A Revised Assessment of Early Modal Reasoning

This research builds upon a previous method from Jones and Call (2024). They gave chimpanzees 2 trial types. In "Inferred 1 grape vs seen half grape” trials (denoted as 1-cup in the paper), chimpanzees chose between a guaranteed full grape and a guaranteed half grape, representing expected values of 1.0 and 0.5 grapes, respectively. The ratio of expected rewards is 2:1. Chimpanzees demonstrated a near-universal preference for the larger reward. In “Unknown 1 grape vs seen half grape” trials (2-cup), chimpanzees chose between a full grape hidden in one of two containers (location unknown) and a visible half grape. Both options offered an equivalent expected value of 0.5 grapes, with a ratio of expected rewards of 1:1. Chimpanzees chose the half grape option more often in “Unknown 1 grape vs seen half grape” trials than in "Inferred 1 grape vs seen half grape” trials. One interpretation of these results is that chimpanzees are calculating the EV ratios of these events and making decisions to maximize their total return. This requires assessing the probability of receiving a reward based on the number of possible hiding locations available in each condition. Alternatively, these results could arise because the chimpanzees were seeking known rewards rather than calculating expected value. In the "Inferred 1 grape vs seen half grape”, the rewards on both sides were known; so, they chose the one with the higher reward. In the “Unknown 1 grape vs seen half grape” trials, however, only the half-grape was a known reward while the location of the full grape remained unknown. Under this interpretation, the chimpanzees preferred the certainty of a visible reward over an unknown reward. Our research shifts the population focus to 4-year-old children. We first determine whether the original findings can be reproduced, and if so, we test whether children are calculating EV or simply seeking known rewards. In this adaptation, we switch the rewards to a bag of gold coins, allowing us to manipulate the expected reward of various decisions by varying the number of coins in the bag. To test for replication, two conditions are implemented: “Inferred 2 coins vs seen 1 coin” (replicating "Inferred 1 grape vs seen half grape”) and “Unknown 2 coins vs seen 1 coin” (replicating “Unknown 1 grape vs seen half grape”). Successful reproduction of the original results would be demonstrated if children choose the visible 1 coin more often in the latter trial type than the former. Finally, we add a third condition: “Unknown 4 coins vs seen 1 coin”. In these trials, the EV ratio is 2:1, mirroring the expected value of the first condition; however, the high-EV option now has 50% probability, matching the uncertainty of “Unknown 2 coins vs seen 1 coin” trials. This condition serves as a critical test between the two hypotheses: if children are calculating expected value, they should prefer the 50% chance of getting 4 coins. But, if they are seeking known reward, they should choose the known low-EV reward more often in “Unknown 4 coins vs seen 1 coin” condition than the “Inferred 2 coins vs seen 1 coin” condition, despite the identical expected values.