The appreciation of humor is a universal phenomenon and a key aspect of cognition. It has been studied in thecontext of jokes, where the incongruity in expected and observed context results in the perception of humor. The present studyexamines how the humor appreciation of single words relates to the humor of the whole joke – is a joke simply a sum of itsparts? Using a novel dataset of single-word humor ratings, collections of jokes from the JESTER database were analyzed. Amultiple regression analysis showed joke length and individual word arousal were the best predictors of joke funniness. Longerjokes with fewer individually arousing words were found funnier. Individual word humor did not contribute to the humor ofthe overall joke. These findings suggest the cognitive aspects of humor are likely driven by broader semantic context, whereasappreciating humor on a per-word basis links to separate factors.

Immigrants always have difficulties in integrating into a local society, and sometimes these difficulties come fromthe fact that they are subconsciously regarded as someone from the outside. How they are represented in the main stream mediacould reflect people’s perceptions towards the ‘otherness’ of different social members. We analyzed the representation of 29US immigrant groups in newspaper articles in 2 related studies. The favorability of an immigrant groups is highly associatedwith its perceived social distance (reflected through usage of concrete language) in our research. To further understand whatcaused the positive or negative image of immigrants, we applied Latent Dirichlet Allocation to identify topics associated withimmigrant groups. We also investigated into how these news topics differ in terms of lingual social distance and favorability.The results provide both qualitative and quantitative insights in how the image of immigrants are reproduced in social media.

Feature distinctiveness is a measure representing the uniqueness of objects’ features. Previous research found linksbetween noun feature distinctiveness and age of acquisition (i.e. nouns referring to objects with relatively unique featuresare learned earlier). The present work investigates the links between feature distinctiveness and age of acquisition in verbs.Using high-dimensional vector space modelling, noun and verb feature distinctiveness was represented as Manhattan distancebetween word nodes. Both nouns and verbs showed negative correlations between feature distinctiveness and age of acquisition(words of more distinctive objects learned earlier), suggesting a general distinctiveness bias. This effect was stronger for nouns.An investigation of child directed speech (CDS) from the CHILDES corpus showed a correlation between child directed wordfrequency and feature distinctiveness for nouns (featurally distinctive nouns are more common in CDS), but not for verbs. Thepossible link between distinctiveness in CDS and age of acquisition effects is discussed.

We search for various things every day – food, information onthe Internet or someone’s name in memory. Despite the dif-ferent nature of these tasks, they all have a common feature –a final goal with an unknown location in a complex environ-ment. This property of the search raises a problem of trade-offbetween exploration of new opportunities and exploitation ofthe known information. We used the data from the semanticfluency task experiment to investigate how humans switch be-tween exploration and exploitation strategies when they searchin memory and whether they do it optimally. On comparingfour different search models, the one that assumes that humansswitch search strategies according to the semantic quality ofthe current neighbourhood best fits the data. Moreover, par-ticipants who set higher thresholds for the words with betterquality of the neighbourhood tend to retrieve more words frommemory. We also used regression analysis to find out whichfactors affect efficiency of both search strategies.

The focus of this paper is to examine differences in semantic network structure of late talkers and typical talkers to elucidate potential learning strategies used by late talking children. To address this question, we conducted network analysis on the vocabularies of 2,912 children, with 566 of those being late talkers. Contrary to previously reported findings, the results show that late talkers have well-connected vocabularies as measured by median degree, clustering coefficient, and mean distance, with more well-connected networks in some cases than their typical talking peers. Further analysis of word order suggests that late talkers may be selecting based on frequency and connectivity of the words in the learning environment, more so than typical talkers. The language processing difficulties in late talkers appear not to be associated with their semantic network properties. In sum, late talkers may initially benefit from using word frequency and word connectivity strategies to build well-connected vocabularies.