The death of a child affects the well-being of parents and families worldwide, but little is known about the scale of this phenomenon. Using a novel methodology from formal demography applied to data from the 2019 Revision of the United Nations World Population Prospects, we provide the first global overview of parental bereavement, its magnitude, prevalence, and distribution over age for the 1950–2000 annual birth cohorts of women. We project that the global burden of parental bereavement will be 1.6 times lower for women born in 2000 than for women born in 1955. Accounting for compositional effects, we anticipate the largest improvements in regions of the Global South, where offspring mortality continues to be a common life event. This study quantifies an unprecedented shift in the timing of parental bereavement from reproductive to retirement ages. Women in the 1985 cohort and subsequent cohorts will be more likely to lose an adult child after age 65 than to lose a young child before age 50, reversing a long-standing global trend. “Child death” will increasingly come to mean the death of adult offspring. We project persisting regional inequalities in offspring mortality and in the availability of children in later life, a particular concern for parents dependent on support from their children after retirement. Nevertheless, our analyses suggest a progressive narrowing of the historical gap between the Global North and South in the near future. These developments have profound implications for demographic theory and highlight the need for policies to support bereaved older parents.

A growing body of research has examined whether birth intervals influence perinatal outcomes and child health as well as long-term educational and socioeconomic outcomes. To date, however, very little research has examined whether birth spacing influences long-term health. We use contemporary Swedish population register data to examine the relationship between birth-to-birth intervals and a variety of health outcomes in adulthood: for men, height, physical fitness, and the probability of falling into different body mass index categories; and for men and women, mortality. In models that do not adjust carefully for family background, we find that short and long birth intervals are clearly associated with height, physical fitness, being overweight or obese, and mortality. However, after carefully adjusting for family background using a within-family sibling comparison design, we find that birth spacing is generally not associated with long-term health, although we find that men born after very long birth intervals have a higher probability of being overweight or obese in early adulthood. Overall, we conclude that birth intervals have little independent effect on long-term health outcomes.

A growing literature has demonstrated a relationship between parity and mortality, but the explanation for that relationship remains unclear. This study aims to pick apart physiological and social explanations for the parity-mortality relationship by examining the mortality of parents who adopt children, but who have no biological children, in comparison with the mortality of parents with biological children. Using Swedish register data, we study post-reproductive mortality amongst women and men from cohorts born between 1915 and 1960, over ages 45-97. Our results show the relative risks of mortality for adoptive parents are always lower than those of parents with biological children. Mortality amongst adoptive parents is lower for those who adopt more than one child, while for parents with biological children we observe a U-shaped relationship, where parity-two parents have the lowest mortality. Our discussion considers the relative importance of physiological and social depletion effects, and selection processes.

Despite the large literature examining predictors of fertility, previous research has not offered a population-level perspective on how health in early adulthood is related to male fertility. Using Swedish population and military conscription registers, we study how body mass index (BMI), physical fitness, and height are associated with total fertility and parity transitions by 2012 among 405,427 Swedish men born 1965-1972, meaning we observe fertility up to age 40 or older. Applying linear regression and sibling fixed effects, we find that these anthropometric measures are strong predictors of fertility, even after accounting for education and cumulative income. Men with a normal BMI and in the highest decile of physical fitness have the most children. Men who were obese at ages 17-20 had a relative probability of childlessness almost twice as high as men who had a normal BMI, and men in the bottom decile of physical fitness had a relatively probability of childlessness more than 50 percent higher than men in the top decile. In sibling comparison models the tallest men have the most children and men in the lowest two deciles of height have significantly lower fertility. Further analyses show that the strong associations persist even among men who married.

We examine the relationship between birth-to-birth intervals and a variety of mid- and long-term cognitive and socioeconomic outcomes, including high school GPA, cognitive ability, educational attainment, earnings, unemployment status, and receiving government welfare support. Using contemporary Swedish population register data and a within-family sibling comparison design, we find that neither the birth interval preceding the index person nor the birth interval following the index person are associated with any substantively meaningful changes in mid- or long-term outcomes. This is true even for individuals born before or after birth-to-birth intervals of less than 12 months. We conclude that in a contemporary high-income welfare state, there appears to be no relationship between unusually short or long birth intervals and adverse long-term outcomes.

While levels of migration within countries have been trending down in a number of advanced economies, Sweden has recorded a rise in internal migration among young adults. An increase in aggregate migration levels can be the result of a decline in immobility (i.e. the absence of migration), an increase in repeat movement or a combination of both. In this paper, we draw on retrospective survey and longitudinal register data to explore the demographic mechanisms underpinning the rise in internal migration among young Swedes born in the 30 years to 1980 and we compare the migration behaviour of the youngest cohort to that of their European counterparts. Of all 25 European countries, Sweden reports the highest level of migration among young adults, which is the result of very low immobility combined with high repeat movement. The increase in migration has been particularly pronounced for inter-county moves for the post-1970 cohorts. Analysis of order-specific components of migration shows that this is the result of a decrease in immobility combined with a modest rise in higher-order moves, whereas it is the rise in higher-order moves that underpins the increase in inter-parish migration. This upswing has been accompanied by a shift in the ages at migration, characterised by an earlier start and later finish leading to a lengthening of the number of years young adults are mobile. The results indicate that change in migration behaviour is order-specific, which underlines the need to collect and analyse migration by move order to obtain a reliable account of migration trends.

Ethno-racial workplace segregation increases already existing ethno-racial inequality. While previous research has identified discriminatory employers as drivers of workplace segregation, this study addresses the role of the employees. Sociological and social psychological theory suggest that people prefer to surround themselves with people who positively confirm their social identity or who contribute with higher group status. Through web-based surveys, we measure employee attitudes and preferences concerning ethno-racial workplace diversity, to what extent they differ by ethnicity/race, and if they contain intersectional patterns. Thereafter, we use simulation models to analyze the consequences for workplace segregation that these preferences would have, if realized. The main survey results showed that all ethno-racial groups favored their own in-group as colleagues, especially European Americans. As a secondary choice, the respondents preferred the out-group with the highest labor market status. Intersectional patterns were identified, as minority women were preferred as colleagues over minority men. Our simulation model, based on the results of two surveys on stated vs. indirectly revealed preferences, showed that employee preferences were at best not diverse enough to desegregate workplaces. When based on the most common preferences (i.e. excluding a few outliers), the simulations even suggested that these preferences can cause segregation. We relate these findings to Schelling's model of segregation.

A widely accepted view in the cultural evolutionary literature is that culture forms a dynamic system of elements (or 'traits') linked together by a variety of relationships. Despite this, large families of models within the cultural evolutionary literature tend to represent only a small number of traits, or traits without interrelationships. As such, these models may be unable to capture complex dynamics resulting from multiple interrelated traits. Here we put forward a systems approach to cultural evolutionary research-one that explicitly represents numerous cultural traits and their relationships to one another. Basing our discussion on simple graph-based models, we examine the implications of the systems approach in four domains: (i) the cultural evolution of decision rules ('filters') and their influence on the distribution of cultural traits in a population; (ii) the contingency and stochasticity of system trajectories through a structured state space; (iii) how trait interrelationships can modulate rates of cultural change; and (iv) how trait interrelationships can contribute to understandings of inter-group differences in realised traits. We suggest that the preliminary results presented here should inspire greater attention to the role of multiple interrelated traits on cultural evolution, and should motivate attempts to formalise the rich body of analyses and hypotheses within the humanities and social science literatures.

To examine the evolutionary basis of a behavior, an established approach (known as the phenotypic gambit) is to assume that the behavior is controlled by a single allele, the fitness effects of which are derived from a consideration of how the behavior interacts, via life-history, with other ecological factors. Here we contrast successful applications of this approach with several examples of an influential and superficially similar line of research on the evolutionary basis of human cooperation. A key difference is identified: in the latter line of research the focal behavior, cooperation, is abstractly defined in terms of immediate fitness costs and benefits. Selection is then assumed to act on strategies in an iterated social context for which fitness effects can be derived by aggregation of the abstractly defined immediate fitness effects over a lifetime. This approach creates a closed theoretical loop, rendering models incapable of making predictions or providing insight into the origin of human cooperation. We conclude with a discussion of how evolutionary approaches might be appropriately used in the study of human social behavior.

Behaving efficiently (optimally or near-optimally) is central to animals' adaptation to their environment. Much evolutionary biology assumes, implicitly or explicitly, that optimal behavioural strategies are genetically inherited, yet the behaviour of many animals depends crucially on learning. The question of how learning contributes to optimal behaviour is largely open. Here we propose an associative learning model that can learn optimal behaviour in a wide variety of ecologically relevant circumstances. The model learns through chaining, a term introduced by Skinner to indicate learning of behaviour sequences by linking together shorter sequences or single behaviours. Our model formalizes the concept of conditioned reinforcement (the learning process that underlies chaining) and is closely related to optimization algorithms from machine learning. Our analysis dispels the common belief that associative learning is too limited to produce ‘intelligent’ behaviour such as tool use, social learning, self-control or expectations of the future. Furthermore, the model readily accounts for both instinctual and learned aspects of behaviour, clarifying how genetic evolution and individual learning complement each other, and bridging a long-standing divide between ethology and psychology. We conclude that associative learning, supported by genetic predispositions and including the oft-neglected phenomenon of conditioned reinforcement, may suffice to explain the ontogeny of optimal behaviour in most, if not all, non-human animals. Our results establish associative learning as a more powerful optimizing mechanism than acknowledged by current opinion.

Although much research has found girls to be less interested in mathematics than boys are, there are many countries in which the opposite holds. I hypothesize that variation in gender differences in interest are driven by a complex process in which national culture promoting high math achievement drives down interest in math schoolwork, with the effect being amplified among girls due to their higher conformity to peer influence. Predictions from this theory were tested in a study of data on more than 500,000 grade 8 students in 50 countries from the 2011 and 2015 waves of TIMSS. Consistent with predictions, national achievement levels were strongly negatively correlated with national levels of math schoolwork interest and this variation was larger among girls: girls in low-achievement, high-interest countries had especially high interest in math schoolwork, whereas girls in high-achievement, low-interest countries had especially low interest in math schoolwork. Gender differences in math schoolwork interest were also found to be related to gender differences in math achievement, emphasizing the importance of understanding them better.

Experiments on economic games typically fail to find positive reputational effects of using peer punishment of selfish behavior in social dilemmas. Theorists had expected positive reputational effects because of the potentially beneficial consequences that punishment may have on norm violators' behavior. Going beyond the game-theoretic paradigm, we used vignettes to study how various social factors influence approval ratings of a peer who reprimands a violator of a group-beneficial norm. We found that ratings declined when punishers showed anger, and this effect was mediated by perceived aggressiveness. Thus the same emotions that motivate peer punishers may make them come across as aggressive, to the detriment of their reputation. However, the negative effect of showing anger disappeared when the norm violation was sufficiently severe. Ratings of punishers were also influenced by social distance, such that it is less appropriate for a stranger than a friend to reprimand a violator. In sum, peer punisher ratings were very high for a friend reprimanding a severe norm violation, but particularly poor for a stranger showing anger at a mild norm violation. We found no effect on ratings of whether the reprimand had the beneficial consequence of changing the violator's behavior. Our findings provide insight into how peer punishers can avoid negative reputational effects. They also point to the importance of going beyond economic games when studying peer punishment.

A requirement of culture, whether animal or human, is some degree of conformity of behavior within populations. Researchers of gene-culture coevolution have suggested that population level conformity may result from frequency-biased social learning: individuals sampling multiple role models and preferentially adopting the majority behavior in the sample. When learning from a single role model, frequency-bias is not possible. We show why a population-level trend, either conformist or anticonformist, may nonetheless be almost inevitable in a population of individuals that learn through social enhancement, that is, using observations of others' behavior to update their own probability of using a behavior in the future. The exact specification of individuals' updating rule determines the direction of the trend. These results offer a new interpretation of previous findings from simulations of social enhancement in combination with reinforcement learning, and demonstrate how results of dynamical models may strongly depend on seemingly innocuous choices of model specifications, and how important it is to obtain empirical data on which to base such choices.

A strategy activated in one task may be transferred to subsequent tasks and prevent activation of other strategies that would otherwise come to mind, a mechanism referred to as procedural priming. In a novel application of procedural priming we show that it can make or break cognitive illusions. Our test case is the 1/k illusion, which is based on the same unwarranted mathematical shortcut as the MPG illusion and the time-saving bias. The task is to estimate distances between values of fractions on the form 1/k. Most people given this task intuitively base their estimates on the distances between the denominators (i.e., the reciprocals of the fractions), which may yield very poor estimations of the true distances between the fractions. As expected, the tendency to fall for this illusion is related to cognitive style (Study 1). In order to apply procedural priming we constructed versions of the task in which the illusion is weak, in the sense that most people do not fall for it anymore. We then gave participants both strong illusion and weak illusion versions of the task (Studies 2 and 3). Participants who first did the task in the weak illusion version would often persist with the correct strategy even in the strong illusion version, thus breaking the otherwise strong illusion in the latter task. Conversely, participants who took the strong illusion version first would then often fall for the illusion even in the weak illusion version, thus strengthening the otherwise weak illusion in the latter task.

One of the many things teachers do is to give feedback on their students' work. Feedback pointing out mistakes may be a key to learning, but it may also backfire. We hypothesized that feedback based on students' mistakes may have more positive effects in cultures where teachers have greater authority over students, which we assume to be cultures that are high on power distance and religiosity. To test this hypothesis we analyzed data from 49 countries taking part in the 2015 wave of the TIMSS assessment, in which students in the 4th and 8th grades were asked whether their teachers in mathematics and science told them how to do better when they had made a mistake. For each country we could then estimate the association between the reported use of mistake-based feedback and student achievement. Consistent with our hypothesis, the estimated effect of mistake-based feedback was positive only in certain countries, and these countries tended to be high on power distance and religiosity. These results highlight the importance of cultural values in educational practice.

Prior research using the Moral Foundations Questionnaire (MFQ) has established that political ideology is associated with self-reported reliance on specific moral foundations in moral judgments of acts. MFQ items do not specify the agents involved in the acts, however. By specifying agents in MFQ items we revealed blatant political double standards. Conservatives thought that the same moral foundation was more relevant if victims were agents that they like (i.e., corporations and other conservatives) but less relevant when the same agents were perpetrators. Liberals showed the same pattern for agents that they like (i.e., news media and other liberals). A UK sample showed much weaker political double standards with respect to corporations and news media, consistent with feelings about corporations and news media being much less politicized in the UK than in the US. We discuss the implications for moral foundations theory.

In markets with asymmetric information, only sellers have knowledge about the quality of goods. Sellers may of course make a declaration of the quality, but unless there are sanctions imposed on false declarations or reputations are at stake, such declarations are tantamount to cheap talk. Nonetheless, in an experimental study we find that most people make honest declarations, which is in line with recent findings that lies damaging another party are costly in terms of the liar’s utility. Moreover, we find in this experimental market that deceptive sellers offer lower prices than honest sellers, which could possibly be explained by the same wish to limit the damage to the other party. However, when the recipient of the offer is a social tie we find no evidence for lower prices of deceptive offers, which seems to indicate that the rationale for the lower price in deceptive offers to strangers is in fact profit-seeking (by making the deal more attractive) rather than moral.

Cultural variants may spread by being more appealing, more memorable, or less offensive than other cultural variants. Empirical studies suggest that such emotional selection is a force to be reckoned with in cultural evolution. We present a research paradigm that is suitable for the study of emotional selection. It guides empirical research by directing attention to the circumstances under which emotions influence the likelihood that an individual will influence another individual to acquire a cultural variant. We present a modeling framework to translate such knowledge into specific and testable predictions of population-level change. A set of already analyzed basic cases can serve as a toolbox.

Violators of cooperation norms may be informally punished by their peers. How such norm enforcement is judged by others can be regarded as a meta-norm (i.e., a second-order norm). We examined whether meta-norms about peer punishment vary across cultures by having students in eight countries judge animations in which an agent who over-harvested a common resource was punished either by a single peer or by the entire peer group. Whether the punishment was retributive or restorative varied between two studies, and findings were largely consistent across these two types of punishment. Across all countries, punishment was judged as more appropriate when implemented by the entire peer group than by an individual. Differences between countries were revealed in judgments of punishers vs. non-punishers. Specifically, appraisals of punishers were relatively negative in three Western countries and Japan, and more neutral in Pakistan, UAE, Russia, and China, consistent with the influence of individualism, power distance, and/or indulgence. Our studies constitute a first step in mapping how meta-norms vary around the globe, demonstrating both cultural universals and cultural differences.

The ultimatum game is a common economic experiment in which some participants reject another's unfair offer of how to split some money, even though it leaves them both worse off. This costly behavior can be seen as enforcement of a fairness norm and has been labeled “altruistic punishment”, suggesting that it is a moral thing to do. But is this behavior viewed as moral by participants? Is it viewed as punishment? And are the payoff consequences of the behavior sufficient to determine the answers to these questions? To investigate this we framed costly punishment in two different ways: either as rejection of an offer (the standard ultimatum game framing) or as reduction of payoff. In a series of paid and hypothetical experiments we found that moral concerns about costly punishment depended on the framing. Specifically, the reduction frame elicited more moral concern about, and less use of, costly punishment than did the rejection frame. Several implications are discussed.

Norm enforcement may be important for resolving conflicts and promoting cooperation. However, little is known about how preferred responses to norm violations vary across cultures and across domains. In a preregistered study of 57 countries (using convenience samples of 22,863 students and non-students), we measured perceptions of the appropriateness of various responses to a violation of a cooperative norm and to atypical social behaviors. Our findings highlight both cultural universals and cultural variation. We find a universal negative relation between appropriateness ratings of norm violations and appropriateness ratings of responses in the form of confrontation, social ostracism and gossip. Moreover, we find the country variation in the appropriateness of sanctions to be consistent across different norm violations but not across different sanctions. Specifically, in those countries where use of physical confrontation and social ostracism is rated as less appropriate, gossip is rated as more appropriate. Little is known about people's preferred responses to norm violations across countries. Here, in a study of 57 countries, the authors highlight cultural similarities and differences in people's perception of the appropriateness of norm violations.

Vaccine hesitancy is a threat to global health, but it is not ubiquitous; depending on the country, the proportion that have confidence in vaccines ranges from a small minority to a huge majority. Little is known about what explains this dramatic variation in vaccine confidence. We hypothesize that variation in religiosity may play a role because traditional religious teachings are likely to be incompatible with the specific magical/spiritual health beliefs that often undergird anti-vaccination sentiments. In analyses of publicly available data in 147 countries, we find that a country measure of religiosity is strongly positively correlated with country measures of confidence in the safety, importance, and effectiveness of vaccines, and these associations are robust to controlling for measures of human development (education, economic development, and health). The underlying mechanism needs to be examined in future research.

Does selfishness pay in the long term? Previous research has indicated that being prosocial (or otherish) rather than selfish has positive consequences for psychological well-being, physical health, and relationships. Here we instead examine the consequences for individuals' incomes and number of children, as these are the currencies that matter most in theories that emphasize the power of self-interest, namely economics and evolutionary thinking. Drawing on both cross-sectional (Studies 1 and 2) and panel data (Studies 3 and 4), we find that prosocial individuals tend to have more children and higher income than selfish individuals. An additional survey (Study 5) of lay beliefs about how self-interest impacts income and fertility suggests one reason selfish people may persist in their behavior even though it leads to poorer outcomes: people generally expect selfish individuals to have higher incomes. Our findings have implications for lay decisions about the allocation of scarce resources, as well as for economic and evolutionary theories of human behavior.

The Maderö wreck was discovered in the 1960s in the Stockholm Archipelago, Sweden. An archaeological investigation undertaken in 2022 included the inspection and documentation of visible ship parts, sampling for dendrochronological analysis and sampling for ICP analysis from the brick cargo. The results show that the wood originates from the Baltic Sea area and was felled after 1467, while the clay for the brick originates from the Mecklenburg-Vorpommern area. The ship's technical analysis shows that it is a large clinker-built merchant ship. Traces of iron on a recovered stone shot indicate that the ship was armed when it sank.

An unusual ski from the Låktatjåhkå / Loktacohkka glacier in Sápmi – Use, wax, 14C and lipid residue analysis: Archaeological skis dated from the Stone Age to today, are not unusual finds in bogs and wetlands. They are found all over Sápmi, from Norway to the Kola peninsula in the Russian federation. Skis are also represented in rock art at different sites in Sápmi and mentioned in written sources; however, skis found at higher altitudes, at glaciers and perennial snow patches, are not as common. In 2018, nine kilometers west of Björkliden, a fragment of a ski was found by the Loktačohkka glacier. The ski fragment was first 14C-dated to the 15th century; however, presence of vax on the fragment presented an interesting problem. A new 14C analysis of the ski fragment, with the wax components removed, now dated the cellulose from the ski to 1645–1916 CE, i.e. the ski could have been used some time from circa 1645 into the first half of the 20th century. This study highlights the importance of regular surveys of melting glaciers and snow patches to retrieve organic material melting out, as well as the importance of investigating what components wood could have been impregnated with.

Bridging associative and normative theories of animal learning, I show that an associative system can behave as if performing probabilistic inference by using the function f(V) = 1 − e^−cV to transform associative strengths (V) into response probabilities. For example, using this function, an associative system can respond normatively to a compound stimulus AB, given previous separate experiences with the components A and B. The CR probability formulae that result from the proposed function have a normative interpretation in terms of statistical decision theory. The formulae also suggest a normative interpretation of stimulus generalization as a heuristic to infer whether different stimuli are likely to convey redundant or independent information about reinforcement. (PsycInfo Database Record (c) 2022 APA, all rights reserved)

Ravignani et al (2013) abituated squirrel monkeys to sound sequences conforming to an ABnA grammar (n = 1, 2, 3), then tested them for their reactions to novel grammatical and non -grammatical sequences. Although they conclude that the monkeys consistently recognized and generalized the sequence AB(n)A, I remark that this conclusion is not robust. The statistical significance of results depends on specific choices of data analysis, namely dichotomization of the response variable and omission of specific data points. Additionally, there is little evidence of generalization to novel patterns (n = 4, 5), which is important to conclude that the monkeys recognized the AB(n)A grammar beyond the habituation patterns. Lastly, many test sequences were perceptually similar to habituation sequences, raising the possibility that the monkeys may have generalized based on perceptual similarity rather than based on grammaticality.

In summation experiments, responding to a compound stimulus is assessed after conditioning a response to each of its components. This simple experiment poses significant challenges to models of associative learning because of substantial variability in results. Here, I introduce a new method to quantify generalization from components to compound in summation experiments, which I apply to over 250 measurements of summation in rabbits, pigeons, rats, and humans. The analysis confirms that more summation occurs with stimuli from different rather than from the same sensory modality, although this is not the sole determinant of summation. A theoretical analysis shows that this finding is best accounted for by a model that includes both element sharing (Rescorla & Wagner, 1972) and element replacement (Brandon et al., 2000) in stimulus representations. I point out remaining gaps in our empirical and theoretical understanding of summation. 

I introduce a simple mathematical method to calculate the associative strengths of stimuli in many models of associative learning, without solving the models' learning equations and without simulating the learning process. The method applies to many models, including the Rescorla and Wagner (1972) model, the replaced elements model of Brandon et al. (2000), and Pearce's (1987) configural model. I illustrate the method by calculating the predictions of these three models in summation and blocking experiments, allowing for a degree of similarity between the training stimuli as well as for the effects of contextual stimuli. The method clarifies the models' predictions and suggests new empirical tests.

The Rescorla and Wagner (1972) model is the first mathematical theory to explain associative learning in the presence of multiple stimuli. Its main theoretical construct is that of associative strength, but this is connected to behavior only loosely. We propose a model in which behavior is described by a collection of Poisson processes, each with a rate proportional to an associative strength. The model predicts that the time between behaviors follows an exponential or hypoexponential distribution. This prediction is supported by two data sets on autoshaped and instrumental behavior in rats.

A defining feature of Pavlovian conditioning is that the unconditioned stimulus (US) is delivered whether or not the animal performs a conditioned response (CR). This has lead to the question: Does CR performance play any role in conditioning? Between the 1930s and 1970s. a consensus emerged that CR acquisition is driven by CS-US (CS, conditioned stimulus) experiences, and that CRs play a minimal role, if any. Here we revisit the question and present 2 new quantitative methods to evaluate whether CRs influence the course of learning. Our results suggest that CRs play an important role in Pavlovian acquisition, in such paradigms as rabbit eye blink conditioning, pigeon autoshaped key pecking, and rat autoshaped lever pressing and magazine entry.

Humans stand out among animals for their unique capacities in domains such as language, culture and imitation, yet it has been difficult to identify cognitive elements that are specifically human. Most research has focused on how information is processed after it is acquired, e.g. in problem solving or 'insight' tasks, but we may also look for species differences in the initial acquisition and coding of information. Here, we show that non-human species have only a limited capacity to discriminate ordered sequences of stimuli. Collating data from 108 experiments on stimulus sequence discrimination (1540 data points from 14 bird and mammal species), we demonstrate pervasive and systematic errors, such as confusing a red-green sequence of lights with green-red and green-green sequences. These errors can persist after thousands of learning trials in tasks that humans learn to near perfection within tens of trials. To elucidate the causes of such poor performance, we formulate and test a mathematical model of non-human sequence discrimination, assuming that animals represent sequences as unstructured collections of memory traces. This representation carries only approximate information about stimulus duration, recency, order and frequency, yet our model predicts non-human performance with a 5.9% mean absolute error across 68 datasets. Because human-level cognition requires more accurate encoding of sequential information than afforded by memory traces, we conclude that improved coding of sequential information is a key cognitive element that may set humans apart from other animals.