It’s all in the delivery: Effects of context valence, arousal, and concreteness on visual word processing

Abstract

Prior research has examined how distributional properties of contexts (number of unique contexts or their informativeness) influence the effort of word recognition. These properties do not directly interrogate the semantic properties of contexts. We evaluated the influence of average concreteness, valence (positivity) and arousal of the contexts in which a word occurs on response times in the lexical decision task, age of acquisition of the word, and word recognition memory performance. Using large corpora and norming mega-studies we quantified semantics of contexts for thousands of words and demonstrated that contextual factors were predictive of lexical representation and processing above and beyond the influence shown by concreteness, valence and arousal of the word itself. Our findings indicate that lexical representations are influenced not only by how diverse the word's contexts are, but also by the embodied experiences they elicit.

Immediate sensorimotor grounding of novel concepts learned from language alone

“…This role of language as a means to establish indirect grounding of (novel) concepts has also been discussed with respect to the notion of semantic prosody, describing the phenomenon that specific words tend to occur in connotated contexts (see Winter, 2019) (for example, utterly is usually used in negative contexts). Moving from this, Snefjella and Kuperman (2016) found influences of the concreteness, valence, and arousal of their typical context words on the representation and processing of words, thus identifying the propagation of semantic properties through context as a general tendency in natural language. In a more recent study, Snefjella et al (in press) further provide direct experimental evidence that originally meaningless novel words acquire emotional connotations from their learning context.…”

“…However, a possible solution is already offered in the work that sparked the grounded cognition debate in the cognitive sciences (Harnad, 1990): Zebra can be easily understood by our medieval European if she knows the meaning of horse and stripe, and is told by the traveller that a zebra is a horse with stripes. In terms of the experiential trace model, zebra would be grounded indirectly, through linguistic co-occurrence with already-grounded concepts -in this case, through being described in terms of already-grounded concepts and therefore incorporating these experiential traces (compare Snefjella, Lana, & Kuperman, in press; see also Snefjella & Kuperman, 2016 for earlier versions of the same argument). Based on this assumption, Günther, Dudschig, and Kaup (2018) investigated whether novel words that were learned from language alone can lead to sensorimotor activation during language processing.…”

Semantic Norm Extrapolation is a Missing Data Problem

“…Indeed, it is routine that when re-analyzing megastudy data, only a subset of items in the megastudy will have complete measurements (cf. large proportions of megastudy items discarded in Brysbaert and Biemiller, 2017;Kuperman, Estes, Brysbaert, and Warriner, 2014;Mandera et al, 2019;Pexman et al, 2019;Snefjella and Kuperman, 2016). Missing semantic norms lead to only portions of megastudy items being included in the analysis, causing at minimum a reduction in statistical power and potentially returning researchers to the problematic, small-sample analyses that have damaged scientific credibility (Ioannidis, 2005).…”

Sensory Linguistics

“…Evidence for the fact that emotional meaning is encoded in lexical representations comes from the fact that a word's emotional valence affects its processing speed in reading and naming tasks (Kousta, Vinson, & Vigliocco, 2009;Kuperman, 2015;Kuperman, Estes, Brysbaert, & Warriner, 2014;Snefjella & Kuperman, 2016). For example, positive words are processed more quickly than negative words, and strongly positive and strongly negative words are overall processed more quickly than neutral words (Kuperman, 2015).…”

“…There are by now dozens of datasets that quantify the extent to which a word is positive or negative (Baccianella, Esuli, & Sebastiani, 2010;Esuli & Sebastiani, 2006;Kiritchenko, Zhu, & Mohammad, 2014;Liu, 2012, Chapter 6;Mohammad & Kiritchenko, 2015;Pang & Lee, 2008, Chapter 7;Warriner et al, 2013). However, only very few researchers have investigated semantic prosody with the help of such datasets (Dilts & Newman, 2006;Snefjella & Kuperman, 2016;Winter, 2016). Doing so advances the reproducibility of semantic prosody research because given the same corpus and the same valence dataset, different researchers will come to the same conclusions.…”