Large language models (LLMs) have shown potential as tools for scientific discovery. This has engendered growing interest in their use in humanistic disciplines, such as historical linguistics and literary studies. These fields often construct arguments on the basis of delineations like genre, or more inflexibly, time period. Although efforts have been made to restrict inference to specific domains via fine-tuning or model editing, we posit that the only true guarantee is domain-restricted pretraining – typically, a data- and compute-expensive proposition. We show that efficient pretraining techniques can produce useful models over corpora too large for easy manual inspection but too small for "typical" LLM approaches. We employ a novel date-attribution pipeline in order to obtain a temporally-segmented dataset of five 10-million-word slices. We train two corresponding five-model batteries over these corpus segments, efficient pretraining and Llama3-8B parameter efficiently finetuned. We find that the pretrained models are faster to train than the finetuned baselines and that they better respect the historical divisions of our corpus. Emphasizing speed and precision over a-historical comprehensiveness enables a number of novel approaches to hypothesis discovery and testing in our target fields. Taking up diachronic linguistics as a testbed, we show that our method enables the detection of a diverse set of phenomena, including en masse lexical change, non-lexical (grammatical and morphological) change, and word sense introduction/obsolescence. We provide a ready-to-use pipeline that allows extension of our approach to other target fields with only minimal adaptation.

Rhetorical devices are difficult to translate, but they are crucial to the translation of literary documents. We investigate the use of multilingual embedding spaces to characterize the preservation of intertextuality, one common rhetorical device, across human and machine translation. To do so, we use Biblical texts, which are both full of intertextual references and are highly translated works. We provide a metric to characterize intertextuality at the corpus level and provide a quantitative analysis of the preservation of this rhetorical device across extant human translations and machine-generated counterparts. We go on to provide qualitative analysis of cases wherein human translations over- or underemphasize the intertextuality present in the text, whereas machine translations provide a neutral baseline. This provides support for established scholarship proposing that human translators have a propensity to amplify certain literary characteristics of the original manuscripts.

We measure the effects of several implementation choices for the Dynamic Embedded Topic Model, as applied to five distinct diachronic corpora, with the goal of isolating important decisions for its use and further development. We identify priorities that will maximize utility in applied scholarship, including the practical scalability of vocabulary size to best exploit the strengths of embedded representations, and more flexible modeling of intervals to accommodate the uneven temporal distributions of historical writing. Of similar importance, we find performance is not significantly or consistently affected by several aspects that otherwise limit the model's application or might consume the resources of a grid search.

We present an ngram model-based logit scaling technique that effectively transfers extreme subword stylistic variation to large language models at inference time. We demonstrate its efficacy by tracking the perplexity of generated text with respect to the ngram interpolated and original versions of an evaluation model. Minimizing the former measure while the latter approaches the perplexity of a text produced by a target author or character lets us select a sufficient degree of adaptation while retaining fluency.

Chiasmus, a debated literary device in Biblical texts, has captivated mystics while sparking ongoing scholarly discussion. In this paper, we introduce the first computational approach to systematically detect chiasmus within Biblical passages. Our method leverages neural embeddings to capture lexical and semantic patterns associated with chiasmus, applied at multiple levels of textual granularity (half-verses, verses). We also involve expert annotators to review a subset of the detected patterns. Despite its computational efficiency, our method achieves robust results, with high inter-annotator agreement and system precision@k of 0.80 at the verse level and 0.60 at the half-verse level. We further provide a qualitative analysis of the distribution of detected chiasmi, along with selected examples that highlight the effectiveness of our approach.

We present a novel combination of dynamic embedded topic models and change-point detection to explore diachronic change of lexical semantic modality in classical and early Christian Latin. We demonstrate several methods for finding and characterizing patterns in the output, and relating them to traditional scholarship in Comparative Literature and Classics. This simple approach to unsupervised models of semantic change can be applied to any suitable corpus, and we conclude with future directions and refinements aiming to allow noisier, less-curated materials to meet that threshold.

In this study, we present a generalizable workflow to identify documents in a historic language with a nonstandard language and script combination, Armeno-Turkish. We introduce the task of detecting distinct patterns of multilinguality based on the frequency of structured language alternations within a document.

We present a dataset of 19th century American literary orthovariant tokens with a novel layer of human-annotated dialect group tags designed to serve as the basis for computational experiments exploring literarily meaningful orthographic variation. We perform an initial broad set of experiments over this dataset using both token (BERT) and character (CANINE)-level contextual language models. We find indications that the "dialect effect" produced by intentional orthographic variation employs multiple linguistic channels, and that these channels are able to be surfaced to varied degrees given particular language modelling assumptions. Specifically, we find evidence showing that choice of tokenization scheme meaningfully impact the type of orthographic information a model is able to surface.

We introduce a graph-aware autoencoder ensemble framework, with associated formalisms and tooling, designed to facilitate deep learning for scholarship in the humanities. By composing sub-architectures to produce a model isomorphic to a humanistic domain we maintain interpretability while providing function signatures for each sub-architectural choice, allowing both traditional and computational researchers to collaborate without disrupting established practices. We illustrate a practical application of our approach to a historical study of the American post-Atlantic slave trade, and make several specific technical contributions: a novel hybrid graph-convolutional autoencoder mechanism, batching policies for common graph topologies, and masking techniques for particular use-cases. The effectiveness of the framework for broadening participation of diverse domains is demonstrated by a growing suite of two dozen studies, both collaborations with humanists and established tasks from machine learning literature, spanning a variety of fields and data modalities. We make performance comparisons of several different architectural choices and conclude with an ambitious list of imminent next steps for this research.

We present a novel corpus consisting of orthographically variant words found in works of 19th century U.S. literature annotated with their corresponding "standard" word pair. We train a set of neural edit distance models to pair these variants with their standard forms, and compare the performance of these models to the performance of a set of neural edit distance models trained on a corpus of orthographic errors made by L2 English learners. Finally, we analyze the relative performance of these models in the light of different negative training sample generation strategies, and offer concluding remarks on the unique challenge literary orthographic variation poses to string pairing methodologies.

Our submission to the MADAR shared task on Arabic dialect identification employed a language modeling technique called Prediction by Partial Matching, an ensemble of neural architectures, and sources of additional data for training word embeddings and auxiliary language models. We found several of these techniques provided small boosts in performance, though a simple character-level language model was a strong baseline, and a lower-order LM achieved best performance on Subtask 2. Interestingly, word embeddings provided no consistent benefit, and ensembling struggled to outperform the best component submodel. This suggests the variety of architectures are learning redundant information, and future work may focus on encouraging decorrelated learning.

This work considers a task from traditional literary criticism: annotating a structured, composite document with information about its sources. We take the Documentary Hypothesis, a prominent theory regarding the composition of the first five books of the Hebrew bible, extract stylistic features designed to avoid bias or overfitting, and train several classification models. Our main result is that the recently-introduced graph convolutional network architecture outperforms structurally-uninformed models. We also find that including information about the granularity of text spans is a crucial ingredient when employing hidden layers, in contrast to simple logistic regression. We perform error analysis at several levels, noting how some characteristic limitations of the models and simple features lead to misclassifications, and conclude with an overview of future work.

Twitter is a ubiquitous source of micro-blog social media data, providing the academic, industrial, and public sectors real-time access to actionable information. A particularly attractive property of some tweets is *geo-tagging*, where a user account has opted-in to attaching their current location to each message. Unfortunately (from a researcher’s perspective) only a fraction of Twitter accounts agree to this, and these accounts are likely to have systematic diffences with the general population. This work is an exploratory study of these differences across the full range of Twitter content, and complements previous studies that focus on the English-language subset. Additionally, we compare methods for querying users by self-identified properties, finding that the constrained semantics of the “description” field provides cleaner, higher-volume results than more complex regular expressions.

Computer Assisted Discovery Extraction and Translation (CADET) is a workbench for helping knowledge workers find, label, and translate documents of interest. It combines a multitude of analytics together with a flexible environment for customizing the workflow for different users. This open-source framework allows for easy development of new research prototypes using a micro-service architecture based atop Docker and Apache Thrift.

When considering a social media corpus, we often have access to structural information about how messages are flowing between people or organizations.This information is particularly useful when the linguistic evidence is sparse, incomplete, or of dubious quality.In this paper we construct a simple model to leverage the structure of Twitter data to help determine the set of languages each user is fluent in.Our results demonstrate that imposing several intuitive constraints leads to improvements in performance and stability.We release the first annotated data set for exploring this task, and discuss how our approach may be extended to other applications.

We discuss factors that affect human agreement on a semantic labeling task in the art history domain, based on the results of four experiments where we varied the number of labels annotators could assign, the number of annotators, the type and amount of training they received, and the size of the text span being labeled. Using the labelings from one experiment involving seven annotators, we investigate the relation between interannotator agreement and machine learning performance. We construct binary classifiers and vary the training and test data by swapping the labelings from the seven annotators. First, we find performance is often quite good despite lower than recommended interannotator agreement. Second, we find that on average, learning performance for a given functional semantic category correlates with the overall agreement among the seven annotators for that category. Third, we find that learning performance on the data from a given annotator does not correlate with the quality of that annotator's labeling. We offer recommendations for the use of labeled data in machine learning, and argue that learners should attempt to accommodate human variation. We also note implications for large scale corpus annotation projects that deal with similarly subjective phenomena.

The CLiMB project investigates semi-automatic methods to extract descriptive metadata from texts for indexing digital image collections. We developed a set of functional semantic categories to classify text extracts that describe images. Each semantic category names a functional relation between an image depicting a work of art historical significance, and expository text associated with the image. This includes description of the image, discussion of the historical context in which the work was created, and so on. We present interannotator agreement results on human classification of text extracts, and accuracy results from initial machine learning experiments. In our pilot studies, human agreement varied widely, depending the labeler's expertise, the image-text pair under consideration, the number of labels that could be assigned to one text, and the type of training, if any, we gave labelers. Initial machine learning results indicate the three most relevant categories are machine learnable. Based on our pilot work, we implemented a labeling interface that we are currently using to collect a large dataset of text that will be used in training and testing machine classifiers.

We present a system which applies text mining using computational linguistic techniques to automatically extract, categorize, disambiguate and filter metadata for image access. Candidate subject terms are identified through standard approaches; novel semantic categorization using machine learning and disambiguation using both WordNet and a domain specific thesaurus are applied. The resulting metadata can be manually edited by image catalogers or filtered by semi-automatic rules. We describe the implementation of this workbench created for, and evaluated by, image catalogers. We discuss the system's current functionality, developed under the Computational Linguistics for Metadata Building (CLiMB) research project. The CLiMB Toolkit has been tested with several collections, including: Art Images for College Teaching (AICT), ARTStor, the National Gallery of Art (NGA), the Senate Museum, and from collaborative projects such as the Landscape Architecture Image Resource (LAIR) and the field guides of the Vernacular Architecture Group (VAG).