Distributional Measures of Semantic Abstraction.

This article provides an in-depth study of distributional measures for distinguishing between degrees of semantic abstraction. Abstraction is considered a "central construct in cognitive science" (Barsalou, 2003) and a "process of information reduction that allows for efficient storage and retrieval of central knowledge" (Burgoon et al., 2013). Relying on the distributional hypothesis, computational studies have successfully exploited measures of contextual co-occurrence and neighbourhood density to distinguish between conceptual semantic categorisations. So far, these studies have modeled semantic abstraction across lexical-semantic tasks such as ambiguity; diachronic meaning changes; abstractness vs. concreteness; and hypernymy. Yet, the distributional approaches target different conceptual types of semantic relatedness, and as to our knowledge not much attention has been paid to apply, compare or analyse the computational abstraction measures across conceptual tasks. The current article suggests a novel perspective that exploits variants of distributional measures to investigate semantic abstraction in English in terms of the abstract-concrete dichotomy (e.g., glory-banana) and in terms of the generality-specificity distinction (e.g., animal-fish), in order to compare the strengths and weaknesses of the measures regarding categorisations of abstraction, and to determine and investigate conceptual differences. In a series of experiments we identify reliable distributional measures for both instantiations of lexical-semantic abstraction and reach a precision higher than 0.7, but the measures clearly differ for the abstract-concrete vs. abstract-specific distinctions and for nouns vs. verbs. Overall, we identify two groups of measures, (i) frequency and word entropy when distinguishing between more and less abstract words in terms of the generality-specificity distinction, and (ii) neighbourhood density variants (especially target-context diversity) when distinguishing between more and less abstract words in terms of the abstract-concrete dichotomy. We conclude that more general words are used more often and are less surprising than more specific words, and that abstract words establish themselves empirically in semantically more diverse contexts than concrete words. Finally, our experiments once more point out that distributional models of conceptual categorisations need to take word classes and ambiguity into account: results for nouns vs. verbs differ in many respects, and ambiguity hinders fine-tuning empirical observations.

Association norms for German noun compounds and their constituents.

We present a collection of association norms for 246 German depictable compound nouns and their constituents, comprising 58,652 association tokens distributed over 26,004 stimulus-associate pair types. Analyses of the data revealed that participants mainly provided noun associates, followed by adjective and verb associates. In corpus analyses, co-occurrence values for compounds and their associates were below those for nouns in general and their associates. The semantic relations between compound stimuli and their associates were more often co-hyponymy and hypernymy and less often hyponymy than for associations to nouns in general. Finally, we found a moderate correlation between the overlap of the associations to compounds and their constituents and the degree of semantic transparency. These data represent a collection of associations to German compound nouns and their constituents that constitute a valuable resource concerning the lexical semantic properties of the compound stimuli and the semantic relations between the stimuli and their associates. More specifically, the norms can be used for stimulus selection, hypothesis testing, and further research on morphologically complex words. The norms are available in text format (utf-8 encoding) as supplemental materials.

Distinct processing of function verb categories in the human brain.

A subset of German function verbs can be used either in a full, concrete, 'heavy' ("take a computer") or in a more metaphorical, abstract or 'light' meaning ("take a shower", no actual 'taking' involved). The present magnetoencephalographic (MEG) study explored whether this subset of 'light' verbs is represented in distinct cortical processes. A random sequence of German 'heavy', 'light', and pseudo verbs was visually presented in three runs to 22 native German speakers, who performed lexical decision task on real versus pseudo verbs. Across runs, verbs were presented (a) in isolation, (b) in minimal context of a personal pronoun, and (c) 'light' verbs only in a disambiguating context sentence. Central posterior activity 95-135 ms after stimulus onset was more pronounced for 'heavy' than for 'light' uses, whether presented in isolation or in minimal context. Minimal context produced a similar heavy>light differentiation in the left visual word form area at 160-200 ms. 'Light' verbs presented in sentence context allowing only for a 'heavy reading' evoked larger left-temporal activation around 270-340 ms than the corresponding 'light reading'. Across runs, real verbs provoked more pronounced activation than pseudo verbs in left-occipital regions at 110-150 ms. Thus, 'heavy' versus 'light readings' of verbs already modulate early posterior visual evoked response even when verbs are presented in isolation. This response becomes clearer in the disambiguating contextual condition. This type of study shows for the first time that language processing is sensitive to representational differences between two readings of one and the same verb stem.

Generation of hydroxyl radicals mediated by streptozotocin in pancreatic islets of mice in vitro.

Type I diabetes is considered a multifactorial autoimmune process initiated by an environmental factor. There is evidence that reactive oxygen species are involved in destructing insulin-producing beta-cells. In mice, reactive oxygen species and nitric monoxide contribute to beta-cell damage in the non-obese diabetic strain developing spontaneously diabetes and in diabetes induced with multiple low doses of streptozotocin. Previously, we found that zinc sulfate induced metallothionein in pancreatic islets, protected beta-cells against streptozotocin toxicity in vitro, and prevented diabetes induced with multiple low doses of streptozotocin. Since metallothionein is known to scavenge hydroxyl radicals in cell-free systems, we hypothesize that the protective effect of zinc sulfate results from metallothionein induction scavenging hydroxyl radicals generated by multiple low doses of streptozotocin. Therefore, we studied whether levels of hydroxyl radicals are increased by streptozotocin in isolated islets in vitro. Here, we demonstrate basal and streptozotocin-stimulated hydroxyl radicals by electron spin resonance spectroscopy in combination with hydroxyl radical-specific spin trapping in islet homogenates. Furthermore, in islet cultures, streptozotocin augmented generation of reactive oxygen species as determined by fluorescence. Of the group of reactive oxygen species, the streptozotocin-augmented generation of hydrogen peroxide was also specifically determined. We conclude that streptozotocin-mediated hydroxyl radicals and generation of reactive oxygen species may be crucial effectors in beta-cell damage.