This study examines a collection of expressions for the taboo topic of menstruation in Dutch, German, and Mandarin Chinese. A model for the identification of conceptualization patterns in taboo verbalizations is set up, analyzing each expression according to the X-phemistic mechanisms and, if applicable, the metaphorical source domains or metonymic vehicles at its origin. The various conceptualizations of menstruation are approached from a socio-cultural perspective; variation in conceptualization is examined through a correspondence regression analysis with three speaker-related explanatory variables (L1 and associated cultural background, menstrual experience, and age group). The underlying interest is linguo-cultural as the study aims to verify whether dominant menstrual attitudes are reflected in the linguistic conceptualization of menstruation within each socio-cultural group. Such correlations are indeed found, although the youngest age-group shows some unexpected linguistic behavior.

Extensive research has shown that children’s early words are learned through sensorimotor experience. Thus, early-acquired words tend to have more concrete meanings. Abstract word meanings tend to be learned later but less is known about their acquisition. We collected meaning-specific concreteness ratings and examined their relationship with age-of-acquisition data from large-scale vocabulary testing with children in grade 2 to college age. Earlier-acquired meanings were rated as more concrete while later-acquired meanings as more abstract, particularly for words typically considered to be concrete. The results suggest that sensorimotor experiences are important to early-acquired word meanings, and other experiences (e.g., linguistic) are important to later-acquired meanings, consistent with a multi-representational view of lexical semantics.

We examine the binary classification of sentiment views for verbal multiword expressions (MWEs). Sentiment views denote the perspective of the holder of some opinion. We distinguish between MWEs conveying the view of the speaker of the utterance (e.g., in “The company reinvented the wheel” the holder is the implicit speaker who criticizes the company for creating something already existing) and MWEs conveying the view of explicit entities participating in an opinion event (e.g., in “Peter threw in the towel” the holder is Peter having given up something). The task has so far been examined on unigram opinion words. Since many features found effective for unigrams are not usable for MWEs, we propose novel ones taking into account the internal structure of MWEs, a unigram sentiment-view lexicon and various information from Wiktionary. We also examine distributional methods and show that the corpus on which a representation is induced has a notable impact on the classification. We perform an extrinsic evaluation in the task of opinion holder extraction and show that the learnt knowledge also improves a state-of-the-art classifier trained on BERT. Sentiment-view classification is typically framed as a task in which only little labeled training data are available. As in the case of unigrams, we show that for MWEs a feature-based approach beats state-of-the-art generic methods.

While relational nouns (cousin) are traditionally delineated in a binary and theory-dependent manner, this article approximates relationality as a continuous, objective corpus metric (Percent Possessive) – allowing for lexicon-wide exploration of which nouns are more or less relational and why. Comparing across nouns and accounting for the ontological class of the noun's referent (focusing on nouns denoting artifacts, natural kinds, occupations, humans and locations), I find that Percent Possessive is positively correlated with a noun's per-million-word frequency. Comparing across different web communities, I find that a noun is more frequent, and shows a greater ratio of definite to indefinite tokens, in the community where its Percent Possessive is significantly higher. I take these findings to be consistent with the claim that a noun is more easily interpreted as relational (as measured by Percent Possessive) when human interaction with its referent is more conventional (as measured by its frequency and definite-to-indefinite ratio). Inspired by the many authors who have suggested a socio-cultural component to relationality and possession, this article explores at scale in English how nouns reflect the conventions of the people who use them.

This paper aims to fill in a long missing piece in the paradigmatic word-formation research: a set of rival affixes whose members are differentiated in meaning. We argue that such a set can be found in English derivational adjectivalization, in the affixal rivalry between the adjectivalizing suffixes -ed and -y. Using the traditional method of doublet comparison (Aronoff 1976, 2020), we reveal that adjectives of the form Xed and those of the form Xy (X standing for the source word) differ in the scale type. Xed adjectives are closed-scale adjectives, but Xy adjectives are totally open-scale adjectives. The scale-type difference explains why Xed adjectives combine with certain degree modifiers, whereas Xy adjectives do not. Furthermore, we show that the rival affixes are doubly differentiated in the deverbal domain in terms of the said output scale type and the input base selection. In this domain, the major sources of the closed-scale -ed adjectives and the open-scale -y adjectives are result and manner verbs, respectively.

Modelling semantic similarity plays a fundamental role in lexical semantic applications. A natural way of calculating semantic similarity is to access handcrafted semantic networks, but similarity prediction can also be anticipated in a distributional vector space. Similarity calculation continues to be a challenging task, even with the latest breakthroughs in deep neural language models. We first examined popular methodologies in measuring taxonomic similarity, including edge-counting that solely employs semantic relations in a taxonomy, as well as the complex methods that estimate concept specificity. We further extrapolated three weighting factors in modelling taxonomic similarity. To study the distinct mechanisms between taxonomic and distributional similarity measures, we ran head-to-head comparisons of each measure with human similarity judgements from the perspectives of word frequency, polysemy degree and similarity intensity. Our findings suggest that without fine-tuning the uniform distance, taxonomic similarity measures can depend on the shortest path length as a prime factor to predict semantic similarity; in contrast to distributional semantics, edge-counting is free from sense distribution bias in use and can measure word similarity both literally and metaphorically; the synergy of retrofitting neural embeddings with concept relations in similarity prediction may indicate a new trend to leverage knowledge bases on transfer learning. It appears that a large gap still exists on computing semantic similarity among different ranges of word frequency, polysemous degree and similarity intensity.

Which normally transitive verbs can omit their objects in English (I ate), and why? This article explores three factors suggested to facilitate object omission: (i) how strongly a verb selects its object (Resnik 1993); (ii) a verb's frequency (Goldberg 2005); (iii) the extent to which the verb is associated with a routine – a recognized, conventional series of actions within a community (Lambrecht & Lemoine 2005; Ruppenhofer & Michaelis 2010; Levin & Rapaport Hovav 2014; Martí 2010, 2015). To operationalize (iii), this article compares the writings of different communities to offer corpus and experimental evidence that verbs omit their objects more readily in the communities in which they are more strongly associated with a routine. More broadly, the article explores how the meaning and syntactic potential of verbs are shaped by the practices of the people who use them.

Both in linguistics and in psycholinguistics there is some debate about how rich or thin lexico-semantic representations are. Traditionally, in formal semantics but also in philosophy of language as well as in cognitive pragmatics, lexical meanings have been thought to be simple stable denotations or functions. In this paper, we present and discuss a number of interpretational phenomena of which the analysis proposed in the literature makes crucial use of rich meanings. The phenomena in question are cases where the assignment of truth-conditional contents to utterances seems to follow rules that do not operate on simple stable denotations or any other kind of ‘thin’ meanings but where composition takes rich structured representations as input. We also discuss problems for such accounts, which are mostly based on the inability of extant rich meanings accounts to explain many other interpretational phenomena, and we discuss the solutions that have been proposed to solve them. Furthermore, we address the discussion whether the informationally rich meanings are part of semantics, and more specifically part of the lexicon, or whether this information should be ascribed to more general world knowledge.

Alleviating pain is good and abandoning hope is bad. We instinctively understand how words like alleviate and abandon affect the polarity of a phrase, inverting or weakening it. When these words are content words, such as verbs, nouns, and adjectives, we refer to them as polarity shifters. Shifters are a frequent occurrence in human language and an important part of successfully modeling negation in sentiment analysis; yet research on negation modeling has focused almost exclusively on a small handful of closed-class negation words, such as not, no, and without. A major reason for this is that shifters are far more lexically diverse than negation words, but no resources exist to help identify them. We seek to remedy this lack of shifter resources by introducing a large lexicon of polarity shifters that covers English verbs, nouns, and adjectives. Creating the lexicon entirely by hand would be prohibitively expensive. Instead, we develop a bootstrapping approach that combines automatic classification with human verification to ensure the high quality of our lexicon while reducing annotation costs by over 70%. Our approach leverages a number of linguistic insights; while some features are based on textual patterns, others use semantic resources or syntactic relatedness. The created lexicon is evaluated both on a polarity shifter gold standard and on a polarity classification task.

Conversion is a key type of word-formation process in English, but the precise nature of the relation between base and derivative in conversion is rarely discussed, even if conversion is considered as a dynamic process. When it is considered explicitly, the relation has been described in terms of paradigmatic relations between lexemes, specifically homonymy or polysemy. This is usually without any specification of how converted words accommodate the conditions set by the definition of each of these relations, and as a special type of one or the other, because conversion-related words violate some of those conditions. This article is intended as a systematic review of the literature that discusses the relation between conversion-related words in English. We show that a wide range of proposals have been made to describe the relation: homonymy, heterosemy, homomorphy, zero-derivation (as a relation), polysemy, lexical extension, synsemy, hyponymy and paronymy. We review the extent to which each of these terms fits the relationship in major types of conversion, and argue that, if a relationship is to be described between conversion-related pairs, then Cruse's (1986) separation of semantic relations of a paradigmatic type from paronymic relations is of special relevance here. We propose that, regardless of the direction and type of meaning, paronymy applies across the various specific semantic patterns that conversion may involve. We emphasize, however, the possibility of several relations according to the type of conversion, i.e. different types of conversion may need description in terms of a different relation.

In this article we explore the range of aspectual and quantificational readings that are available to two kinds of deverbal nominalizations in English, conversion nouns and -ing nominals. Using data gathered from the Corpus of Contemporary American English (COCA) and the British National Corpus (BNC), we examine the range of readings available for the conversion and -ing forms of 106 English verbs in context. We distinguish eventive versus referential readings, looking at instances of both count and mass quantification for the two kinds of nominalizations. Within the eventive readings we also distinguish bounded versus unbounded aspectual readings, and within bounded readings two types that we call ‘completive’ and ‘package’. We argue that the quantificational properties and aspectual intepretation of both conversion and -ing nominalizations are not rigidly or even loosely determined by the form of the nominalization, but that the lexical aspect of the base verb (state, activity, accomplishment, achievement, semelfactive) plays some role in circumscribing aspectual readings. We argue that the strongest role in determining quantificational and aspectual readings is played by factors arising from the context in which conversion forms and -ing nominalizations are deployed. The aspectual interpretation of conversion and -ing nominalizations can be influenced by the presence of temporal and quantificational modifiers, by surrounding tenses, as well as by encyclopedic knowledge. We conclude with a consideration of the theoretical implications of our findings.

This article is concerned with plural marking in two English-lexified creoles, Jamaican Patwa and Tok Pisin. In addition to bare plurals, these creoles possess two overt strategies of plural marking—a free-standing morpheme and the suffix -s. The analytic and inflectional plural markers occur according to different linguistic constraints. It appears that the creoles use two conceptually and typologically different number marking systems — that of set noun languages, based on the opposition between singleton and collective sets, and that of singular object noun languages, based on the opposition between singular and plural individuals. This poses problems for the definition of the lexical semantics of the creole nouns if one assumes the existence of cross-linguistic differences. The analysis proposed here is based on the universalist approach to lexical semantics. Under this approach, individuated and collective (set) interpretations of plurals are encoded in the noun phrase structure.

This paper looks at a hitherto unexplored aspect of taxonomically organized concepts which has to do with word distributions in corpora of actual language use. In parallel to the psychological informativeness claim of the differentiation explanation, the question is addressed if concepts are internally more similar than their higher-ranked taxonomical relatives. This internal similarity is measured by making use of token-based vector space models. For each occurrence of a concept in the corpus a context vector can be calculated, which then serves as input for the internal similarity measure. Experiments are conducted for taxonomies taken from the Dutch counterparts of the English semantic domains animal and means of transportation. Results do not wholeheartedly agree with the imposition of a strict taxonomical order, but give rise to a new behavioural measure of the basic level.

The kinds of entities that can be described as causing an event depend, in part, on the language one speaks. Whereas in English and Chinese it is possible to say The knife cut the bread or The key opened the door, in Korean and many other languages, such sentences sound very odd. According to the initiator hypothesis, languages fall into two major groups with respect to possible external arguments in causal expressions: those that require that the causer be capable of generating its own energy and those that require only that the causer participate in the causal chain leading up to a particular result. In support of this hypothesis, we show that ability to self-energize has a larger impact on acceptability ratings in Korean than in either English or Chinese (Exp. 1). We also show that restrictions on possible causers extend to the semantics of possible causes in the descriptions of animations depicting causal chains (Exp. 2). Finally, we show that cross-linguistic differences in the linguistic coding of causers may have consequences for the way people conceptualize animations of causal chains in terms of number of events (Exp. 3). Implications for the representation of verb meaning and the semantics of external arguments in other languages are discussed.

The property generation task (i.e. “feature listing”) is often assumed to measure concepts. Typically, researchers assume implicitly that the underlying representation of a concept consists of amodal propositions, and that verbal responses during property generation reveal their conceptual content. The experiments reported here suggest instead that verbal responses during property generation reflect two alternative sources of information: the linguistic form system and the situated simulation system. In two experiments, properties bearing a linguistic relation to the word for a concept were produced earlier than properties not bearing a linguistic relation, suggesting the early properties tend to originate in a word association process. Conversely, properties produced later tended to describe objects and situations, suggesting that late properties tend to originate from describing situated simulations. A companion neuroimaging experiment reported elsewhere confirms that early properties originate in language areas, whereas later properties originate in situated simulation areas. Together, these results, along with other results in the literature, indicate that property generation is a relatively complex process, drawing on at least two systems somewhat asynchronously.