Morphologically complex forms are related to their bases in two ways: some alteration (perhaps vacuous) of form is correlated with some alteration (also possibly vacuous) of meaning. In a sort of ideal case, this is representable by the addition of a morpheme bearing both phonological and semantic content, comparable to a Saussurean (minimal) sign. Criticism of the claim that this is indeed the general case has focused on formal relations that cannot be seen as strictly additive: ablaut, umlaut, consonant shifts, metathesis, truncation, and other markers that apparently change the shape of the base rather than simply adding material to it. The present paper brings into the discussion the opposite side of this coin: instances in which it is the semantics, rather than the phonology of the base form, that is altered in a non-additive way in a morphologically derived form. Specific cases involve the removal of a component of meaning associated with one argument of the base, correlated with the addition of a component of form. These examples constitute a challenge for morpheme-based views of morphology, comparable to that posed by non-concatenative formal markers, and a challenge to the claim that morphology is always semantically monotonic.

The paper investigates placeholders, such as German Dings(bums) or English thingy. They are used in informal speech particularly for person or place names, when the speaker has forgotten them or doesn’t know them. As it turns out, in a sample of twenty-nine languages, more than half of them show only phrases or phrasal compounds of the question type (e.g. what is s/he/it called) or the deictic type (e.g. that/this). The other half use simple words or word formations, usually with a negative meaning. Compounds and derivations exist solely in Romanic and Germanic languages, however. Therefore, in a second step, I will take a closer look at placeholders in Modern German.

As noted by Bauer, real dvandva compounds – that is, coordinative compounds that properly express the aggregation of two different entities, not the intersection of properties in one entity – are extremely rare in English and Spanish. This article explores the empirical domain of dvandva compounding in Spanish, and notes that they are productive when not used as heads within their phrases. We propose that the explanation for this is that Spanish can only productively build dvandva compounds using flat structures without internal hierarchy. This causes the compound to look externally for a head noun that defines the interpretation of the relation established between the two members of the dvandva. The proposal also explains why proper names are preferred in dvandva compounding, given that they do not denote properties.

Recent research on the acoustic properties of S-final words in English has revealed unexpected effects of morphology on phonetic realization, especially on acoustic duration (Zimmermann 2016, Plag et al. 2017, Seyfarth et al. 2017, Tomaschek et al. 2019). In this paper, we investigate 462 plural tokens and 417 genitive-plural tokens elicited in a production experiment. The statistical analyses show that plural S is significantly shorter than genitive-plural S. The duration effect is, however, not restricted to the final S, but extends over the whole word. We find that word-form frequency is predictive of duration, resulting in shorter durations for (the more frequent) plurals, and longer durations for (the less frequent) genitive plurals.The empirical findings also have implications for morphological theory. Word-form frequency effects for regularly inflected words in speech production are at odds with theories in which only morphologically irregular words, or highly frequent regular words, are assumed to be stored (e.g. Pinker 1999, Alegre & Gordon 1999). The word-form frequency effect can be more naturally accounted for in word-and-paradigm models of morphology (e.g. Matthews 1974, Blevins 2016), in which individual word-forms may have representations in a network of morphologically related forms.

The paper researches all the occurrences of a stratified sample of lemmas tagged both as noun and as verb in the British National Corpus, plus a set of parallel samples of denominal verbs derived by affixation intended for comparison. This design records the distribution of semantic categories not according to the lemmas in the sample, but according to how much and where they are actualized in each occurrence of the lemma (i.e. it separates the various senses within the lemmas). The sample starts out from the lowest frequency of occurrence (1). In the case of conversion, the sample spans from frequency 1 to frequency 960. Semantic categories are recorded according to the analysis of the concordances available under each of the lemmas recorded in the sample, and based on the terms of the definitions available in the Oxford English Dictionary for each sense of each lemma. The results show that the patterns described for noun-to-verb conversion have a rather dissimilar relevance with respect to each other. The distribution of semantic categories according to their occurrences in the senses attested by each of the occurrences reported for the lemmas of the corpus sample in conversion and in affixation is quite uneven.

I compare and contrast the properties of uninflecting words such as almost, and uninflectable words such as the indeclinable Russian noun kenguru ‘kangaroo’ or the German adjective rosa ‘pink’, which behave like ordinary nouns/adjectives syntactically, but lack the expected inflectional forms. I relate such lexically uninflectable lexemes to the case of inflectable lexemes in constructions which do not permit inflected forms, such as the noun-head noun position in English N-N compounds, or the German predicative adjective (constructional uninflectability). Uninflecting words raise the question of what it is exactly that gets lexically inserted into syntactic representations. I provide a uniform solution to all these questions by appealing to Stump's distinction between content and form paradigms. Uninflecting lexemes have neither type of paradigm; uninflectable lexemes have a content but no form paradigm; construction uninflectability specifies that both content and form paradigms are bypassed. Lexical insertion of a lexeme lacking, or constructionally deprived of, a form paradigm is defined by a maximally general Default Exponence Principle: 'use the root form'. This approach solves all the descriptive and conceptual problems outlined above.

Blevins (2018) presents a new reconstruction of Proto-Basque, the mother language of modern Basque varieties, historical Basque, and Aquitanian, grounded in traditional methods of historical linguistics. Building on a long tradition of Basque scholarship, the comparative method and internal reconstruction, informed by the phonetic bases of sound change and phonological typology, are used to explain previously underappreciated alternations and asymmetries in Basque sound patterns, resulting in a radically new view of the proto-language. One aspect of this new reconstuction involves word-internal structural elements unrecognized in previous work. This paper builds on the derivational patterns hypothesized for Proto-Basque in Blevins (2018), focusing on the following formatives: root-extension *-r; acategorial *s-; nominalizating *-s; nominal *ha-; and collective *hi-. Under this analysis, lexemes may have the maximally complex structure.

Not only formal word families are important for the organization of the morphologically complex lexicon, but also their semantic (or rather conceptual) subfamilies. This is shown with eighteen Austrian German diminutive compound families with the diminutive suffix -chen, and for their correspondent formal word families with the diminutive suffix -erl. For all of them, their competition (leading to mutual avoidances and semantic specializations), degree of morphosemantic transparency/opacity, productivity and left- vs. right-branching are accounted for. Our electronic database consists of nearly 13,000 different diminutive types which have at least five tokens within the Austrian Media Corpus.

This paper discusses the ways in which formal or model theoretic semantics has difficulty addressing questions of lexical semantics, specifically questions that arise in accounting for the pervasive polysemy of deverbal nominalizations like German Bemalung (painting) and English blending. Formal or model theoretic semantics has largely neglected issues in which analysis of complex words comes to the fore. I briefly review the history of model theoretic semantics and its relation to lexical semantics and illustrate the tendency for conceptual semantics to be quietly integrated into referential semantics, as, for example, in a recent article by Pross (2019). I conclude with a brief sketch of how Lieber’s (2004, 2016) Lexical Semantic Framework treats the polysemy of deverbal nominalizations.

In many domains of linguistics, theoretical differences have led to entrenchment and a certain degree of fragmentation. Morphology seems to be different. Theoretical positions differ substantially, but the differences never get in the way of informing oneself about the reasons for adhering to a different framework, making use of it. In this volume, the following frameworks are discussed: a-morphous morphology (Anderson 1992), word and paradigm morphology (Blevins 2016), paradigm function morphology (Stump 2001, 2016), onomasiological approaches (Dokulil 1962, Štekauer 1998), construction morphology (Booij 2010), lexical semantic framework (Lieber 2004, 2016), and neo-constructionist approaches (Hale & Keyser 2002) such as distributed morphology (Embick 2015). This volume thus contains a wealth of theoretical approaches, methodologies, and descriptive issues: a fitting tribute to a linguist who made it his hallmark to serve the linguistic community with a broad range of textbooks, monographs, and research articles.

According to word and paradigm morphology (Matthews 1974, Blevins 2016), the word is the basic cognitive unit over which paradigmatic analogy operates to predict form and meaning of novel forms. Baayen et al. (2019b, 2018) introduced a computational formalization of word and paradigm morphology which makes it possible to model the production and comprehension of complex words without requiring exponents, morphemes, inflectional classes, and separate treatment of regular and irregular morphology. This computational model, Linear Discriminative Learning (LDL), makes use of simple matrix algebra to move from words’ forms to meanings (comprehension) and from words’ meanings to their forms (production). In Baayen et al. (2018), we showed that LDL makes accurate predictions for Latin verb conjugations. The present study reports results for noun declension in Estonian. Consistent with previous findings, the model’s predictions for comprehension and production are highly accurate. Importantly, the model achieves this high accuracy without being informed about stems, exponents, and inflectional classes.The speech errors produced by the model look like errors that native speakers might make. When the model is trained on incomplete paradigms, comprehension accuracy for unseen forms is hardly affected, but production accuracy decreases, reflecting the well-known asymmetry between comprehension and production.

This paper provides a contrastive analysis of English and Slovak onomatopoeia with regard to both its theoretical comprehension in two different linguistic traditions and its characteristics in two typologically different languages. The emphasis is on (i) the onomatopoeia-founded word-formation processes and (ii) the semantic characteristics of complex words derived from onomatopoeia. The latter will focus on the identification of metaphorical and metonymical shifts from the cognitive-semantic point of view. Our sample includes 120 Slovak and 115 English onomatopoetic words. In addition, we argue that (i) onomatopoeia differs from interjections and therefore should not be classified within this class of words, and (ii) the scope of onomatopoeia should not be extended beyond its inherent feature of being imitations of sounds by means of a language-specific system of phonemes. By implication, no complex word derived from onomatopoeia should be treated as genuine onomatopoeia.

This article investigates the presence of intermediate morphological categories between stems and affixes, the so-called affixoids (prefixoids and suffixoids) in a number of Modern Greek varieties. It argues that affixoids can exist as an independent morphological category in morphologically rich languages, such as Modern Greek, on the condition that these languages base their word-formation processes on stems. It also shows that affixoids can vary within the same language, depending on several factors, such as the original items the affixoids come from (Greek affixoids may emerge from lexemes or affixes), the word-formation process which gives birth to them, or the occurrence of a grammaticalization or a degrammaticalization process. Proposals and arguments are supported by evidence consisting of two prefixoids, plako- and mata-, and one suffixoid, -opulo, drawn from the Modern Greek varieties.