Cross-disciplinary perspectives on lexical blending is a selection of papers from the International Conference on Lexical Blending held in Lyon in 2010, plus one invited contribution.Footnote ¹ Lexical blending (marked by ←), generally understood as fusion of two lexemes into one involving partial loss of the phonological and/or graphical material of at least one of them (e.g. staycation←stay+vacation, shress←shirt+dress) has recently become a widely discussed topic. The volume under review presents a slice of diverse views on blending on the basis of data from typologically different languages and from the perspective of different linguistic disciplines. The volume consists of an introduction, in which the editors give a brief summary of the problem and of the approaches represented in the book, followed by twelve chapters, information about the contributors, and an index. In this review I will comment on each contribution separately and try to map them all onto a unified picture of the current situation in the linguistic research of lexical blending.

The opening chapter, ‘Blends: Core and periphery’, by Laurie Bauer outlines criteria for defining blends represented in the linguistic literature. Due to the fact that the linguists often ‘disagree about the classification’ (11), Bauer lists a set of ‘defeasible constraints’ (11) to describe prototypical phonological, structural and semantic features used by different researchers to classify blends and to distinguish them from other word formation categories (for example, clipping compounds such as sci-fi). As the examples of prototypical or less prototypical blends are introduced, the author comes to a conclusion that what is considered a blend according to one approach, may not be regarded as a typical blend within a different framework, i.e. that ‘the category is a fuzzy one’ (21). This provides the reader with an initial perception of the complexity of the issue discussed in the volume, as well as with a healthy portion of scepticism concerning the findings reported in the following chapters.

The contribution by López Rúa, ‘Beyond all reasonable transgression: Lexical blending in alternative music’, presents a description of blends used as alternative music band names. Blends are classified according to their level of integration (or phonological overlap), the type of the constituents (full words, splinters, and various combinations of those), grammatical and semantic structure using the criteria from López Rúa (Reference López Rúa2004) . The chapter contains a number of stunning examples of blends that involve not only unusual combinations of source material, e.g. RJD2←R2D2+RJ (26), but sometimes intricate, multi-level play on words as in the Foxymorons which can be analysed as either foxy+oxymorons or as foxy+morons (29). The analysis itself does not go far beyond confirming what has already been stated in the literature about blends in general (e.g. that two source forms are blended more often than three or more) and does not give any account of how this very specific corpus is similar to or different from blends coming from other sources and therefore to what extent the results of the analysis can be applied to other blends.

‘Blend formation in Modern Greek’ (Angela Ralli & George J. Xydopoulos) is analysed with regard to the similarly of the blends' structure and meaning to ‘stem word’ compounds which are productive in this language (37). The formal classification of blends is made with respect to their syllabic structure and the number of phonological segments preserved from each source word. The authors relate the form to meaning by stating that ‘the extent of form reduction varies, depending on the speaker's willingness to communicate a smaller or greater part of the meaning of the combination’ (43). It is worth noting that the extent of reduction in some of the blends is estimated without consideration of the similar prosodic contours of the source words, as in vlά.mα ‘extremely stupid’←vlά.kα ‘stupid’+ vlί.mα ‘thick’ (42), or the overlap, as in kré.vo.me ‘burping while having a haircut’←ku.ré.vo.me ‘have a haircut’+ ré.vo.me ‘burp’ (43), which is significant for recognisability of the source words. A different approach is provided, for example, in the chapter by Stefan Th. Gries (see below).

‘Lexical blending in Polish: A result of the internationalisation of Slavic languages’ (Ewa Konieczna) begins with the recent increase of productivity of clipping, compounding and blending in Polish under English influence. The classification of structural and semantic types of Polish blends, derived from Lehrer's (Reference Lehrer2003) classification of English blends, marks some differences in the frequencies of various structural types of blends in the two languages. One of the differences is a relatively high proportion of ‘graphical blends’ in Polish, which may be due to the fact that a large amount of data was collected from printed mass media. An example cited is POstęp←PO (the acronym of Poland's major political party) + postęp ‘progress’ (63). On the other hand, the most productive types of blends in both English and Polish are reported to be similar. In terms of surface structure it is the initial splinter of one word followed by a full word, and in terms of the underlying structure it is that of a subordinate endocentric compound. The paper also looks at a number of foreign elements in Polish blends, coming primarily from English (e.g. fitfala ‘fitness wave’←fitness+fala ‘wave’) or from classical languages via Russian (e.g. immulinia ‘immunological line’←immunologiczna ‘immunological’+linia ‘line’), as a sign of internationalisation of Polish.

The chapter by Susanne R. Borgwaldt, Tetyana Kulish & Arpita Bose, ‘Ukrainian blends: Elicitation paradigm and structural analysis’, focusses on blends in another Slavic language, experimentally induced in a psycholinguistic study using hybrid-object naming task with the same stimuli as were used in Borgwaldt & Benczes (Reference Borgwaldt, Benczes and Schönefeld2011) for German and Hungarian. The authors draw the reader's attention to the fact that the majority of the hybrid names in their data are blends and complex clippings, rather than compounds (in contrast to German and Hungarian). The analysis of the Ukrainian data aims to investigate the structure of the blends and to compare their features to the ones in other corpora of lexical blends. Although blends are explicitly distinguished from clipping compounds in this study (with the main analysis focussing on blends only), in some of the tables reporting the results of the data analysis these two categories are not separated. Some features of blends produced by Ukrainian speakers seem to differ from what is described in Bat-El (Reference Bat-El and Brown2006), Gries (see below) and other work, in terms of the relative length of the source words and the contribution of the second source word to the body of the blend. These differences may be either a feature of Ukrainian blends in general, or only of the spoken blends induced using a particular experimental technique; therefore the authors suggest replicating the experiment with speakers of different languages in order to make cross-linguistic comparisons.

Most contributions to the volume either exclude clipping compounds or complex clippings from the analysis, or regard them as a subtype of blends. In contrast to this, Giorgio Francesco Arcodia & Fabio Montermini, in the chapter ‘Are reduced compounds compounds? Morphological and prosodic properties of reduced compounds in Russian and Mandarin Chinese’, refer to blends as a type of ‘reduced compounds’ (93). The formations that are studied, namely ‘stump compounds’ in Russian and ‘reduced compounds’ in Mandarin Chinese, are not ‘canonical’ blends of the brunch←breakfast+lunch type but are formally closer to complex clippings. Moreover, according to the subsequent chapter in this volume, they cannot be analysed as blends at all (119ff.). The structure and formation of reduced compounds is explained using phonological constraints which are responsible for the formation of Minimal Prosodic Word and operate on the input level for Russian and on the output level for Mandarin (110). Semantically, reduced compounds in both languages, mostly collected from earlier publications, are analysed as conventionalisations of ‘regular’ compounds.

The key notion in the chapter ‘Blending between grammar and universal cognitive principles: Evidence from German, Farsi and Chinese’, by Elke Ronneberger-Sibold, is the semantic transparency of blends. The author distinguishes between four transparency levels, the so-called ‘telescope blends’, e.g. German Kamelefant←Kamel ‘camel’ + Elefant ‘elephant’ (115) seen as the most transparent, and ‘fragment blends’, e.g. Persil←Wasserstoffperoxyd ‘hydrogen peroxide’ + Silikat ‘silicate’ (118) as the least transparent. The degree of transparency is determined not only by the amount of material preserved from the beginning or from the end of the source words, but also by their prosodic contour. Blends which are formed by ‘inserting one word in to the abstract sound shape of the other’ (23; see also Piñeros (Reference Piñeros2004) for a similar approach to Spanish blends) are termed ‘contour blends’ and are claimed to be attested in all three languages under consideration, unlike ‘telescope blends’ which, according to Ronneberger-Sibold, are not attested in Farsi or in Mandarin. The productivity of ‘contour blends’ is interpreted as evidence of universal cognitive ability to recognise a word by its prosodic shape. This generalisation, however promising, might need more evidence not only because the data from Farsi and Mandarin are quite scarce in comparison to German (27 Farsi and 17 Mandarin blends vs. 612 German blends), but also because, as blending in Mandarin is possible only for compounds (132), it is not always clear whether a form should be interpreted as a blend or simply as a compound.

An analysis of a large corpus of English blends is provided in the contribution by Stefan Th. Gries, entitled ‘Quantitative corpus data on blend formation: Psycho- and cognitive-linguistic perspectives’. The author brings together data from several case studies (partly replicated from his earlier research) and, using a variety of statistical methods, draws conclusions regarding various properties of intentional blends as compared to speech error blends, on the one hand, and to complex clippings, on the other. The differences that are reported are related to relative length of the source words, their similarity to the blend, and the position of the ‘split point’, which, in turn, relates to the degree of recognisability of the source words. The evidence from all the case studies leads Gries to well-grounded conclusions about the production of blend words, although its ‘stages’ (source word selection, their ordering and blending) are, as the author himself cautions, not to be treated as ‘a characterization that is isomorphic to the actual psycholinguistic processes’ (147).

The chapter by Christian Bassac, ‘A Combinatory Logic and formal-semantic account of lexical blending’, addresses the structure and meaning of blends from the position of formal logic. First, the form of blends is described using logical operators controlling the number of preserved or deleted letters. However, this description does not seem to suggest any predictions about the formation of blends. In the next section, an attempt is made to decompose the semantics of blends relying on the Qualia structure, in the framework of the Generative Lexicon Theory (Pustejovsky Reference Pustejovsky1995). The analysis implies that, semantically, blends are similar to compounds, both in what can be predicted about their meanings and what cannot.

The last three chapters of the volume present analyses of the form of blends from the point of view of varieties of Optimality Theory (OT). The first of these chapters, ‘Stress in English blends: A constraint-based analysis’, by Outi Bat-El & Evan Gary Cohen, aims to explain and predict the position of stress in polysyllabic English blends (deliberately excluding from the analysis clipping compounds and fully overlapping blends). A set of faithfulness constraints is used to explain the position of stress in blends with regard to the stress and the length of the source words. This is done separately for the cases where the blend has the same number of syllables as at least one of the source words and where it has a different number of syllables. To account for the situation when the right-hand source word is monosyllabic and, despite the expected right-hand stress on this element, a blend acquires the default word stress, the authors ‘assume that monosyllabic words are not lexically stressed’ (207), and the model based on this assumption successfully explains nearly all the observed cases.

‘Output-to-output faithfulness in the phonological structure of English blends’ is discussed by Ewa Tomaszewicz. The regular patterns in the phonological structure of blends are viewed as the output of the process of mapping the phonological material of two source words onto a template of a single prosodic word. The main finding reported in this chapter is that blending is, by nature, not a subtractive process, in contrast to complex shortening, resulted in the formation of clipping compounds. An analysis of the phonology of Polish blends lends further support to the presented model.

A similar finding is reported by Jochen Trommer & Eva Zimmermann, in the chapter ‘Portmanteaus as generalized templates’, analysing a specific type of Spanish blends referred to as portmanteaus in the sense of Piñeros (Reference Piñeros2004). Following the analysis of English truncations in Lappe (Reference Lappe2007) as mapping the phonological material of a word onto the template of a Minimal Prosodic Word, the authors explain blending as mapping the phonological material of one source word onto the prosodic template of another by substituting the required segments. Their model is claimed to have more explanatory power than that of Piñeros (Reference Piñeros2004), but it applies only to a specific type of blends – those that meet its criteria – which makes the main argument circular.

The book is well set out, except for a few cases of incorrect transliterations (for example, of Russian and Ukrainian data) and typographical errors. As a whole, the volume under review provides a deep insight into blending in a number of languages from the point of view of a variety of theoretical frameworks, which makes it one of the key contemporary publications concerning this topic. The collection provides well-selected and diverse coverage of different aspects of the phenomenon under consideration. On the other hand, demonstrating multiple paths to approaching blends, it seems to contain more questions than answers, which will be a source of inspiration for further research.