Testing a computational model of causative overgeneralizations: Child judgment and production data from English, Hebrew, Hindi, Japanese and K’iche’

How do language learners avoid the production of verb argument structure overgeneralization errors (*The clown laughed the man c.f. The clown made the man laugh), while retaining the ability to apply such generalizations productively when appropriate? This question has long been seen as one that is both particularly central to acquisition research and particularly challenging. Focussing on causative overgeneralization errors of this type, a previous study reported a computational model that learns, on the basis of corpus data and human-derived verb-semantic-feature ratings, to predict adults’ by-verb preferences for less- versus more-transparent causative forms (e.g., * The clown laughed the man vs The clown made the man laugh) across English, Hebrew, Hindi, Japanese and K’iche Mayan. Here, we tested the ability of this model (and an expanded version with multiple hidden layers) to explain binary grammaticality judgment data from children aged 4;0-5;0, and elicited-production data from children aged 4;0-5;0 and 5;6-6;6 (N=48 per language). In general, the model successfully simulated both children’s judgment and production data, with correlations of r=0.5-0.6 and r=0.75-0.85, respectively, and also generalized to unseen verbs. Importantly, learners of all five languages showed some evidence of making the types of overgeneralization errors – in both judgments and production – previously observed in naturalistic studies of English (e.g., *I’m dancing it). Together with previous findings, the present study demonstrates that a simple learning model can explain (a) adults’ continuous judgment data, (b) children’s binary judgment data and (c) children’s production data (with no training of these datasets), and therefore constitutes a plausible mechanistic account of the acquisition of verbs’ argument structure restrictions.

Not only size matters: limits to the Law of Three Consonants in French phonology

Grammont’s (1914) influential Law of Three Consonants (LTC) states that French schwa is obligatorily pronounced in any CC_C sequence to avoid three-consonant clusters. Later works have shown that schwa presence is also sensitive to the nature of the consonants involved, at least at the word and phrase levels. However the LTC is still generally considered as accurate under its original formulation to describe schwa-zero alternations at the stem level. The goal of the paper is to test whether the LTC should be relaxed even in this context. The paper presents two studies using judgment data to compare the behavior of schwa in derived words (stem-level phonology) and in inflected words (word-level phonology). The results of the two studies show that the nature of consonants involved in the CC_C sequence plays a role at both stem and word levels. Moreover, the same phonotactic asymmetries among consonant clusters are found in both contexts. The data therefore support a weaker version of the stem-level vs. word-level divide than what is usually assumed for French. This conclusion is strengthened by the results of a modeling study showing that a constraint-based grammar with the same phonotactic constraints across stem- and word-level phonologies provides a better fit to the judgment data from Study 1 and Study 2 than a grammar with different phononotactic constraints in the two morphosyntactic domains. The paper also replicates a number of earlier findings on the role of morphosyntactic domains, clash avoidance, and dialectal variation in schwa-zero alternations.

Analyzing variability in L2 ultimate attainment

Ultimate attainment is typically more heterogeneous among second-language (L2) learners than among native speakers (e.g. Bley-Vroman, 1990). The present study offers a suite of simple analytical procedures aimed at exploring types and loci of variability in L2 attainment vis-à-vis those in the corresponding first language (L1), with special attention to certain learner-external factors that might condition such variabilities. To demonstrate the methods and their potential, we apply these procedures to learner and native acceptability judgment data published in Birdsong (1992). Under means analyses of item ratings and coefficients of variation (CV), a group of adult Anglophone learners of L2 French (ENS) are found to resemble native French controls (FNS). In contrast, under correlational analyses of ratings and CVs, ENS resemble FNS on grammatical items, but diverge on ungrammatical items. Correlational and means analyses of both CV and acceptability ratings reveal that ENS-FNS convergence is not predictable from the degree of FNS ratings variability, contra DeKeyser (2012). For both groups, we observe an interaction between FNS ratings variability and the grammatical status of items (ungrammatical vs. grammatical). Finally, for neither group do we find a relationship between the order of item presentation and ratings severity or CVs. We present our perspectives as a road map for future analyses of variabilities inherent in language learning outcomes.

Against the Law of Three Consonants in French: Evidence from Judgment Data

Grammont's Law of Three Consonants (LTC) states that French schwa is obligatorily pronounced in any CC_C sequence to avoid three-consonant clusters. Although schwa presence has been shown to be sensitive not only to cluster size but also to the nature of consonants in post-lexical phonology, the LTC is still considered as accurate to describe schwa-zero alternations in lexical phonology. The paper uses judgment data from French speakers in France and Switzerland to compare the behavior of schwa in derived words (lexical phonology) and inflected words (post-lexical phonology). The results show that schwa-zero alternations are conditioned not only by cluster size but also by cluster type in lexical phonology. Moreover, the same phonotactic asymmetries among consonant clusters are found in lexical and post-lexical phonologies. The data therefore support a weaker version of the lexical-phonology hypothesis than what is usually assumed for French. Lexical and post-lexical phonologies do not require different phonotactic constraints but only different weights for the same constraints.

Big-M-Small-N Temporal-Order Judgment Data

We present a large and precise data set of temporal-order judgments on visual stimuli.Stimulus asynchronies ranged from 0 to 80 ms in steps of 6.67 ms. The data setincludes an attention manipulation driven by one target’s orientation compared tobackground elements (either zero or 90 degrees). Each of 25 stimulus asynchronies wassampled with at least 196 repetitions (and more than 400 times in two participants).Furthermore, fixation, an important concern in studies on covert attention, wasmonitored. Precise data are helpful for answering theoretical questions in psychology.For some questions such as model comparisons, they may even be necessary. Threedifferent exemplary models are fitted to the data.

Testing a computational model of causative overgeneralizations: Child judgment and production data from English, Hebrew, Hindi, Japanese and K’iche’

How do language learners avoid the production of verb argument structure overgeneralization errors (*The clown laughed the man c.f. The clown made the man laugh), while retaining the ability to apply such generalizations productively when appropriate? This question has long been seen as one that is both particularly central to acquisition research and particularly challenging. Focussing on causative overgeneralization errors of this type, a previous study reported a computational model that learns, on the basis of corpus data and human-derived verb-semantic-feature ratings, to predict adults’ by-verb preferences for less- versus more-transparent causative forms (e.g., *The clown laughed the man vs The clown made the man laugh) across English, Hebrew, Hindi, Japanese and K’iche Mayan. Here, we tested the ability of this model to explain binary grammaticality judgment data from children aged 4;0-5;0, and elicited-production data from children aged 4;0-5;0 and 5;6-6;6 (N=48 per language). In general, the model successfully simulated both children’s judgment and production data, with correlations of r=0.5-0.6 and r=0.75-0.85, respectively, and also generalized to unseen verbs. Importantly, learners of all five languages showed some evidence of making the types of overgeneralization errors – in both judgments and production – previously observed in naturalistic studies of English (e.g., *I’m dancing it). Together with previous findings, the present study demonstrates that a simple discriminative learning model can explain (a) adults’ continuous judgment data, (b) children’s binary judgment data and (c) children’s production data (with no training of these datasets), and therefore constitutes a plausible mechanistic account of the retreat from overgeneralization.

A note on the equality judgment experiment in Itthipuripat, et al. (2019)

Itthipuripat, Chang, Bong, & Serences (2019) investigated the effects of attention on appearance and decision criteria. Using a series of comparative and equality judgment tasks, they determined that most of the effects of attention could be explained as changes in decision criteria; however, their equality judgment experiment did reveal an effect of attention on appearance at low contrasts. Because their analysis reported Cmax, the contrast at which the equality judgement is maximum, and because Cmax is not independent of potential low-contrast response biases, I decided to reanalyze their equality judgment data. Based on this reanalysis, I was not able to rule out a potential effect of attention on appearance at low contrast. But, due to the small number and asymmetric range of contrasts probed, their experiment was not well designed to test this contrast regime, and I am not confident about their result. For the higher contrasts they tested, the measured effects of attention trended in the opposite direction.

Good for what? Power, character, and moral judgment.

The person-centered account of moral judgment suggests an evolutionary explanation for two puzzling findings from moral psychology: that people judge an act that had bad side effects as more intentional than an act that had good side effects, and that people judge an accident as being more under the control of the target who caused it if that target has good values. In both cases, people’s intuitions are biased by their “fundamental need” to judge a person’s character. People have evolved to judge people not just by specific actions, but by considering whether they would be good cooperative partners and members of a social group. We tested implications of this explanation: people alter their judgments of intention and causal control to benefit people with greater value to the group. In two studies, we examined the moderating role of social value on moral judgment using highly cited and often replicated experimental paradigms. Results do not corroborate this influential theory of moral judgment. High status individuals are not given “moral license.” Instead, results suggest that dominant, but not prestigious, individuals are judged as more efficacious, which may have downstream consequences for moral judgment. Data and analysis scripts available at: https://osf.io/u9pfr/.

There Is No Primacy Effect in Interpersonal Perception: A Series of Preregistered Analyses Using Judgments of Actual Behavior

First impressions are commonly assumed to be particularly important: Information about a person that we obtain early on may shape our overall impression of that person more strongly than information obtained later. In contrast to previous research, the present series of preregistered analyses uses actual person judgment data to investigate this so-called primacy effect: Perceivers ( N = 1,395) judged the videotaped behavior of target persons ( N = 200) in 10 different situations. Separate subsamples of about 200 perceivers each were used in moving from exploratory to increasingly confirmatory analyses. Contrary to our expectations, no primacy effect was found. Instead, judgments of the targets in later situations were more strongly associated with overall impressions, indicating an acquaintance effect. Relying on early information seems unreasonable when more comprehensive information is readily available. Early information may, however, affect perceivers’ behavioral reactions to the targets and thus their future interactions, if such interactions are possible.