A database of orthography-semantics consistency (OSC) estimates for 15,017 English words

Abstract

Orthography-semantics consistency (OSC) is a measure that quantifies the degree of semantic relatedness between a word and its orthographic relatives. OSC is computed as the frequency-weighted average semantic similarity between the meaning of a given word and the meanings of all the words containing that very same orthographic string, as captured by distributional semantic models. We present a resource including optimized estimates of OSC for 15,017 English words. In a series of analyses, we provide a progressive optimization of the OSC variable. We show that computing OSC from word-embeddings models (in place of traditional count models), limiting preprocessing of the corpus used for inducing semantic vectors (in particular, avoiding part-of-speech tagging and lemmatization), and relying on a wider pool of orthographic relatives provide better performance for the measure in a lexical-processing task. We further show that OSC is an important and significant predictor of reaction times in visual word recognition and word naming, one that correlates only weakly with other psycholinguistic variables (e.g., family size, word frequency), indicating that it captures a novel source of variance in lexical access. Finally, some theoretical and methodological implications are discussed of adopting OSC as one of the predictors of reaction times in studies of visual word recognition.

Morphological processing without semantics: An ERP study with spoken words

“…The interaction between embedded word frequency and LSA semantic relatedness was not significant, suggesting that embedded words were activated independently of whether or not they shared a semantic relationship with the whole word. The main effect of embedded word frequency is consistent with a growing body of evidence (e.g., Beyersmann et al, 2015;Beyersmann, Cavalli, Casalis, & Colé, 2016;Marelli & Amenta, 2018;Taft, Li, & Beyersmann, 2018), suggesting that embedded words are always activated independently of whether they are accompanied by an affix (as in farm + er or corn + er) or a non-affix (as in cash + ew). The activation of embedded words (e.g., cash) then generates lexical competition with the whole word (e.g., cashew), thus leading to an increase in response times across…”

Consistency measures individuate dissociating semantic modulations in priming paradigms: A new look on semantics in the processing of (complex) words

“…OSC was computed as the frequency-weighted average of the cosine proximity (based on the semantic space just described) between the target vector and each of its orthographic relatives. These latter were defined as any words embedding the target (differently from the original 2015 paper, we did not include an onset-specific positional constraint in the selection of relatives, as we observed that it leads to a worse measure performance; see Marelli & Amenta, 2018).…”

“…and regardless of the relationship between the relative and the target (i.e., relatives are not necessarily morphologically or semantically related to the target, as seen in the examples above). For example, the string widow is contained in widower, widowed , and widowhood; therefore, all these words will be considered orthographic relatives of “widow” (for more details on how the orthographic relatives are defined and validation of this procedure, see Marelli & Amenta, 2018). Because all these words are associated to the meaning of WIDOW, OSC will be high (i.e., the semantic similarity between the target— widow in this case—and all its neighbours is high).…”

“…Formally, OSC is the frequency-weighted 2 semantic similarity between the semantic representation of a word (generated using methods from distributional semantics, similar to latent semantic analysis (LSA); see below for details) and the semantic representations of its orthographic relatives (see Marelli & Amenta, 2018; Marelli et al, 2015 for details on this method):…”

“…Words that present higher degrees of consistency between their meanings and those of their orthographic relatives (i.e., high OSC) are easier to recognise, whereas words with less consistency between their meaning and those of their orthographic relatives (i.e., low OSC values) are harder to recognise. It does seem, then, that the visual identification system is sensitive to how consistent the mapping is between orthographic form and meaning (Marelli & Amenta, 2018; Marelli et al, 2015).…”