu.

Furthermore, urban studies provide a perspective for identifying how urban transformations, interactions among different social groups, and the impact of technological progress on city life are reflected in literature. Themes of alienation, progress, and public and private space are frequently explored in the novels of this era, particularly through the prism of urban transport. The expansion of the range of urban vehicles is depicted not only as a technological triumph but also as a source of existential anxiety and social fragmentation. The historical context of rapid urbanization and events leading up to World War II imparts a sense of urgency and transformation to these themes.

6. Data Availability

Data can be found here: https://dom.lndb.lv/data/obj/1554666.

7. Software Availability

Software can be found here: https://github.com/ValRCS/lnb_transports/releases/tag/v1.0.0 (DOI: https://doi.org/10.6084/m9.figshare.31156879).

8. Author Contributions

Eva Eglāja-Kristsone: Investigation, Writing - original draft

Anda Baklāne: Investigation, Writing - original draft, Writing - review & editing

Valdis Saulespurēns: Software, Visualization

Notes

  1. In this study, the term ‘horse-drawn cab’ (in Latvian ormanis) will be used to refer to carriages commonly employed during the 19th century for passenger transportation. This term refers to vehicles such as the ‘hansom cab’ and ‘hackney carriage’, which were widely used in urban areas of the period. [^]
  2. In the context of Latvian studies of literary geography, Zita Kārkla and Eva Eglāja Kristsone have studied the geographical places in women’s prose fiction; see Kārkla and Eglāja-Kristsone 2022. [^]
  3. The corpus analysis platform of the National Library of Latvia: https://korpuss.lnb.lv. [^]
  4. An earlier stage of this research that did not include the usage of the Gemini 1.5 model was documented in a publication in Latvian (Eglāja-Kristsone et al. 2024). [^]
  5. For a more detailed account of creation of the corpus, see Baklāne et al. 2024. [^]
  6. Comparative analysis of GPT-4o and Gemini 1.5 results was presented at the DHNB 2025 conference; see: Baklāne and Saulespurēns 2025. [^]
  7. No definite guidelines exist to indicate what size of the corpus is sufficient for acquiring satisfactory results for various tasks performed based on Word2Vec embeddings, however, the larger size of the training data is known to increase the performance (Rodman 2020). [^]
  8. Initial list: horse (zirgs), railway (dzelzceļš), horse-drawn cab (ormanis), rig (pajūgs), coach (kariete), sledge (kamanas), bicycle or velocipede (velosipēds), machine (mašīna), motorcar (automašīna) train (vilciens), locomotive (lokomotīve), tram (tramvajs), dinky line (bānītis). [^]
  9. The best practices of formulating instructions for language models are discussed in Törnberg 2024. [^]
  10. Only tentatively, can we call all non-valid terms false positives. The range of terms considered relevant in this study is highly dependent on the arbitrary boundaries established for its scope. For instance, it remains debatable whether various types of horses should be included within the semantic domain of transportation studies. In addition to the terms considered valid in this study, the results from all models included other transportation-related terms, such as those referring to air and water transport, vehicle parts, infrastructure elements, drivers, passengers, and others. [^]
  11. The percentages reported for the Word2Vec and Gemini 1.5 results are not directly comparable, as the lists were generated using different methodologies; however, these figures serve a descriptive purpose. [^]
  12. A quantitative analysis of the concept ‘train’ is complicated by the fact that the Latvian word vilciens primarily means ‘train’ (a railway vehicle), though it also has a secondary, metaphorical meaning derived from the root ‘vilkt’ (to pull or draw). This figurative usage refers to a motion, stroke, or sweeping action and appears in expressions like elpas vilciens (‘the flow or rhythm of breath’) and zīmuļa vilciens (‘the stroke of a pencil’). These uses evoke the idea of a continuous or pulling motion, adding depth and poetic nuance to the term. [^]
  13. Although relative frequencies are more appropriate for studying the importance of a particular term in an author’s oeuvre, the absolute frequencies used here offer insight into the extent to which an author may have hypothetically exposed their audience to specific vocabulary, whether in the course of a single work or across several. With relative frequencies, this signal would be subdued for authors with several comparatively long works. [^]
  14. The most comprehensive source for understanding Riga’s historical transport system is Andris Caune, Riga’s Transport 100 Years Ago (Caune 2020). [^]
  15. The news that a bus route was opened between Sarkandaugava and Jaunmīlgrāvis can be found in the 1913 issue of the newspaper Dzimtenes Vēstnesis (Dzimtenes Vēstnesis 1913). [^]
  16. The Gemini 1.5 inquiries identified 46 maritime and 30 air transportation terms. These domains were not researched further in this study. [^]
  17. “The city council has just ordered six taximeters from abroad for horse-drawn cab drivers, to whom they will sell them. A taximeter is a device attached in a certain way to the carriage, showing how far has been traveled.” See: Baltijas Vēstnesis 1902. [^]
  18. For the concept of speed in interwar modernist poetry in connection with European metropolises (Ostups 2024). [^]

References

Aizstrauts, Vilis (1933). Mi̇las un ārprāta varā. Latgrāmata.

Akuraters, Jānis (1921). Pēteris Danga. A. Gulbja apgādi̇bā.

Ameel, Lieven (2014). Helsinki in early twentieth-century literature: urban experiences in Finnish prose fiction 1890–1940. Studia Fennica Litteraria, 8. Finnish Literature Society/SKS.

Anonymous (1926). Kapteinis Tālivaldis un viņa bri̇nišķi̇gie piedzi̇vojumi jeb No kuģapuikas par kapteini. Viļņi.

Baklāne, Anda, Artis Ozols, and Eduards Skvireckis (2024). “Latviešu senāko romānu atkārtotie izdevumi – datu kopas izveide un anali̇ze”. In: Proceedings of the National Library of Latvia XXX11.12, 113–141.  http://doi.org/10.52197/AQKR3401.

Baklāne, Anda and Valdis Saulespurēns (2025). “Extracting Semantically Related Concepts from the Corpus of Latvian Novels: A Comparative Analysis of Word2Vec, GPT-4o, and Gemini-1.5 Results”. In: DHNB 2025. Programme. Book of Abstracts. Presented at DHNB2025 ‘Digital Dreams and Practices’. Estonian Literary Museum. https://dhnb.eu/conferences/dhnb2025/.

Baltijas Vēstnesis (1902). “Ri̇ga”. In: Baltijas Vēstnesis 244, 3.

Biedriņš, Andris (2021). “Ri̇gas elektriskais tramvajs”. In: Transports. Eiropas kultūras mantojuma dienas. Mantojums, 56–57.

Bologna, Federica (2020). “A computational approach to urban space in science fiction”. In: Journal of Cultural Analytics 5.2, 37–60.  http://doi.org/10.22148/001c.18120.

Budiloviča, Eveli̇na (2024). Pilsētas transports Latvijā. Nacionālā enciklopēdija. https://enciklopedija.lv/skirklis/89597-pils%C4%93tas-transports-Latvij%C4%81 (visited on 01/20/2026).

Čaks, Aleksandrs (1930). Poēma par ormani. Zelta grauds.

Caune, Andris (2020). Ri̇gas satiksme pirms 100 gadiem. Zinātne.

Deglavs, Augusts (1891). Starp divām uguni̇m. F. Gēliņš.

Deglavs, Augusts (1897–1899). Jaunā pasaule. Ernsts Plāte.

Dzimtenes Vēstnesis (1913). “Vietējās ziņas”. In: Dzimtenes Vēstnesis 123, 2.

Eglāja-Kristsone, Eva, Anda Baklāne, and Valdis Saulespurēns (2024). “Pilsētas transportli̇dzekļi latviešu senākajos romānos”. In: Proceedings of the National Library of Latvia XXX11.12, 65–96.  http://doi.org/10.52197/DIOW5621.

Erss, Ādolfs (1927). Aglonas dievmātes atgriešanās. Grāmatu Draugs.

Fan, Yaxin, Feng Jiang, Peifeng Li, and Haizhou Li (2023). “Uncovering the Potential of ChatGPT for Discourse Analysis in Dialogue: An Empirical Study”. In: Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024), 16998–17010.  http://doi.org/10.48550/arXiv.2305.08391.

Finch, Jason (2023). “Powered modernity, contested space: literary modernism and the London tram”. In: European Journal of English Studies 27.2, 288–308.  http://doi.org/10.1080/13825577.2024.2307035.

Fridlund, Mats, Michael Azar, Daniel Brodén, and Michael McGuire (2023). “The Cultural Imaginary of Terrorism: Close and Distant Readings of Political Terror in Swedish News and Fiction During the Cold War”. In: DHNB2023 Conference Proceedings. Vol. 5. 1, 90–104.  http://doi.org/10.5617/dhnbpub.10653.

Gavin, Adrienne E. (2015). “‘I Saw a Great Deal of Trouble Amongst the Horses in London’: Anna Sewells’ ”Black Beauty” and the Victorian Cab Horse”. In: Transport in British Fiction: Technologies of Movement, 1840–1940. Ed. by Adrienne E. Gavin and Humphries. Andrew F. Palgrave Macmillan, 101–122.

Gemini Team (2024). Gemini 1.5: Unlocking Multimodal Understanding Across Millions of Tokens of Context. Tech. rep. Google DeepMind.  http://doi.org/10.48550/arXiv.2403.05530.

Ģirupnieks, Jānis (1939). Ekspedicija zemes dziļumos. Valters un Rapa.

Graham, Shawn, Milligan Ian, and Scott Weingart (2016). Exploring Big Historical Data: The Historian’s Macroscope. Imperial College Press.

Gulbis, Ansis (1925). Fantastu mantojums. A. Gulbis.

Jaunsudrabiņš, Jānis (1911). Aija. J. Brigadera apgādi̇ba.

Jockers, Matthew (2013). Macroanalysis: Digital Methods in Literary History. University of Illinois Press.

Kaija, Ivande (1913). Iedzimtais grēks. Valters un Rapa.

Kaija, Ivande (1919). Jūgā. A. Valtera, J. Rapas un biedru komisijā.

Karjus, Andres (2025). “Machine-assisted Quantitizing Designs: Augmenting Humanities and Social Sciences with Artificial Intelligence”. In: Humanities and Social Sciences Communication 12.277, 1–18.  http://doi.org/10.1057/s41599-025-04503-w.

Kārkla, Zita and Eva Eglāja-Kristsone (2022). “Liriskās ģeogrāfijas: literārās telpas kartēšana latviešu rakstnieču romānos un i̇sprozā”. In: Letonica 47, 104–127.

Kaudzi̇te, Mati̇ss and Reinis Kaudzi̇te (1879). Mērnieku laiki. H. Allunans.

Kellerman, Robin (2019). “Waiting for Railways (1830–1914)”. In: Timescapes of Waiting: Spaces of Stasis, Delay and Deferral. Ed. by Christoph Singer, Robert Wirth, and Olaf Berwald. Brill Rodopi, 35–57.

Kloto, Amanda (1927). Sievietes uzvara. Nākotnes Sieviete.

Kohlrausch, Martin and J. C. Behrends (2014). Races to Modernity: Metropolitan Aspirations in Eastern Europe, 1890–1940. Central European University Press.

Lapiņš, Kārlis (1934a). Pagrimušie. Grāmatu Draugs.

Lapiņš, Kārlis (1934b). Studenti fermā. Autora izdevums.

Livesey, Ruth (2016). Writing the Stage Coach Nation: Locality on the Move in Nineteenth-Century British Literature. Oxford University Press.

Mansbach, Steven A. (2014). “Capital Modernism in the Baltic Republics: Kaunas, Tallinn, and Riga”. In: Races to Modernity: Metropolitan Aspirations in Eastern Europe, 1890–1940. Ed. by Jan Behrends and Martin Kohlrausch. Central European University Press, 309–328.

Māters, Juris (1879). Sadzi̇ves viļņi. M. Māters.

McNees, Eleanor (2010). “Public Transport in Woolf’s City Novels: The London Omnibus”. In: Woolf and the City: Selected Papers from the Nineteenth Annual Conference on Virginia Woolf. Ed. by Elizabeth Evans and Sarah Cornish. Clemson University Press, 31–39.

McShane, Clyde and Joel Tarr (2003). “The Decline of the Urban Horse in American Cities”. In: The Journal of Transport History 24.2, 177–198.  http://doi.org/10.7227/TJTH.24.2.4.

Mežsēts, Augusts (1929). Apburtā pilsēta. Grāmatu Draugs.

Mikolov, Tomas, Kai Chen, Greg Corrado, and Jeffrey Dean (2013). “Efficient Estimation of Word Representations in Vector Space”. In: arXiv preprint, 1–12.  http://doi.org/10.48550/arXiv.1301.3781.

Moretti, Franco (1998). Atlas of the European Novel: 1800–1900. Verso.

Moretti, Franco (2005). Graphs, Maps, Trees: Abstract Models for Literary History. Verso.

Moretti, Franco (2013). Distant Reading. Verso.

Ostups, Artis (2024). Laicena kunga ātruma izjūta: par dzejnieka Linarda Laicena (1883–1938) Berli̇nes dzeju. https://www.punctummagazine.lv/2024/03/27/laicena-kunga-atruma-izjuta/ (visited on 01/20/2026).

Paulockis, Miķelis (1936). Vecās bākas noslēpums. Laikmets.

Paulockis, Miķelis (1938a). Pasaules glābējs. Senatne.

Paulockis, Miķelis (1938b). Profesora Sūnas bri̇nišķi̇gais eliksi̇rs. Senatne.

Piper, Andrew (2018). Enumerations. The University of Chicago Press.

Plaudis, Jānis (1935). Ģimnāzisti. A. Gulbis.

Poruks, Jānis (1895). Pērļu zvejnieks. E. Plāte.

Prūsa, Emi̇lija (1936). Tāle vilina. A. Gulbis.

Purapuķe, Jānis (1892). Pasaules lāpi̇tājs. M. Jakobsons.

Řehůřek, Radim and Petr Sojka (2010). “Software Framework for Topic Modelling with Large Corpora”. In: Proceedings of the LREC 2010 Workshop on New Challenges for NLP Frameworks. ELRA, 45–50. https://repozitar.cz/publication/15725/?lang=en (visited on 01/20/2026).

Rodman, Emma (2020). “A Timely Intervention: Tracking the Changing Meanings of Political Concepts with Word Vectors”. In: Political Analysis 28.1, 87–111.

Rodriguez, Pedro L., Arthur Spirling, and Brandon M. Stewart (2023). “Embedding Regression: Models for Context-Specific Description and Inference”. In: American Political Science Review 117.4, 1255–1274.  http://doi.org/10.1017/S0003055422001228.

Rozi̇tis, Pāvils (1928). Ceplis. Valters un Rapa.

Rozi̇tis, Pāvils (1936). Valmieras puikas. Valters un Rapa.

Schmucki, Barbara (2012). “The Machine in the City: Public Appropriation of the Tramway in Britain and Germany, 1870–1915”. In: Journal of Urban History 38.6, 1060–1093.  http://doi.org/10.1177/0096144211435121.

Schöch, Christof, Roxana Patras, Tomaž Erjavec, and Diana Santos (2021). “Creating the European Literary Text Collection (ELTeC): Challenges and Perspectives”. In: Modern Languages Open 2021.1, 1–19.  http://doi.org/10.3828/mlo.v0i0.364.

Skuja, Fri̇dis (1912). Zem saules. K. Ķeirans.

Skuja, Fri̇dis (1924). Sidrabota saule lec. Latvijas Aizsardzi̇bas Biedri̇ba.

Stirna, Li̇ga (2024). Ātruma ierobežojums – 12 kilometri stundā. Kā pa Ri̇gu sāka joņot pirmie auto. https://www.delfi.lv/a/120011444 (visited on 01/20/2026).

Tenen, Dennis Yi (2018). “Toward a Computational Archaeology of Fictional Space”. In: New Literary History 49.1, 119–147.  http://doi.org/10.1353/nlh.2018.0005.

Törnberg, Petter (2024). “Best Practices for Text Annotation with Large Language Models”. In: arXiv preprint.  http://doi.org/10.48550/arXiv.2402.05129.

Underwood, Ted (2019). Distant Horizons. The University of Chicago Press.

Upi̇tis, Andrejs (1913). Pēdējais latvietis. A. Golta apgādi̇bā.

Upi̇ts, Andrejs (1921). Zelts. D. Zeltiņa un A. Golta apgāds.

Veselis, Jānis (1931). Dienas krusts. Valters un Rapa.

Veselis, Jānis (1934). Cilvēku sacelšanās. Valters un Rapa.

Ziems, Calet, William Held, Omar Shaikh, Jiaao Chen, Zhehao Zhang, and Diyi Yang (2023). “Can Large Language Models Transform Computational Social Science?” In: Computational Linguistics 50.1, 1–53.  http://doi.org/10.1162/coli_a_00502.