This essay reviews the following works:

El caso Padilla/The Padilla Affair. Dir. Pavel Giroud. Prod. Lia Rodríguez, Alejandro Hernández. Spain and Cuba, 2022. 78 mins. Distributed by Amazon Prime Video.

Footage of Heberto Padilla’s “confession” recorded by Cuban authorities on April 27 at the National Union of Writers and Artists of Cuba. The footage is available on YouTube.

This study examines detrital garnet compositions from five samples spanning a Late Miocene–Pliocene section of Himalayan sedimentary rocks (Surai Khola, Siwalik Group and central Nepal) to assess provenance and tectonic implications. A total of 100 detrital garnets were analysed for edge-to-edge compositional zoning, revealing distinct groups linked to specific hinterland regions. Manual classification identified garnet Groups 1 and 2 as grossular, Group 3 as spessartine, and Groups 4 through 7 as almandine, varying in XCa, XMg and XMn. Most garnets exhibit low XMg and flat zoning, with Groups 6* and 7* containing slightly higher XMg. Statistical clustering aligns broadly with manual groupings, which strengthens provenance interpretations. Comparisons with hinterland garnet compositions expand provenance options to magmatic garnets and rocks outside the Himalayan core units. Eight Siwalik Group garnets were modelled for pressure–temperature conditions and paths. Group 4 and 6 garnets, potentially linked to blueschist/eclogites or metamorphosed arc/Himalayan core rocks, record conditions of 510–538°C and 4.6–6.8 kbar, with isothermal burial over 0.5–2 kbar. Group 2 garnet, resembling compositions from North Himalayan granitic enclaves, yields core conditions of 480°C and 6 kbar and an N-shaped pressure–temperature path. Two Group 5 garnets with zoning like those in the High Himalayan leucogranites yield 520–528°C at 3.2–3.6 kbar. These findings provide insights into Himalayan erosion dynamics, hinterland exhumation and sediment transport pathways. Integrating garnet compositional zoning with statistical clustering and thermodynamic modelling is valuable for provenance studies of garnet-bearing sedimentary sections.

Folk magic practices were common across the early modern Spanish Empire, including in seventeenth-century Manila where dozens of Asian herbalists and other practitioners of magic offered magical solutions in affairs of the heart and matters of fortune and divination to their mostly Spanish clients. At the centre of these folk magic activities were a group of Ternaten captives of war, relatives of the Sultan Saïd Berkat Syah, who was taken hostage by the Spanish during their invasion of Ternate in 1606. While the capture of Sultan Saïd by the Spanish in 1606 is well known within the historiography of the Maluku Islands, the presence of the Ternaten hostages within Manila in the early seventeenth century remains absent from the history of the port city. This article explores the lives of these Ternaten hostages, arguing that their spellcasting activities represent a hidden transcript of politics and power among previously marginalised historical subjects.

Italian ryegrass [Lolium perenne L. ssp. multiflorum (Lam.) Husnot] is a significant weed in winter wheat (Triticum aestivum L.), corn (Zea mays L.), soybean [Glycine max (L.) Merr.], and peanut (Arachis hypogaea L.) crops in Alabama. In response to reports of herbicide failure, field surveys were conducted in these cropping systems across Alabama in 2023. The objectives were to document the distribution of herbicide resistance in the collected L. perenne ssp. multiflorum populations. Populations were evaluated in a greenhouse for sensitivity to herbicides representing three modes of action: an acetolactate synthase (ALS) inhibitor (pyroxsulam), two acetyl-coenzyme A carboxylase (ACCase) inhibitors (fluazifop-butyl and clethodim), and a 5-enolpyruvylshikimate-3-phosphate synthase (EPSPS) inhibitor (glyphosate). Herbicide screenings were followed by dose–response assays of the most resistant L. perenne ssp. multiflorum population for each herbicide at eight rates (0.5, 1, 2, 4, 8, 16, 32, and 64×) compared with a susceptible population at six rates (0.0625, 0.125, 0.25, 0.5, 1, and 2×). Out of 44 populations evaluated, 21%, 11%, 25%, and 2% were found resistant to glyphosate, fluazifop-butyl, pyroxsulam, and clethodim, respectively. Resistance levels were confirmed to be 192-, 14-, 90-, and 738-fold for glyphosate, fluazifop-butyl, pyroxsulam, and clethodim, respectively. Mutation detection studies revealed specific mutations: Asp-2078-Gly in the ACCase gene, Pro-106-Ser in the EPSPS gene, and a novel Arg-421-Thr mutation in the ALS gene.

“La tierra está enferma”—“the territory is sick” as the result of the armed conflict, say the elders of the Awá people, one of Colombia’s Indigenous groups primarily living in the region of Nariño, who preserve their native language, the Awapit, and with it an entire spiritual and cultural relationship with its territory.1 For the Awá people, the prolonged internal armed conflict in Colombia affected not only human beings but also their Mother Earth, the Katsa Su, la gran casa—the big house. They consider their territory a living being that can experience pain and is currently sick and in need of reparation. In 2019, the Colombian Special Jurisdiction for Peace—as the court in charge of dealing with cases related to the armed conflict—embodied these views and recognized the Katsa Su as a victim of the armed conflict, with its own right to reparation.2 But what should reparations to the Katsa Su look like? And more broadly, how does the concept of repairing Indigenous territories challenge traditional anthropocentric assumptions of reparation in Colombia and beyond?

This study aimed to evaluate different precision hoeing methods on re-compacted ridges. It also aimed to evaluate the impact of ridge re-compaction on soil temperature and moisture retention. Five weeding trials were conducted in corn fields from 2022 to 2024 using two different ridge cultivators, Glühfosator and Damm Profi. The treatments included hoeing (HOE-2), hoeing combined with band herbicide spraying (HOE-1), hoeing combined with living mulch sown in the ridge valleys (HOE-3), and hoeing combined with postemergence harrow (HOE-4). Nontreated control and broadcast herbicide plots were included as controls. Soil moisture and temperature were recorded at 20-min intervals from May to September. Weed species composition, weed biomass, and corn silage yield were measured. Broadleaf weeds were the dominant weed species observed in all corn trials. In most trials, the hoeing treatments were not significantly different from that of applying a broadcast herbicide. Interrow areas treated with side-cut knives and ridge re-builders (HOE-1, HOE-2, and HOE-3) produced a significantly reduced weed biomass (4 to 55 g m−2) and exhibited high (80% to 96% weed control efficacy (WCE) across all trials. Intrarow-treated areas (i.e., tops of ridges) with a band herbicide (HOE-1), no-till sweeps (HOE-2, HOE-3), and postemergence harrow (HOE-4) resulted in 88% to 100%, 30% to 63%, and 17% WCE, respectively. Depending on corn cultivar and ridge cultivator, the HOE-1, HOE-4, and HOE-2 treatments resulted in corn silage yield that was similar to or greater than that of a broadcast herbicide. Yield was increased by 2000 to 9000 kg ha−1 after the HOE-1 treatment, by 2000 to 5000 kg ha−1 after the HOE-4 treatment, and by 3000 to 6000 kg ha−1, after the HOE-2 treatment. When rainfall was limited, re-compacted ridges demonstrated moisture conservation, which resulted in higher day-warming and lower night-cooling of ridge valleys (compared to ridge areas and flat-tilled beds), whereas when rain is heavy, ridges drained moisture and exhibited higher day-warming and lower night-cooling of ridge areas. These results suggest that precision hoeing on ridges could alternate broadcast herbicide use, while re-compacted ridges prove resilient to extreme rainfall events.

Hospital food services and the resulting food waste impact patient satisfaction, health outcomes, healthcare costs, and the environment. This cross-sectional study assessed food waste and patient satisfaction in five public hospitals in Cyprus, involving 844 inpatients. Patient characteristics and responses to the 21-item Acute Care Hospital Foodservice Patient Satisfaction Questionnaire (ACHFPSQ) were recorded. Plate waste was evaluated using photographs and a five-point visual scale (0 to 1) to estimate food consumption. Hunger and overall satisfaction were also assessed. While 77.8% rated food services as good or very good, food quality received the most negative feedback. Only 31.2% finished their main dish entirely; 29.5% and 26.3% left ¼ and ½, respectively. For dessert, 48.2% finished it, while 13.3% left it untouched. These findings reveal a gap between general satisfaction and perceived food quality, underscoring the need for targeted public health strategies to enhance food quality and reduce waste in hospitals.

We introduce the “Fork Game,” a graphical interface designed to elicit higher-order risk preferences. In this game, participants connect forked pipes to create a final structure. A ball is then dropped into the top opening of this structure and follows a downward path, randomly turning left or right at each forked joint. This construction is effectively isomorphic to the apportionment of binary-outcome lotteries, allowing participants to construct complex gambles. Furthermore, the game is easily comprehensible, highly modular, and provides a flexible means of assessing risk aversion, prudence, temperance, and even higher-order risk preferences.

Results obtained by transformer-based token classification models are now considered to be a benchmark for the Automatic Terminology Extraction (ATE) task. However, the unsatisfactory results (they rarely exceed 0.7 of the F1 value) raise the question of whether this approach is correct and of what text features are being remembered or inferred by the model trained on this type of annotation. In the paper, we describe a number of experiments using the fine-tuned RoBERTa base model on the ACTER data, RD-TEC, and three Wikipedia articles, which proved that the results of the ATE task obtained by such models depend considerably on the type of texts being processed and their relationship to the training data. While the results are relatively good for some texts with highly specialized vocabulary, the poor results seem to correlate with the high frequency (in general English texts) of tokens that are part of terms in a particular domain. Another property that affects the results is the degree of overlap between the vocabulary of the test data and the vocabulary of terms from the training data. Words that have been labeled as terms in the training data are usually labeled as terms in other, unrelated domains as well. Moreover, we show that the results obtained by these models are unstable—models trained on more data do not include all the items identified by models trained on a smaller dataset and can present substantially lower performance.

Where dual-numbers forward-mode automatic differentiation (AD) pairs each scalar value with its tangent value, dual-numbers reverse-mode AD attempts to achieve reverse AD using a similarly simple idea: by pairing each scalar value with a backpropagator function. Its correctness and efficiency on higher-order input languages have been analysed by Brunel, Mazza and Pagani, but this analysis used a custom operational semantics for which it is unclear whether it can be implemented efficiently. We take inspiration from their use of linear factoring to optimise dual-numbers reverse-mode AD to an algorithm that has the correct complexity and enjoys an efficient implementation in a standard functional language with support for mutable arrays, such as Haskell. Aside from the linear factoring ingredient, our optimisation steps consist of well-known ideas from the functional programming community. We demonstrate the use of our technique by providing a practical implementation that differentiates most of Haskell98. Where previous work on dual numbers reverse AD has required sequentialisation to construct the reverse pass, we demonstrate that we can apply our technique to task-parallel source programs and generate a task-parallel derivative computation.

While speakers are theorized to ideally not include unnecessary information (redundancy) in their utterances, in reality, they often do so. One potential reason is that linguistic redundancy facilitates language communication, especially when the addressee (interlocutor) is linguistically less competent (e.g., an artificial system). In three experiments, we examined whether linguistic redundancy may arise as a result of people’s tendency to use similar linguistic features as their interlocutor does during communication (i.e., linguistic alignment) and whether redundancy alignment (if any) differs with a human interlocutor versus a computer interlocutor. We also examined whether redundancy alignment is affected by the perceived competency of the interlocutor, participants’ abilities in theory of mind (ToM), and if redundancy alignment varied across time during the experiment. Participants carried out a picture matching and naming task with a human or computer interlocutor who either always or never included redundancies in their utterances. Redundancy alignment was found across all experiments, in that speakers produced more redundancies with a redundant interlocutor compared to a non-redundant one. This alignment was also modulated by the perceived competency of the interlocutor, the time course of the interaction, and ToM abilities, suggesting that redundancy usage is affected by both automatic and strategic mechanisms of linguistic alignment.

In 1919, the League of Nations and the Mandate System were established through the Treaty of Versailles. Shy of 100 years later, the International Rights of Nature Tribunal gathered in the same city to establish itself as an international peoples’ tribunal, taking form outside the international legal order. For all their differences, these institutions shared some commonalities. Both institutions claimed to be concerned with the ‘wellbeing’ of ‘peoples’ through international law. Both also claimed to represent some kind of international legal community. For the League, this was a community of states. For the Tribunal, it was a community of peoples. This article reads these institutional moments together and considers what they tell us about the discipline of international law. It traces how both institutions constituted and authorized themselves as if speaking for an already given international legal community – and how they did so precisely by mobilizing competing ideas of ‘nature’, ‘peoples’, and ‘statehood’. The article argues that the institutions deployed similar legal techniques and narratives that limited what the international legal domain might, and might not, look like. Namely, they presented their ideas of ‘nature’ and ‘peoples’ as part of a natural order of things, authorizing themselves as vanguards for whatever form of international legal order they saw as ‘natural’. Ultimately, the piece complicates our understanding of the international legal domain and peoples’ tribunal in it, inviting a reflexivity over what it means to speak law in the name of an international community – be it as ‘states’, ‘peoples’ or ‘nature’ itself.