Early human cultural dynamics underpin the Plio-Pleistocene archaeological record and impact how we understand some of our earliest identifiable behaviours. One major outstanding question is whether Early Stone Age material culture represents a single lineage of cultural information, or did we ever lose the knowledge required to make stone tools? No single approach satisfactorily addresses this problem, but to date, objective analyses of temporal data have been absent from the conversation. Here, using a comprehensive database of dated African Oldowan archaeological sites, we demonstrate that there are no temporal breaks large enough, on a relative basis, to infer a loss of stone-tool-making cultural information. Therefore, alongside previously published data, we infer a continuous record of early human stone tool production in Africa from c. 3.3 to 1.5 million years ago. Stone tool-associated behavioural adaptations and evolutionary selective pressures were, therefore, likely to have been ever present during this period.

Preventing human-caused extinctions is a foundational aim of conservation. However, in addition to causing extinctions, humans have moved numerous species to new areas. A considerable percentage of these are threatened in their native ranges. Broadening our conservation ethos to include introduced species is contentious and requires critical thinking in empirical and normative dimensions to negotiate between conflicting conservation goals. Here, we present a series of questions to inspire critical thinking in the negotiation of these conflicts. Empirically, we suggest that conservationists should consider whether the effects of introduced species are due to their non-nativeness per se or are simply a consequence of the organism having a metabolism and taking up space. Importantly, this requires proper scientific comparison to the effects of similar native organisms – otherwise many claims of ‘harm’ are unfalsifiable and could be used to justify the eradication of any organism. We further propose questions to help conservationists sort facts from normative values, which often wear empirical clothes. Through empirical rigor, value transparency and critical justification of these values, we believe that twenty- first century conservation can become a future-facing and pluralistic discipline with a heightened ability to prevent extinctions in an increasingly unpredictable and novel biosphere.

Scientists still debate whether small groups of Paleolithic hunter-gatherers caused the extinction of large Ice Age animals like prehistoric elephants, giant sloths and cave lions. Beyond paleontology, this question has deep sociological implications and is relevant for how we understand the role of humankind in today’s environmental crisis. A human-driven megafauna extinction has often fostered the idea of a naturalization of human environmental impacts and the belief that all people (modern or ancient, rich or poor, from any part of the world) share responsibility for the current crisis. But is that true? In the review, I discuss whether a long evolutionary history of impacts really makes us inevitably destructive, compelling humanity to accept a devastating anthropocentric dominance as the fateful destiny natural selection built for us. In contrast, I argue that, while our exceptional ability to shape environments has made us a ‘hyper-keystone’ species, benefiting only a few species and humans, this same ability also has the potential to help us restore balance to the world. That requires rejecting anthropocentric supremacy and placing ecosystems at the center stage of our relationship with nonhuman nature. We may have wiped out the mammoths and mastodons, but human destructiveness is not fate.

Rewilding is a transformative conservation strategy that emphasises the restoration of ecological processes and ecosystem resilience. This perspective article addresses limitations in current rewilding monitoring practices, which predominantly rely on quantitative metrics. We argue for the integration of qualitative approaches to better capture the complexity and interdependence of human–nature interactions that shape rewilding outcomes. Drawing on social–ecological systems (SES) thinking, we propose a holistic monitoring framework that combines quantitative and qualitative measures. This approach reflects emerging shifts in conservation mindsets – recognising humans as part of nature, valuing biodiversity intrinsically as well as instrumentally, and embracing uncertainty and complexity over control. We suggest that Holling’s panarchy – a framework for understanding adaptive cycles and cross-scale interactions – can support the design of rewilding interventions and guide monitoring. By focusing on fast- and slow-changing variables, panarchy enhances adaptive management and supports context-sensitive theories of change. This article contributes to the rewilding discourse by offering practical guidance for practitioners and policymakers, promoting a shift towards inclusive, adaptive and transformative monitoring practices.

Generation length is a key parameter for extinction risk assessment in the IUCN Red List, defining the biological horizon over which population decline is measured. Generation length data are largely unavailable, and a naive time frame of 10 years is often used to represent the three-generation horizon required for assessing decline in short-lived species. This assumption risks introducing biases in the assessment. Here, we quantified how using a 10-year time frame, instead of one based on predicted generation lengths, affects the extinction risk assessments of amphibians: one of the most imperiled groups worldwide. We measured past deforestation trends for 1,278 forest-dependent amphibian species and predicted future climate-driven habitat change for 1,559 species. Using predicted generation lengths systematically increased the percentage of declining species from deforestation (3–29%) and climate change (1–20%). This was also reflected in a higher proportion of species potentially assessed as threatened. Relying on a 10-year period to assess extinction risk can lead to substantial underestimations across both past and future time frames, compared to using predicted generation length. Predictions of generation length can allow applying IUCN Red List criteria in the absence of empirical data and improve consistency of extinction risk assessments.

This paper examines the stone tool evidence associated with extinctions among Australopithecus, Paranthropus, and Homo in Eastern Africa between 0.8 and 3.5 Ma. It does this using Stoneworking Modes A-I, a relatively new framework for comparing stone tool evidence, and data from the Eastern African Prehistoric Stoneworking Survey. While the evolutionary transition between early Homo and H. ergaster/erectus seems correlated with a shift from occasional to habitual stoneworking, Australopithecus’ and Paranthropus’ last appearance dates do not appear correlated with major changes in the archaeological record.

Cooperation is widely viewed as a stabilizing force because it pools risk and buffers against environmental uncertainty. Demographic theory predicts that volatility in per-capita population growth declines rapidly with population size, implying that larger populations should be more stable. Yet human societies of all sizes eventually collapse. Here, we propose a general mechanism that helps explain this pattern. We show that cooperation generates demographic synchrony—correlated survival and reproductive outcomes among individuals—which inflates aggregate volatility. Using more than 200 longitudinal population records from indigenous societies in the Brazilian Amazon, we find that volatility declines with population size substantially slower than predicted under demographic independence. This pattern implies pervasive within-population correlation, such that census populations behave demographically as if composed of far fewer independent modules. We formalize this process with a synchrony-extended stochastic model in which shared environmental exposure, social coupling, and spatial overlap generate demographic correlations that erode the stabilizing effects of scale. We term this dynamic the cooperation–synchrony paradox: the same institutions that reduce risk for individuals can increase vulnerability to population-wide shocks. Further, we propose this is likely a general mechanism across species, where cooperation stabilizes locally but generates systemic fragility by reducing effective demographic dimensionality.

Faunal extinctions correlate with human colonization worldwide during the Quaternary, yet the extent to which human hunting drove these losses remains debated. Resolving these debates depends on accurately dating both human arrival and animal extinction events. Herein I present a simulated overkill scenario modeling interaction between a founding human population and a single prey species. The simulation evaluates expected age offsets, the differences between the actual and observed dates, for extinction and colonization as functions of sample size and human population growth rate. Results show that when extinction is driven by human hunting, age offsets for extinction are expected to be smaller than those for colonization. This difference in age offset is expected because overkill produces a rapid demographic collapse of prey populations and taphonomic processes concentrate late-surviving individuals in the fossil record. In contrast, colonization dates are more offset because human populations are smallest immediately after arrival, and archaeological rarity is compounded by taphonomic effects. Although colonization age offsets are consistently larger than extinction age offsets across sample sizes, they are not extreme, such that archaeological sites should be detectable within five centuries of initial human settlement under modest rates of human population growth.

Cambridge Prisms: Extinction was launched to provide an open, synthetic forum for research on the patterns, processes and consequences of extinction. Four years on, the release of the journal’s first impact factor offers a useful moment to look back at what has begun to take shape and to look ahead to what it should now become. The early content shows a young journal whose first visible scholarly impact has been driven largely by synthesis. That was deliberate: a review-led beginning helped define the journal’s intellectual map, draw together conservation biology, palaeontology, modelling, policy, social science and the humanities, and create pathways for more primary and data-driven work. The next stage should build on that foundation through more empirical research and through Extinction Accounts, a new article format designed to analyse individual extinctions.

Whether the last common ancestor (LCA) of Homo and Pan (ca. 6–7 Ma) used stone tools is a foundational question in palaeoanthropology. However, it is often treated as known rather than as a hypothesis requiring substantiation. Lacking fossil evidence for the LCA, this assumption is based on data from extant primates and the available hominin fossil record. Phylogenetic approaches invoking parsimony argue that the presence of stone tool use in both Pan and Homo reflects homology rather than convergent evolution. While methodologically legitimate, we argue that this framework has been applied without sufficient attention to the specific criteria for this to be true. For homology to hold, stone tool use must have been continuously expressed across generations of both lineages from the LCA to the present. We evaluate this claim against three lines of evidence: behavioural data from extant primates, showing that stone tool use is heterogeneous within Pan; the hominin fossil and archaeological record, which leaves a substantial evidential gap between the LCA and the earliest known documented technologies; and comparative data from phylogenetically distant tool using animals, which show that similar behaviours can arise independently under comparable ecological conditions. Together, these lines of evidence favour a model of multiple, context-dependent origins of stone tool use over unbroken inheritance from a tool using LCA.

Does one language really disappear every two weeks? The statistic is widely cited in discussions of language endangerment, yet its empirical basis remains unclear. This article traces the origins of the claim and reviews the evidence on which it rests. We first revisit the conjecture that half of the world’s languages were already no longer transmitted to the next generation in the early 1990s and examine how this assumption gave rise to the “one language every two weeks” figure. Using assessments from major linguistic catalogues, including the UNESCO Atlas and Ethnologue, we show that the proportion of languages no longer transmitted to younger generations is substantially lower than originally assumed. We then review more recent attempts to estimate the rate of language extinction from documented disappearances, rather than from current intergenerational transmission statuses. These studies suggest a much lower rate of roughly one language every six to thirteen weeks but rely on methods that are subject to important – yet difficult to quantify – sources of bias. We conclude that the widely cited claim that one language disappears every two weeks is probably too high, while the accuracy of more recent estimates cannot presently be determined. Drawing on approaches developed in biological extinction research, we conclude by outlining how methods that account for the population at risk, supported by data documenting the demographic trajectories of the world’s languages, could enable more robust estimates of the current rate of language extinction.

Global diversity patterns delineate the Serpukhovian (Upper Mississippian, Carboniferous) mass extinction as the sixth-largest of the Phanerozoic, with 26–28% of genera extinct. Studies of Serpukhovian marine assemblages from North America, however, exhibit stable diversity patterns across the Serpukhovian–Bashkirian boundary. These studies provide evidence against an extinction event during the Serpukhovian, with observed taxonomic turnover attributable to extirpation and facies preservation rather than true extinction. Here, we provide further insight into the Serpukhovian extinction at the mid-Carboniferous Global Boundary Stratotype Section and Point in Arrow Canyon, southern Nevada, US. Assessing brachiopod assemblages within their facies context, we test for evidence of the Serpukhovian mass extinction. Results reveal that 39% of Serpukhovian genera are absent from the Bashkirian at Arrow Canyon. However, global stratigraphic ranges reveal that all but two Serpukhovian taxa persisted into the Bashkirian. Multivariate analysis reveals similar taxonomic composition between the Serpukhovian and Bashkirian assemblages, with a notable lithological gradient driving local variation between geologic units. These results further suggest the absence of a late Serpukhovian extinction event in North America and highlight how working with the structure of the stratigraphic record can separate biological signals from stratigraphic patterns.

Time to extinction (TTE) is an estimate of how long a population, species or wider biological group persists before reaching a stated endpoint, such as extinction, local loss or a defined level of severe decline. The estimate depends on where the clock starts, what endpoint stops it, what evidence is available and how uncertainty is handled. In conservation biology, TTE is usually estimated by projecting possible population futures under specified threats and management choices. In historical and fossil records, TTE is inferred backwards from incomplete evidence such as sightings, specimens, fossils, dated remains, environmental DNA and non-detections. TTE complements extinction probability by showing how quickly loss is expected or inferred to occur, not just whether it occurs within a chosen period. The concept is useful for comparing collapse rates, intervention windows, extinction debts and uncertainty, provided each estimate states the biological entity, spatial scale, start point, endpoint, evidence base, assumptions and uncertainty.