Do neural machine translation models learn language-universal conceptual representations, or do they merely cluster languages by surface similarity? We investigate this question by probing the representation geometry of Meta's NLLB-200, a 200-language encoder-decoder Transformer, through six experiments that bridge NLP interpretability with cognitive science theories of multilingual lexical organization. Using the Swadesh core vocabulary list embedded across 135 languages, we find that the model's embedding distances significantly correlate with phylogenetic distances from the Automated Similarity Judgment Program ($ρ= 0.13$, $p = 0.020$), demonstrating that NLLB-200 has implicitly learned the genealogical structure of human languages. We show that frequently colexified concept pairs from the CLICS database exhibit significantly higher embedding similarity than non-colexified pairs ($U = 42656$, $p = 1.33 \times 10^{-11}$, $d = 0.96$), indicating that the model has internalized universal conceptual associations. Per-language mean-centering of embeddings improves the between-concept to within-concept distance ratio by a factor of 1.19, providing geometric evidence for a language-neutral conceptual store analogous to the anterior temporal lobe hub identified in bilingual neuroimaging. Semantic offset vectors between fundamental concept pairs (e.g., man to woman, big to small) show high cross-lingual consistency (mean cosine = 0.84), suggesting that second-order relational structure is preserved across typologically diverse languages. We release InterpretCognates, an open-source interactive toolkit for exploring these phenomena, alongside a fully reproducible analysis pipeline.
Tjaša Arčon, Matej Klemen, Marko Robnik-Šikonja
et al.
LLMs are routinely evaluated on language use, yet their explicit knowledge about linguistic structure remains poorly understood. Existing linguistic benchmarks focus on narrow phenomena, emphasize high-resource languages, and rarely test metalinguistic knowledge - explicit reasoning about language structure. We present a multilingual evaluation of metalinguistic knowledge in LLMs, based on the World Atlas of Language Structures (WALS), documenting 192 linguistic features across 2,660 languages. We convert WALS features into natural-language multiple-choice questions and evaluate models across documented languages. Using accuracy and macro F1, and comparing to chance and majority-class baselines, we assess performance and analyse variation across linguistic domains and language-related factors. Results show limited metalinguistic knowledge: GPT-4o performs best but achieves moderate accuracy (0.367), while open-source models lag. Although all models perform above chance, they fail to outperform the majority-class baseline, suggesting they capture broad cross-linguistic patterns but lack fine-grained distinctions. Performance varies by domain, partly reflecting differences in online visibility. At the language level, accuracy correlates with digital language status: languages with greater digital presence and resources are evaluated more accurately, while low-resource languages perform worse. Analysis of predictive factors confirms that resource-related indicators (Wikipedia size, corpus availability) are more informative than geographic, genealogical, or sociolinguistic factors. Overall, LLM metalinguistic knowledge appears fragmented and shaped mainly by data availability, rather than broadly generalizable grammatical competence. We release the benchmark as an open-source dataset to support evaluation across languages and encourage greater global linguistic diversity in future LLMs.
This essay examines relationships between adoptees and the (extended) adoptive family, focusing on the inheritance rights of adopted persons as entry points into levels and cycles of their belonging and un-belonging. The essay contextualizes a case report (or summary reports) on the kind of estrangement in the adoptee world that is fueled by inheritance disputes. It delves into postadoption perceptions and thus into the “unwritten” truths about adoption and its possible fallout. It draws from archival sources, semi-structured interviews (life-story interviewing), and life writing by adoptees, and also from a sequence of real-life exchanges dating back to 2018. All these sources focus on the contested inheritance of children, now older adults, who were adopted from Greece in the 1950s–60s and who became (or should have become) subsequent heirs to the estates of their adoptive parents and/or relatives. The Greek out-of-country adoptions of the postwar and early Cold War era involved more than 4000 children, most of whom were sent to the United States. The various testimonies and sections reflect critically on the continuing trend to infantilize the adopted persons, forever the adopted children, to push their origins back into the past and into geographical distance, to untie the family connections they have forged over the course of half a century. The examples take the reader from the adoptive family’s pre-adoption attempts at disowning the child through the postadoption stage of the end of an adopted lifetime, including cases of the extended adoptive family’s attempts at “de-adopting” the adopted person. This essay includes various sources of life-cycle documentation, among them an extensive case study and online obituaries. It adheres to truth and authenticity by incorporating fairly long original quotations, which, in the case study of the second half especially, assist the reader in comprehending much historical information in a question-and-answer format. This bolder structure offers the advantage of taking the reader step by step through the transactions of a prominent Greek adoption scheme (Rebecca and Maurice Issachar) and also through the various layers of the postadoption mindset and minefield. The material presented here is intended to raise awareness that change can and must still benefit the Greek adoptees today, whose lives may have been permeated by conditionality and nonlinearity. I conclude that, in the cases discussed here, the child’s orphanhood may well be a perpetual state, with the adoptee being orphaned of individuality and of a protective family on more than just one occasion.
The violent dispossession of land in South Africa disrupted more than just homes—it severed Black South Africans from a sacred, ancestral connection to land as a source of identity, belonging, and spiritual dwelling. This article examines how forced removals displaced not only bodies but also histories, memories, and the deep-rooted sense of ikhaya (home). Rooted in the concept of endleleni (being on the road/along the road), this study explores how amaXhosa navigate the in-between journey of landlessness and homecoming. Using indigenous storytelling methodology, it reveals how land is not merely for shelter or sustenance but is intricately tied to birth, the umbilical cord, and death, making its reclamation a fight for existence itself.
This study analyzes displaced families’ disintegration amidst multifaceted crises in the Far North Region of Cameroon. The focus is on displaced families in four divisions where host communities along the border of Nigeria have fled Boko Haram, due to sociopolitical instability and environmental degradation. Both quantitative and qualitative analyses show how insecurity and precarity have resulted in disintegrating, dispersing and sometimes recomposing the traditional structure of the family institution by disrupting marriage traditions, gender roles and intergenerational relations. Such shifts in the family occurred as a result of the challenging contexts in their hometowns, during flight and after settling in the host towns. The findings demonstrate the effects of the multifaceted crises, which created intergenerational, but also inter- and intra-community rifts that are challenging to repair amidst mass distrust and continued insecurity. This study contributes especially to understanding how these crises affect family cohesion and intergenerational relations as part of dynamics of great social change.
Most star clusters dissolve into the Galaxy over tens to hundreds of millions of years after they form. While recent Gaia studies have honed our view of cluster dispersal, the exact chronology of which star formation events begat which star cluster remnants remains unclear. This problem is acute after 100 Myr, when cluster remnants have spread over hundreds of parsecs and most age estimates for main sequence stars are too imprecise to link the stars to their birth events. Here we develop a Bayesian framework that combines TESS stellar rotation rates with Gaia kinematics to identify diffuse remnants of open clusters. We apply our method to the Pleiades, which previous studies have noted shows kinematic similarities to other nearby young stellar groups. We find that the Pleiades constitutes the bound core of a much larger, coeval structure that contains multiple known clusters distributed over 600 pc. We refer to this structure as the Greater Pleiades Complex. On the basis of uniform ages, coherent space velocities, detailed elemental abundances, and traceback histories, we conclude that most stars in this complex originated from the same giant molecular cloud. This work establishes a scalable approach for tracing the genealogies of nearby clusters and further cements the Pleiades as a cornerstone of stellar astrophysics. We aim to apply this methodology to other associations as part of the upcoming TESS All-Sky Rotation Survey.
A fundamental concern in linguistics has been to understand how languages change, such as in relation to word order. Since the order of words in a sentence (i.e. the relative placement of Subject, Object, and Verb) is readily identifiable in most languages, this has been a productive field of study for decades (see Greenberg 1963; Dryer 2007; Hawkins 2014). However, a language's word order can change over time, with competing explanations for such changes (Carnie and Guilfoyle 2000; Crisma and Longobardi 2009; Martins and Cardoso 2018; Dunn et al. 2011; Jager and Wahle 2021). This paper proposes a general universal explanation for word order change based on a theory of communicative interaction (the Min-Max theory of language behavior) in which agents seek to minimize effort while maximizing information. Such an account unifies opposing findings from language processing (Piantadosi et al. 2011; Wasow 2022; Levy 2008) that make different predictions about how word order should be realized crosslinguistically. The marriage of both "efficiency" and "surprisal" approaches under the Min-Max theory is justified with evidence from a massive dataset of 1,942 language corpora tagged for parts of speech (Ring 2025), in which average lengths of particular word classes correlates with word order, allowing for prediction of basic word order from diverse corpora. The general universal pressure of word class length in corpora is shown to give a stronger explanation for word order realization than either genealogical or areal factors, highlighting the importance of language corpora for investigating such questions.
Marjan Hosseini, Ella Veiner, Thomas Bergendahl
et al.
Computing haplotypes from sequencing data, i.e. haplotype assembly, is an important component of molecular and population genetics problems, including interpreting the effects of genetic variation on complex traits and reconstructing genealogical relationships. Assembling the haplotypes of polyploid genomes remains a significant challenge due to the exponential search space of haplotype phasings and read assignment ambiguity; the latter challenge is particularly difficult for haplotype assemblers since the information contained within the observed sequence reads is often insufficient for unambiguous haplotype assignment in polyploid genomes. We present pHapCompass, probabilistic haplotype assembly algorithms for diploid and polyploid genomes that explicitly model and propagate read assignment ambiguity to compute a distribution over polyploid haplotype phasings. We develop graph theoretic algorithms to enable statistical inference and uncertainty quantification despite an exponential space of possible phasings. Since prior work evaluates polyploid haplotype assembly on synthetic genomes that do not reflect the realistic genomic complexity of polyploidy organisms, we develop a computational workflow for simulating genomes and DNA-seq for auto- and allopolyploids. Additionally, we generalize the vector error rate and minimum error correction evaluation criteria for partially phased haplotypes. Benchmarking of pHapCompass and several existing polyploid haplotype assemblers shows that pHapCompass yields competitive performance across varying genomic complexities and polyploid structures while retaining an accurate quantification of phase uncertainty. The source code for pHapCompass, simulation scripts, and datasets are freely available at https://github.com/bayesomicslab/pHapCompass.
We generalize the celebrated coagulation-fragmentation duality of Pitman (1999), originally established for the PD$(α,θ)$ laws of Pitman and Yor (1997), resolving a two-decade open problem. Our framework extends the duality to processes driven by arbitrary non-negative L'evy subordinators and, for the first time, to multi-group settings with coupled dynamics. The solution is a novel four-component system built from the PHIBP, a framework developed for modeling complex microbiome species sampling (arXiv:2502.01919), which circumvents intractable analysis on traditional partition spaces. Crucially, this architecture embeds naturally within Bertoin's (2006) continuous-time fragmentation framework, resolving a foundational impasse he highlighted [Ch. 4, p. 213]: where time-reversal fails, we introduce simultaneous structural duality, where fragmentation and coalescence evolve in physical time while maintaining a pointwise dual relationship via coupled subordinators. This provides a new generative modeling framework enabling continuous-time representations of coupled genealogical and mutational dynamics -- opening avenues for complex ancestral processes such as recombination graphs. The architecture yields exact compound Poisson representations providing tractable paths to explicit joint EPPF laws, with exact sampling inherited from the PHIBP framework. Section 7 develops the $h$-biased Lévy-Itô coupled duality constructor, extending our framework to arbitrary Polish spaces by adapting the $h$-biased PRM structure of Pitman and Yor to Cox processes driven by a common PRM. This demonstrates that duality arises from point-process regrouping rather than features specific to interval partitions.
Investigating linguistic relationships on a global scale requires analyzing diverse features such as syntax, phonology and prosody, which evolve at varying rates influenced by internal diversification, language contact, and sociolinguistic factors. Recent advances in machine learning (ML) offer complementary alternatives to traditional historical and typological approaches. Instead of relying on expert labor in analyzing specific linguistic features, these new methods enable the exploration of linguistic variation through embeddings derived directly from speech, opening new avenues for large-scale, data-driven analyses. This study employs embeddings from the fine-tuned XLS-R self-supervised language identification model voxlingua107-xls-r-300m-wav2vec, to analyze relationships between 106 world languages based on speech recordings. Using linear discriminant analysis (LDA), language embeddings are clustered and compared with genealogical, lexical, and geographical distances. The results demonstrate that embedding-based distances align closely with traditional measures, effectively capturing both global and local typological patterns. Challenges in visualizing relationships, particularly with hierarchical clustering and network-based methods, highlight the dynamic nature of language change. The findings show potential for scalable analyses of language variation based on speech embeddings, providing new perspectives on relationships among languages. By addressing methodological considerations such as corpus size and latent space dimensionality, this approach opens avenues for studying low-resource languages and bridging macro- and micro-level linguistic variation. Future work aims to extend these methods to underrepresented languages and integrate sociolinguistic variation for a more comprehensive understanding of linguistic diversity.
While much research in artificial intelligence (AI) has focused on scaling capabilities, the accelerating pace of development makes countervailing work on producing harmless, "aligned" systems increasingly urgent. Yet research on alignment has diverged along two largely parallel tracks: safety--centered on scaled intelligence, deceptive or scheming behaviors, and existential risk--and ethics--focused on present harms, the reproduction of social bias, and flaws in production pipelines. Although both communities warn of insufficient investment in alignment, they disagree on what alignment means or ought to mean. As a result, their efforts have evolved in relative isolation, shaped by distinct methodologies, institutional homes, and disciplinary genealogies. We present a large-scale, quantitative study showing the structural split between AI safety and AI ethics. Using a bibliometric and co-authorship network analysis of 6,442 papers from twelve major ML and NLP conferences (2020-2025), we find that over 80% of collaborations occur within either the safety or ethics communities, and cross-field connectivity is highly concentrated: roughly 5% of papers account for more than 85% of bridging links. Removing a small number of these brokers sharply increases segregation, indicating that cross-disciplinary exchange depends on a handful of actors rather than broad, distributed collaboration. These results show that the safety-ethics divide is not only conceptual but institutional, with implications for research agendas, policy, and venues. We argue that integrating technical safety work with normative ethics--via shared benchmarks, cross-institutional venues, and mixed-method methodologies--is essential for building AI systems that are both robust and just.
A genealogy of anti-Ahmadi narratives in Pakistan can be traced from the Zia regime (1977–88), through Maulana Maududi, and back to Muhammad Iqbal. The Zia regime criminalised and de facto excluded Ahmadis from the Muslim community through the 1984 Ordinance, rationalised in a pamphlet that plagiarised Maududi’s work. Before their criminalisation, Ahmadis were constitutionally excluded from the Pakistani Muslim community through the 1974 Second Amendment, which, like Iqbal, defined a Muslim according to belief in Khatam-e-Nabuwwat (the finality of prophethood). One argument to separate Ahmadis was thus theological, that they misinterpreted Khatam-e-Nabuwwat by believing that the founder of the community was a prophet. Other arguments were socioeconomic and political, with Iqbal and Maududi both perceiving Ahmadis as declaring themselves separate from Muslims, while benefitting from their association with the Muslim community. This will be contextualised within Iqbal’s desire to maintain the unity of the Muslim community in the face of socioeconomic challenges prior to Pakistan’s creation but also the dismantling of Muslim political power in both pre-Partition India and the wider world, an inherent concern of the anti-colonial Khilafat movement (1919–22), from which other prominent anti-Ahmadi figures emerged. Maududi’s and the Zia regime’s presentations of Ahmadis as un-Islamic and historically sympathetic to British colonialists functioned to assert that Ahmadis were anti-Pakistan, too.
Philosophy of religion. Psychology of religion. Religion in relation to other subjects
Recent research in social psychology underscores the role of language and its intersection with other identity markers, including ethnic visibility, in exploring social perceptions and biases. This paper examines the physical visibility of people of Middle Eastern or North African (MENA) descent in the U.S., and the linguistic visibility of a concentrated MENA American community in Dearborn, Michigan. Relying on headshots, Study 1 shows that MENA could be an ambiguous ethnic community based solely on physical appearance, while religiously affiliated attire proves to be a significant ethnic marker for MENA. Using audio cues, Study 2 shows that the English variety spoken in Dearborn is a recognizable variety with masculinity associations. As such, Dearborn English is argued to be an ethnolinguistic repertoire that can be used to project ethno-local visibility. The results highlight the importance of the linguistic visibility of Dearborn and future research on language attitudes towards this variety.
To understand the changing trends in Jewish Genealogy over the past 40 years, the author has interviewed more than one hundred genealogists around the world. All of them are connected to the two most important genealogy organisations, JewishGen and JRI-Poland. They range from hobbyists researching their own families to professionals researching specific prewar Polish shtetls and those serving the entire genealogical community. Based on their responses to 26 questions, the author has identified two important features of contemporary Jewish genealogy: its democratisation and institutionalisation. The democratisation of genealogical research has contributed to a great expansion of the field. The focus of interest is no longer limited to only rabbinical families but is also concerned with the common man. Thus, genealogists today speak not only on behalf of <i>sheyne yidn</i> and otherwise distinguished families but also on behalf of the millions of murdered „ordinary” Jews who once lived in Poland. The institutionalisation of genealogy refers to the degree to which genealogical research organisations like JewishGen or JRI-Poland now provide some of the same functions provided years ago by the <i>landsmanshaft</i> institutions. Today, descendants of a particular shtetl often discover and connect to each other through genealogical researchers and these genealogical organisations. How these Jewish genealogical practices can be/are used to strengthen the <i>landsmanshaft</i>-like function will be examined.
Ajith Harish, Fernando A. Lopes Pinto, Susanne Eriksson
et al.
Abstract Several indigenous cattle breeds in Sweden are endangered. Conservation of their genetic diversity and genomic characterization is a priority. Whole-genome sequences (WGS) with a mean coverage of 25X, ranging from 14 to 41X were obtained for 30 individuals of the breeds Fjällko, Fjällnära, Bohuskulla, Rödkulla, Ringamåla, and Väneko. WGS-based genotyping revealed 22,548,028 variants in total, comprising 18,876,115 single nucleotide polymorphisms (SNPs) and 3,671,913 indels. Out of these, 1,154,779 SNPs and 304,467 indels were novel. Population stratification based on roughly 19 million SNPs showed two major groups of the breeds that correspond to northern and southern breeds. Overall, a higher genetic diversity was observed in the southern breeds compared to the northern breeds. While the population stratification was consistent with previous genome-wide SNP array-based analyses, the genealogy of the individuals inferred from WGS based estimates turned out to be more complex than expected from previous SNP-array based estimates. Polymorphisms and their predicted phenotypic consequences were associated with differences in the coat color phenotypes between the northern and southern breeds. Notably, these high-consequence polymorphisms were not represented in SNP arrays, which are used routinely for genotyping of cattle breeds. This study is the first WGS-based population genetic analysis of Swedish native cattle breeds. The genetic diversity of native breeds was found to be high. High-consequence polymorphisms were linked with desirable phenotypes using whole-genome genotyping, which highlights the pressing need for intensifying WGS-based characterization of the native breeds.
The article argues that Michel Foucault’s frequently stated opposition between his “genealogy of norms” and the Marxist “critique of ideology” is, in many ways, unfounded. By analyzing some of his texts, I claim that Foucault’s objections to the critique of ideology conceal his creative use of Karl Marx’s and Friedrich Engels’ insights. To support this claim, I first identify some “Marxist echoes” in Foucault’s work. I then demonstrate that these echoes lead to original reformulations of the critique of ideology, which are not reducible to Louis Althusser’s theory of the “ideological apparatuses”. Thus, the article highlights that Foucault’s “genealogy of norms” is surprisingly connected to Marx’s and Engels’ hypotheses presented in the manuscripts known as The German Ideology.
This work introduces a construction of conformal processes that combines the theory of branching processes with chordal Loewner evolution. The main novelty lies in the choice of driving measure for the Loewner evolution: given a finite genealogical tree $\mathcal{T}$, we choose a driving measure for the Loewner evolution that is supported on a system of particles that evolves by Dyson Brownian motion at inverse temperature $β\in (0,\infty]$ between birth and death events. When $β=\infty$, the driving measure degenerates to a system of particles that evolves through Coulombic repulsion between branching events. In this limit, the following graph embedding theorem is established: When $\mathcal{T}$ is equipped with a prescribed set of angles, $\{θ_v \in (0,π/2)\}_{v \in \mathcal{T}}$ the hull of the Loewner evolution is an embedding of $\mathcal{T}$ into the upper half-plane with trivalent edges that meet at angles $(2θ_v,2π-4θ_v,2θ_v)$ at the image of each edge $v$. We also study the scaling limit when $β\in (0,\infty]$ is fixed and $\mathcal{T}$ is a binary Galton-Watson process that converges to a continuous state branching process. We treat both the unconditioned case (when the Galton-Watson process converges to the Feller diffusion) and the conditioned case (when the Galton-Watson tree converges to the continuum random tree). In each case, we characterize the scaling limit of the driving measure as a superprocess. In the unconditioned case, the scaling limit is the free probability analogue of the Dawson-Watanabe superprocess that we term the Dyson superprocess.
Kala G. Pradeep, Kulinder Pal Singh, G. C. Dewangan
et al.
We present multi-wavelength temporal and spectral characteristics of a magnetic cataclysmic variable (MCV) Swift J0503.7-2819, using far ultraviolet (FUV) and X-ray data from AstroSat, supplemented with optical data from the Southern African Large Telescope and X-ray data from the XMM-Newton and Swift observatories. The X-ray modulations at 4897.6657 s and 3932.0355 s are interpreted as the orbital ($P_Ω$) and spin ($P_ω$) period, respectively, and are consistent with prior reports. With a spin-orbit period ratio of 0.8 and $P_Ω$ falling below the period gap (2-3 hrs) of CVs, Swift J0503.7-2819 would be the newest addition to the growing population of nearly synchronous MCVs, which we call EX Hya-like systems. Hard X-ray luminosity of $<$ $2.5\times10^{32} erg s^{-1}$, as measured with the Swift Burst Alert Telescope, identifies it to be a low-luminosity intermediate polar, similar to other EX Hya-like systems. The phenomenology of the light curves and the spectral characteristics rule out a purely disc-fed/stream-fed model and instead reveal the presence of complex accretion structures around the white dwarf. We propose a ring-like accretion flow, akin to EX Hya, using period ratio, stability arguments, and observational features. An attempt is made to differentiate between the asynchronous polar/nearly-synchronous intermediate polar nature of Swift J0503.7-2819. Further, we note that with the advent of sensitive surveys, a growing population of MCVs that exhibit characteristics of both polars and intermediate polars is beginning to be identified, likely forming a genealogical link between the two conventional classes of MCVs.
Anyone of African descent or with African ancestry who engages in a genealogy project soon learns that the U.S. Census is a helpful yet frustrating tool. In 2016, equipped with my history degree and an online ancestry search engine, I searched for my great-grandfather Leroy in census records after I saw a picture of him as a young man at work in Philadelphia. This image would have been unremarkable had it not been for the fact that my African American ancestor was so light skinned that he seemed to blend in with his co-workers at Kramer’s Fruit and Vegetables. I thought there had to be a story behind this. Classified as, “Mu”, for mulatto in most of his records, Leroy became “Black” on the census in 1930. My first thought was to question whether this categorization changed for other folks like him. My research led me to my master’s thesis “From ‘Mulatto’ to ‘Negro’: How Fears of ‘Passing’ Changed the 1930 United States Census”. Through this research, I also became closer to my father’s family. This piece will take you through this journey of discovery and my frustrations along the way.
In genealogy, tracing names and dates is often the initial goal, but, for many, desire soon turns to learning about the embodied lives of those who came before them. This type of texture is hard for any genealogist to locate, but excruciatingly hard for those seeking to trace family histories that include ancestors who were enslaved in the northern parts of the colonies that would become the United States. Often, records thin to nearly nothing and frame all lived experiences through the lens of an enslaver. This is true especially of public records, created, maintained, and curated by the state apparatus. By adhering to the proposition that even materials that do not immediately reveal much about Black life may be useful if we consider what is missing and left out, this article suggests that these types of documents might help breathe some fullness into the individual and collective lives of those Black ancestors whose humanity the state denied. Emerging from a larger project to locate stories and histories of Black residents of one of the first colonized spaces in British North America, this article focuses on the ways in which the publicly available Massachusetts pre-1850 Vital Records—which have specific “Negroes” sections—serve as an unexpected source of useful, if fragmentary, evidence of not only individual lives, but collective histories of the communities in which Black ancestors lived. Highlighting creative approaches to analyzing these particular vital records, and centering women’s lives throughout, this article demonstrates what is possible to learn about patterns of childbearing, relationships between and among enslaved persons owned by different families, the nature of religious lives or practices, relationships between enslavers and enslaved, and the movements, over time, of individuals and families. Alongside these possibilities, the violence, limitations, and challenges of the vital records are identified, including issues related to Afro-indigenous persons, the conflation of birth and baptismal records, and differential access to details of the lives of enslaved men vs. women.