The Language of Creativity: Evidence from Humans and Large Language Models.

Recent developments in computerized scoring via semantic distance have provided automated assessments of verbal creativity. Here, we extend past work, applying computational linguistic approaches to characterize salient features of creative text. We hypothesize that, in addition to semantic diversity, the degree to which a story includes perceptual details, thus transporting the reader to another time and place, would be predictive of creativity. Additionally, we explore the use of generative language models to supplement human data collection and examine the extent to which machine-generated stories can mimic human creativity. We collect 600 short stories from human participants and GPT-3, subsequently randomized and assessed on their creative quality. Results indicate that the presence of perceptual details, in conjunction with semantic diversity, is highly predictive of creativity. These results were replicated in an independent sample of stories (n = 120) generated by GPT-4. We do not observe a significant difference between human and AI-generated stories in terms of creativity ratings, and we also observe positive correlations between human and AI assessments of creativity. Implications and future directions are discussed.

Evidence for Spinozan "Unbelieving" in the Right Inferior Prefrontal Cortex.

Humans can think about possible states of the world without believing in them, an important capacity for high-level cognition. Here, we use fMRI and a novel "shell game" task to test two competing theories about the nature of belief and its neural basis. According to the Cartesian theory, information is first understood, then assessed for veracity, and ultimately encoded as either believed or not believed. According to the Spinozan theory, comprehension entails belief by default, such that understanding without believing requires an additional process of "unbelieving." Participants (n = 70) were experimentally induced to have beliefs, desires, or mere thoughts about hidden states of the shell game (e.g., believing that the dog is hidden in the upper right corner). That is, participants were induced to have specific "propositional attitudes" toward specific "propositions" in a controlled way. Consistent with the Spinozan theory, we found that thinking about a proposition without believing it is associated with increased activation of the right inferior frontal gyrus. This was true whether the hidden state was desired by the participant (because of reward) or merely thought about. These findings are consistent with a version of the Spinozan theory whereby unbelieving is an inhibitory control process. We consider potential implications of these results for the phenomena of delusional belief and wishful thinking.

Dual-process moral judgment beyond fast and slow.

De Neys makes a compelling case that the sacrificial moral dilemmas do not elicit competing "fast and slow" processes. But are there even two processes? Or just two intuitions? There remains strong evidence, most notably from lesion studies, that sacrificial dilemmas engage distinct cognitive processes generating conflicting emotional and rational responses. The dual-process theory gets much right, but needs revision.

Balancing selection shapes density-dependent foraging behaviour.

The optimal foraging strategy in a given environment depends on the number of competing individuals and their behavioural strategies. Little is known about the genes and neural circuits that integrate social information into foraging decisions. Here we show that ascaroside pheromones, small glycolipids that signal population density, suppress exploratory foraging in Caenorhabditis elegans, and that heritable variation in this behaviour generates alternative foraging strategies. We find that natural C. elegans isolates differ in their sensitivity to the potent ascaroside icas#9 (IC-asc-C5). A quantitative trait locus (QTL) regulating icas#9 sensitivity includes srx-43, a G-protein-coupled icas#9 receptor that acts in the ASI class of sensory neurons to suppress exploration. Two ancient haplotypes associated with this QTL confer competitive growth advantages that depend on ascaroside secretion, its detection by srx-43 and the distribution of food. These results suggest that balancing selection at the srx-43 locus generates alternative density-dependent behaviours, fulfilling a prediction of foraging game theory.

Veil-of-ignorance reasoning favors the greater good.

The "veil of ignorance" is a moral reasoning device designed to promote impartial decision making by denying decision makers access to potentially biasing information about who will benefit most or least from the available options. Veil-of-ignorance reasoning was originally applied by philosophers and economists to foundational questions concerning the overall organization of society. Here, we apply veil-of-ignorance reasoning in a more focused way to specific moral dilemmas, all of which involve a tension between the greater good and competing moral concerns. Across 7 experiments (n = 6,261), 4 preregistered, we find that veil-of-ignorance reasoning favors the greater good. Participants first engaged in veil-of-ignorance reasoning about a specific dilemma, asking themselves what they would want if they did not know who among those affected they would be. Participants then responded to a more conventional version of the same dilemma with a moral judgment, a policy preference, or an economic choice. Participants who first engaged in veil-of-ignorance reasoning subsequently made more utilitarian choices in response to a classic philosophical dilemma, a medical dilemma, a real donation decision between a more vs. less effective charity, and a policy decision concerning the social dilemma of autonomous vehicles. These effects depend on the impartial thinking induced by veil-of-ignorance reasoning and cannot be explained by anchoring, probabilistic reasoning, or generic perspective taking. These studies indicate that veil-of-ignorance reasoning may be a useful tool for decision makers who wish to make more impartial and/or socially beneficial choices.

Two Ways to Build a Thought: Distinct Forms of Compositional Semantic Representation across Brain Regions.

To understand a simple sentence such as "the woman chased the dog", the human mind must dynamically organize the relevant concepts to represent who did what to whom. This structured recombination of concepts (woman, dog, chased) enables the representation of novel events, and is thus a central feature of intelligence. Here, we use functional magnetic resonance (fMRI) and encoding models to delineate the contributions of three brain regions to the representation of relational combinations. We identify a region of anterior-medial prefrontal cortex (amPFC) that shares representations of noun-verb conjunctions across sentences: for example, a combination of "woman" and "chased" to encode woman-as-chaser, distinct from woman-as-chasee. This PFC region differs from the left-mid superior temporal cortex (lmSTC) and hippocampus, two regions previously implicated in representing relations. lmSTC represents broad role combinations that are shared across verbs (e.g., woman-as-agent), rather than narrow roles, limited to specific actions (woman-as-chaser). By contrast, a hippocampal sub-region represents events sharing narrow conjunctions as dissimilar. The success of the hippocampal conjunctive encoding model is anti-correlated with generalization performance in amPFC on a trial-by-trial basis, consistent with a pattern separation mechanism. Thus, these three regions appear to play distinct, but complementary, roles in encoding compositional event structure.

Concepts and Compositionality: In Search of the Brain's Language of Thought.

Imagine Genghis Khan, Aretha Franklin, and the Cleveland Cavaliers performing an opera on Maui. This silly sentence makes a serious point: As humans, we can flexibly generate and comprehend an unbounded number of complex ideas. Little is known, however, about how our brains accomplish this. Here we assemble clues from disparate areas of cognitive neuroscience, integrating recent research on language, memory, episodic simulation, and computational models of high-level cognition. Our review is framed by Fodor's classic language of thought hypothesis, according to which our minds employ an amodal, language-like system for combining and recombining simple concepts to form more complex thoughts. Here, we highlight emerging work on combinatorial processes in the brain and consider this work's relation to the language of thought. We review evidence for distinct, but complementary, contributions of map-like representations in subregions of the default mode network and sentence-like representations of conceptual relations in regions of the temporal and prefrontal cortex.

A long time ago in a galaxy far, far away: How temporal are episodic contents?

A prominent feature of mental event (i.e. 'episodic') simulations is their temporal orientation: human adults can generate episodic representations directed towards the past or the future. Here, we investigated how the temporal orientation of imagined events relates to the contents of these events. Is there something intrinsically temporal about episodic contents? Or does their temporality rely on a distinct set of representations? In three experiments (N = 360), we asked participants to generate and later recall a series of imagined events differing in (1) location, (2) time of day, (3) temporal orientation, and (4) weekday. We then tested to what extent successful recall of episodic content (i.e. (1) and (2)) would predict recall of temporality and/or weekday information. Results showed that recall of temporal orientation was only weakly predicted by recall of episodic contents. Nonetheless, temporal orientation was more strongly predicted by content recall than weekday recall. This finding suggests that episodic simulations are unlikely to be intrinsically temporal in nature. Instead, similar to other forms of temporal information, temporal orientation might be determined from such contents by reconstructive post-retrieval processes. These results have implications for how the human ability to 'mentally travel' in time is cognitively implemented.

Spontaneous giving and calculated greed.

Cooperation is central to human social behaviour. However, choosing to cooperate requires individuals to incur a personal cost to benefit others. Here we explore the cognitive basis of cooperative decision-making in humans using a dual-process framework. We ask whether people are predisposed towards selfishness, behaving cooperatively only through active self-control; or whether they are intuitively cooperative, with reflection and prospective reasoning favouring 'rational' self-interest. To investigate this issue, we perform ten studies using economic games. We find that across a range of experimental designs, subjects who reach their decisions more quickly are more cooperative. Furthermore, forcing subjects to decide quickly increases contributions, whereas instructing them to reflect and forcing them to decide slowly decreases contributions. Finally, an induction that primes subjects to trust their intuitions increases contributions compared with an induction that promotes greater reflection. To explain these results, we propose that cooperation is intuitive because cooperative heuristics are developed in daily life where cooperation is typically advantageous. We then validate predictions generated by this proposed mechanism. Our results provide convergent evidence that intuition supports cooperation in social dilemmas, and that reflection can undermine these cooperative impulses.

An architecture for encoding sentence meaning in left mid-superior temporal cortex.

Human brains flexibly combine the meanings of words to compose structured thoughts. For example, by combining the meanings of "bite," "dog," and "man," we can think about a dog biting a man, or a man biting a dog. Here, in two functional magnetic resonance imaging (fMRI) experiments using multivoxel pattern analysis (MVPA), we identify a region of left mid-superior temporal cortex (lmSTC) that flexibly encodes "who did what to whom" in visually presented sentences. We find that lmSTC represents the current values of abstract semantic variables ("Who did it?" and "To whom was it done?") in distinct subregions. Experiment 1 first identifies a broad region of lmSTC whose activity patterns (i) facilitate decoding of structure-dependent sentence meaning ("Who did what to whom?") and (ii) predict affect-related amygdala responses that depend on this information (e.g., "the baby kicked the grandfather" vs. "the grandfather kicked the baby"). Experiment 2 then identifies distinct, but neighboring, subregions of lmSTC whose activity patterns carry information about the identity of the current "agent" ("Who did it?") and the current "patient" ("To whom was it done?"). These neighboring subregions lie along the upper bank of the superior temporal sulcus and the lateral bank of the superior temporal gyrus, respectively. At a high level, these regions may function like topographically defined data registers, encoding the fluctuating values of abstract semantic variables. This functional architecture, which in key respects resembles that of a classical computer, may play a critical role in enabling humans to flexibly generate complex thoughts.

Integrative moral judgment: dissociating the roles of the amygdala and ventromedial prefrontal cortex.

A decade's research highlights a critical dissociation between automatic and controlled influences on moral judgment, which is subserved by distinct neural structures. Specifically, negative automatic emotional responses to prototypically harmful actions (e.g., pushing someone off of a footbridge) compete with controlled responses favoring the best consequences (e.g., saving five lives instead of one). It is unknown how such competitions are resolved to yield "all things considered" judgments. Here, we examine such integrative moral judgments. Drawing on insights from research on self-interested, value-based decision-making in humans and animals, we test a theory concerning the respective contributions of the amygdala and ventromedial prefrontal cortex (vmPFC) to moral judgment. Participants undergoing fMRI responded to moral dilemmas, separately evaluating options for their utility (Which does the most good?), emotional aversiveness (Which feels worse?), and overall moral acceptability. Behavioral data indicate that emotional aversiveness and utility jointly predict "all things considered" integrative judgments. Amygdala response tracks the emotional aversiveness of harmful utilitarian actions and overall disapproval of such actions. During such integrative moral judgments, the vmPFC is preferentially engaged relative to utilitarian and emotional assessments. Amygdala-vmPFC connectivity varies with the role played by emotional input in the task, being the lowest for pure utilitarian assessments and the highest for pure emotional assessments. These findings, which parallel those of research on self-interested economic decision-making, support the hypothesis that the amygdala provides an affective assessment of the action in question, whereas the vmPFC integrates that signal with a utilitarian assessment of expected outcomes to yield "all things considered" moral judgments.

Response to anticipated reward in the nucleus accumbens predicts behavior in an independent test of honesty.

This study examines the cognitive and neural determinants of honesty and dishonesty. Human subjects undergoing fMRI completed a monetary incentive delay task eliciting responses to anticipated reward in the nucleus accumbens. Subjects next performed an incentivized prediction task, giving them real and repeated opportunities for dishonest gain. Subjects attempted to predict the outcomes of random computerized coin-flips and were financially rewarded for accuracy. In some trials, subjects were rewarded based on self-reported accuracy, allowing them to gain money dishonestly by lying. Dishonest behavior was indexed by improbably high levels of self-reported accuracy. Nucleus accumbens response in the first task, involving only honest rewards, accounted for ∼25% of the variance in dishonest behavior in the prediction task. Individuals showing relatively strong nucleus accumbens responses to anticipated reward also exhibited increased dorsolateral prefrontal activity (bilateral) in response to opportunities for dishonest gain. These results address two hypotheses concerning (dis)honesty. According to the "Will" hypothesis, honesty results from the active deployment of self-control. According to the "Grace" hypothesis, honesty flows more automatically. The present results suggest a reconciliation between these two hypotheses while explaining (dis)honesty in terms of more basic neural mechanisms: relatively weak responses to anticipated rewards make people morally "Graceful," but individuals who respond more strongly may resist temptation by force of Will.

Free will and punishment: a mechanistic view of human nature reduces retribution.

If free-will beliefs support attributions of moral responsibility, then reducing these beliefs should make people less retributive in their attitudes about punishment. Four studies tested this prediction using both measured and manipulated free-will beliefs. Study 1 found that people with weaker free-will beliefs endorsed less retributive, but not consequentialist, attitudes regarding punishment of criminals. Subsequent studies showed that learning about the neural bases of human behavior, through either lab-based manipulations or attendance at an undergraduate neuroscience course, reduced people's support for retributive punishment (Studies 2-4). These results illustrate that exposure to debates about free will and to scientific research on the neural basis of behavior may have consequences for attributions of moral responsibility.

Bayesian multivariate Poisson abundance models for T-cell receptor data.

A major feature of an adaptive immune system is its ability to generate B- and T-cell clones capable of recognizing and neutralizing specific antigens. These clones recognize antigens with the help of the surface molecules, called antigen receptors, acquired individually during the clonal development process. In order to ensure a response to a broad range of antigens, the number of different receptor molecules is extremely large, resulting in a huge clonal diversity of both B- and T-cell receptor populations and making their experimental comparisons statistically challenging. To facilitate such comparisons, we propose a flexible parametric model of multivariate count data and illustrate its use in a simultaneous analysis of multiple antigen receptor populations derived from mammalian T-cells. The model relies on a representation of the observed receptor counts as a multivariate Poisson abundance mixture (m PAM). A Bayesian parameter fitting procedure is proposed, based on the complete posterior likelihood, rather than the conditional one used typically in similar settings. The new procedure is shown to be considerably more efficient than its conditional counterpart (as measured by the Fisher information) in the regions of m PAM parameter space relevant to model T-cell data.

Boosting the impact of charitable giving with donation bundling and micromatching.

The most effective charities are hundreds of times more impactful than typical charities. However, most donors favor charities with personal/emotional appeal over effectiveness. We gave donors the option to split their donations between their personal favorite charity and an expert-recommended highly effective charity. This bundling technique increased donors' impact without undermining their altruistic motivation, boosting effective donations by 76%. An additional boost of 55% was achieved by offering matching donations with increasing rates for allocating more to the highly effective charity. We show further that matching funds can be provided by donors focused on effectiveness through a self-sustaining process of micromatching. We applied these techniques in a new online donation platform (GivingMultiplier.org), which fundraised more than $1.5 million in its first 14 months. While prior applied research on altruism has focused on the quantity of giving, the present results demonstrate the value of focusing on the effectiveness of altruistic behavior.

You see, the ends don't justify the means: visual imagery and moral judgment.

We conducted three experiments indicating that characteristically deontological judgments--here, disapproving of sacrificing one person for the greater good of others--are preferentially supported by visual imagery. Experiment 1 used two matched working memory tasks-one visual, one verbal-to identify individuals with relatively visual cognitive styles and individuals with relatively verbal cognitive styles. Individuals with more visual cognitive styles made more deontological judgments. Experiment 2 showed that visual interference, relative to verbal interference and no interference, decreases deontological judgment. Experiment 3 indicated that these effects are due to people's tendency to visualize the harmful means (sacrificing one person) more than the beneficial end (saving others). These results suggest a specific role for visual imagery in moral judgment: When people consider sacrificing someone as a means to an end, visual imagery preferentially supports the judgment that the ends do not justify the means. These results suggest an integration of the dual-process theory of moral judgment with construal-level theory.

Patterns of neural activity associated with honest and dishonest moral decisions.

What makes people behave honestly when confronted with opportunities for dishonest gain? Research on the interplay between controlled and automatic processes in decision making suggests 2 hypotheses: According to the "Will" hypothesis, honesty results from the active resistance of temptation, comparable to the controlled cognitive processes that enable the delay of reward. According to the "Grace" hypothesis, honesty results from the absence of temptation, consistent with research emphasizing the determination of behavior by the presence or absence of automatic processes. To test these hypotheses, we examined neural activity in individuals confronted with opportunities for dishonest gain. Subjects undergoing functional magnetic resonance imaging (fMRI) gained money by accurately predicting the outcomes of computerized coin-flips. In some trials, subjects recorded their predictions in advance. In other trials, subjects were rewarded based on self-reported accuracy, allowing them to gain money dishonestly by lying about the accuracy of their predictions. Many subjects behaved dishonestly, as indicated by improbable levels of "accuracy." Our findings support the Grace hypothesis. Individuals who behaved honestly exhibited no additional control-related activity (or other kind of activity) when choosing to behave honestly, as compared with a control condition in which there was no opportunity for dishonest gain. In contrast, individuals who behaved dishonestly exhibited increased activity in control-related regions of prefrontal cortex, both when choosing to behave dishonestly and on occasions when they refrained from dishonesty. Levels of activity in these regions correlated with the frequency of dishonesty in individuals.