Blending in Morphology
Blending in Morphology
- Natalia BeliaevaNatalia BeliaevaSchool of Linguistics and Applied Language Studies, Victoria University of Wellington
Blending is a type of word formation in which two or more words are merged into one so that the blended constituents are either clipped, or partially overlap. An example of a typical blend is brunch, in which the beginning of the word breakfast is joined with the ending of the word lunch. In many cases such as motel (motor + hotel) or blizzaster (blizzard + disaster) the constituents of a blend overlap at segments that are phonologically or graphically identical. In some blends, both constituents retain their form as a result of overlap, for example, stoption (stop + option). These examples illustrate only a handful of the variety of forms blends may take; more exotic examples include formations like Thankshallowistmas (Thanksgiving + Halloween + Christmas). The visual and audial amalgamation in blends is reflected on the semantic level. It is common to form blends meaning a combination or a product of two objects or phenomena, such as an animal breed (e.g., zorse, a breed of zebra and horse), an interlanguage variety (e.g., franglais, which is a French blend of français and anglais meaning a mixture of French and English languages), or other type of mix (e.g., a shress is a type of clothes having features of both a shirt and a dress).
Blending as a word formation process can be regarded as a subtype of compounding because, like compounds, blends are formed of two (or sometimes more) content words and semantically either are hyponyms of one of their constituents, or exhibit some kind of paradigmatic relationships between the constituents. In contrast to compounds, however, the formation of blends is restricted by a number of phonological constraints given that the resulting formation is a single word. In particular, blends tend to be of the same length as the longest of their constituent words, and to preserve the main stress of one of their constituents. Certain regularities are also observed in terms of ordering of the words in a blend (e.g., shorter first, more frequent first), and in the position of the switch point, that is, where one blended word is cut off and switched to another (typically at the syllable boundary or at the onset/rime boundary). The regularities of blend formation can be related to the recognizability of the blended words.
1. Blends in Word Formation: An Overview
Brexit (Britain + exit), Trustafarians (trust fund + rastafarians), affluenza (affluent + influenza), and other blends were listed among top words of the first fifteen years of the 21st century, according to the Global Language Monitor (2015). These and other numerous blends in advertisements, headlines, blogs, and various other media, illustrate that blend words occupy an important niche in contemporary vocabulary. This is not to say that blending is a new phenomenon in word formation. Examples of early attested blends are rebuse (rebuke + abuse) from Early Modern English (Pound, 1914, p. 6), écornifler ‘to nose about and steal’ (écorner ‘to cut off’ + nifler ‘to sniff’) from Middle French, and pstrokaty ‘motley’ (pstry ‘multicoloured’ + srokaty ‘piebald’) from Middle Polish (Renner, Maniez, & Arnaud, 2012, p. 1).
Blends underwent a boost in popularity in the late 19th century. The novel Through the Looking Glass (1871) by Lewis Carroll, containing the now classical blends like mimsy (miserable + flimsy) and galumph (gallop + triumph), catalyzed the use of blends, particularly in English. Carroll’s work also gave rise to the term ‘portmanteau word’, which is used in morphological studies either as a synonym of ‘blend’, for example, in Pound (1914) and Thurner (1993), or as its hyponym, denoting a certain type of blend, for example, in Algeo (1977), Piñeros (2004), or Tomaszewicz (2012) (see section 2.2 for details).
It is worth noting that the academic works of the late 19th century utilized the term ‘blend’ mainly for speech errors. For example, in Meringer and Mayer (1895) the term ‘blend’ labels slips of the tongue such as evoid, a blend of phonologically and semantically similar words avoid and evade. However, in early-20th-century literature the term ‘blend’ acquired the meaning it has in contemporary morphology, that is, to name a new word formed by fusing two known words into one. Blends listed in Wood (1911), such as brunch (breakfast + lunch) and canoodle (canoe + paddle), suggest intentional formation of blends to convey a specific meaning (e.g., a meal that includes both breakfast and lunch), or to use one word for a complex action (paddle a canoe). The distinction between speech errors and intentional blends is discussed in Kelly (1998), and a comprehensive summary of phonological and semantic differences between speech errors and intentional blends is provided in Gries (2004b, 2012). Henceforth, the distinction between error blends and intentional formations will not be pursued here, the main focus of this article being on blending as a morphological phenomenon.
Blending as a type of word formation is fascinating because blends are incredibly diverse (as emphasized in Dressler, 2000; López Rúa, 2004; Mattiello, 2013; and other studies), and at the same time surprisingly predictable (as argued by Bat-El, 1996; Plag, 2003; Gries, 2012; and other scholars). Particular patterns have been observed in terms of the selection of the words that are blended and the formal structure and the semantics of blends. The formal and semantic properties of the words that become blend constituents are considered in detail in section 2. The process of blending and the structure of blends are elaborated on in section 3. The domains of use are outlined in section 4, and section 5 provides an overview of blending across different languages. Section 6 characterises blends as a morphological phenomenon.
2. What Can Be Blended?
2.1 Formal Properties
Blends are formed by fusing two (or sometimes more) words together, so that part of the material of these words is lost. This distinguishes blends from other types of word formation, although using just this definitional feature may not be enough to discriminate between blends and, for example, initialisms. The boundaries of the category are, therefore, fuzzy, and determining whether a given formation is a blend requires considering multiple factors, which are discussed in section 3. The words that are blended will henceforth be referred to as source words, as it is the most widely accepted term in the relevant literature (Cannon, 1986; Kubozono, 1990; Lehrer, 1998; Kemmer, 2003; Gries, 2004b), some of the alternatives being component words (Kelly, 1998), input words (Brdar-Szabó & Brdar, 2008), formatives (Fradin, 2000), or source forms (López Rúa, 2004). The latter term is justified to use when a blend is formed using units larger (see 1c–d) or smaller (1f) than words.
As exemplified in 1, various language units may participate in blending: abbreviations (1a–b), phrases (1c), complex names (1d–e), and neoclassical combining forms (1f). These cases are marginal in the sense that their constituents deviate from the majority of attested blends (note that 1d–f, presumably creative coinages, could have been formed to stand out in order to attract more attention).
The words to be blended are not selected at random. As have been shown in Kubozono (1990), Kelly (1998), Gries (2004a), and other studies, the source words of many blends display a certain degree of phonological or graphical similarity. In most cases, the locus of similarity is where the switch point occurs, that is, where the switch from one source word to the other takes place. The source words of blends like motel (motor + hotel), mockbuster (mock + blockbuster) and jumbrella (jumbo + umbrella) contain identical segments (indicated henceforth by bold type), and this is where the two splinters overlap in the resulting blends. In some blends, the overlap is purely phonological, as in burqini (burqa+ bikini), or in fauxbia (faux + phobia), which, in fact, can be recognized as a blend only in written form because, phonologically, it fully coincides with its second source word. On the other hand, graphical overlap may not be entirely reflected in pronunciation, as in mascary (mascara + scary), where the second vowel is pronounced as in scary /ˈskeə.ri/, and not as in mascara /mæsˈkɑː.rə/. The similarity between the source words may not necessarily be limited to the switch point. In fact, there are many blends where similar phoneme / letter strings, are located elsewhere, as shown in (2).
Blends in 2a–c and 2e are formed from words with similar initial segments. The source words of blends in 2a–b have more than one initial letter and phoneme in common, and the source words in 2c and 2e share the same first letter plus the stressed vowel. In 2c, more similarity is observed on the graphical level because the initial phonemes of song and shoplifting are different. The contrast between phonological and graphical similarity is even more obvious in 2d, where the first source form is the abbreviation of ‘World Wide Web’ that is used in web addresses and is often pronounced as /ˌdʌb.dʌbˈdʌb/, a shorter variant of /ˈdʌb.l ̩.juːˌdʌb.l ̩.juːˈdʌb.l ̩.juː/. Thus, the source forms in 2d share the first phoneme, although no graphical similarity is observed. Different locus of similarity is exemplified in 2f–h: the source words of these blends share the final phoneme / letter strings. The locus of similarity in blends can therefore be different, although the analysis in Gries (2012) demonstrates that in most blends the similarity tends to be concentrated around the switch point.
Importantly, in 2g–h the similarity is due to the fact that the source words belong to the same syntactic category and therefore contain the same suffixes (the present participle suffix -ing in 2g, and the adverbial suffix ‑nie added to loan base ending in ‑jal in both source words in 2h). As pointed out, for example, in Kubozono (1990), Kelly (1998), and Renner (2006), most blends are formed from the words of the same part of speech, though other types of grammatical structure are also popular, for example, adjective + noun. In English, noun is the most frequent category of blends, followed by adjective and verb (Bauer, Lieber, & Plag, 2013, pp. 459–460). Noun dominance is also observed in other languages such as Hebrew (Bat-El, 1996), Greek (Ralli & Xydopoulos, 2012), and Russian (Hrushcheva, 2011). Most frequent combinations of source words according to their grammatical category are exemplified in (3), based on the classification in Bauer et al. (2013, p. 459):
Among other syntactic structures that are attested are, for example, adverb + adverb in 2h, or adverb + verb in the Greek blend ipulegízo ‘approach in an insidious manner’ (ípula ‘insidiously’ + prosegízo ‘approach’) (Ralli & Xydopoulos, 2012, p. 39). As noted in Renner (2015a, p. 126), sometimes blends allow grammatical structures that are otherwise illicit in the relevant language, for example, noun object preceding the verb in the French blend cadonner ‘give as a present’ (cadeau ‘present’ + donner ‘give’). Such cases, however, are marginal, and the order of the blend constituents may then be determined by factors other than syntactic, such as the locus of similarity or phonotactics (see section 3.1 for discussion).
2.2 Semantic Properties
The selection of the source words of blends is also subject to restrictions on the semantic relationship between them. In Algeo (1977), blends are classified into two main categories: syntagmatic and associative (see Figure 1). According to Algeo, syntagmatic or telescope blends result from “a combination of two forms that occur sequentially in the speech chain” (1977, p. 56), while the source words of associative or portmanteau blends are “linked in the word-maker’s mind and thence in his language” (1977, p. 57). The latter category is further subdivided into: (1) synonymic blends, (2) blends that combine words from the same paradigmatic class—paradigmatic or dvandva blends, and (3) jumble blends, the source words of which “are associated with one another, but not by paradigmatic equivalence” (1977, p. 58). An example of an alternative, but largely compatible, classification scheme is Renner (2006), where the category of tautologous blends denotes synonymic relations between the source words. Renner’s taxonomy also includes different kinds of semantic relations of words belonging to the same paradigm, in particular, additional relations, as in smog (both smoke and fog), and hybrid relations, as in zorse (a hybrid between zebra and horse).
The problem with the semantic taxonomy is that in some cases it may be unclear whether the blend is a telescope or a portmanteau, and also that several semantic relations may be realized in one blend, as shown in (4). The blends in 4a–b originate from phrases and therefore can be classified as telescope blends. On the other hand, 4c–e are portmanteau blends realizing different kinds of paradigmatic relations. However, it is likely that 4c originates from the phrase meat and potatoes, that is, has syntagmatic origin in addition to realizing additive relations. This may also be the case with 4e. An example of a blend with an unclear categorical status is 4f, which can be analyzed as either a telescope of sham amateur meaning ‘one who pretends to be an amateur but is really a professional’, or as a portmanteau of sham and amateur meaning ‘one who tries to deceive but is amateurish’. Finally, 4g has a metaphorical, rather than purely syntagmatic, origin, although its source words are in determinative relations, rather than paradigmatic.
The distinction between telescope and portmanteau blends has been reconsidered by many linguists. Some, among them Bauer (1983), Devereux (1984), and Cannon (1986), include both portmanteau and telescope words in the category of blends. Other researchers, such as Kubozono (1990) and Berg (1998), restrict the category to portmanteaus only. Such categorical restriction is problematic because the semantic type of a given blend may be ambiguous. The quantitative analysis in Gries (2012) shows that the source words of most blends have paradigmatic relations, for example, are synonyms, co-hyponyms, or belong to the same semantic frame. Moreover, the analysis of frequencies of source word combinations in Beliaeva (2014) demonstrates that blends are less likely to be formed as contractions of existing combinations of their source words than clipping compounds such as digicam (which is a clipping of digital camera— see also section 3.2 on the distinction between blends and clipping compounds). As noted in Bauer (2012, p. 18), the problem of interpreting blends as either having a semantic head (i.e., telescope) or coordinative (portmanteau) is of the same nature as the problem of interpreting compounds like fighter-bomber as either headed or coordinative (see also Bauer & Tarasova, 2013 for a discussion of the semantic relationships between the elements in a compound). Nevertheless, it is justified to distinguish between the syntagmatic origin and paradigmatic origin blends, as suggested in Bauer (2012), because such semantic analysis provides insights into the onomasiological nature of blends and can be related to the purposes and contexts of their use.
3. How Are the Source Words Blended?
3.1 Ordering of the Source Words
The order in which the source words are arranged in a blend is conditioned by the semantic properties of the blend (as defined in section 2.2) and the properties of the source words themselves. In syntagmatic blends, the order of the source words is likely to be determined by the position of the semantic head, as in floordrobe (floor + wardrobe), which is a form of storage for clothing (Urban Dictionary, n.d., ‘Floordrobe’, n.p.) rather than a type of floor. The position of the head, in its turn, varies across languages, for example, English, Serbian, and Korean blends are predominantly right-headed, while in French and Modern Hebrew left-headed blends are as common as right-headed ones (Renner, 2015b, p. 106). In the absence of predetermined order, as is the case with paradigmatic blends, other factors may determine the order of the constituents. As discussed in Kelly (1998) and recapitulated in Bauer (2012), these properties of the source words are relevant (examples from Bauer, 2012, p. 12):
Length: shorter first, e.g., donkephant (donkey + elephant);
Frequency: more frequent first, e.g., smog (smoke + fog);
Prototypicality: more prototypical first, e.g., spork (spoon + fork);
Pragmatics: e.g., in brunch, the order of constituents corresponds to temporal order of breakfast and lunch.
Another factor that influences the ordering of the source words is the tendency to maximize similarity between the blend and each of its source words. That is, if the source words have identical segments they tend to be blended in such a way that those segments are retained in the blend. In some cases, the considerations of maximal similarity overrun the influence of other factors (e.g., syntagmatic order or source word length). For example, in the English blend epiphanot ‘not an epiphany’ and in the Polish blend reformatoł ‘stupid reformer’ (reforma ‘reform’ + matoł ‘moron’) the order of the elements does not correspond to the syntagmatic order. Instead, in both blends the source words are arranged in such a way that identical segments overlap (henceforth, overlapping segments in the examples are in bold type) Similarity is not limited to the phonemic level but also encompasses the prosodic level (Piñeros, 2004). For example, in a Spanish blend crucidrama ‘a serious mess’ (crucigrama ‘crossword’ + drama ‘drama’) the shorter source word is integrated into the prosodic structure of the longer word such that the similarity between the blend and its source words is maximal. Thus, more than one relevant factor is at play in determining the ordering in each particular case (see, e.g., Piñeros, 2004 for an optimality theory [OT] analysis of the ordering of constituents in Spanish blends, and Tomaszewicz, 2012 for a similar analysis of English blends).
3.2 Structural Classification
In a typical blend, the beginning of the first source word is merged with the ending of the second one, as expressed as a formula in Plag (2003, p. 123): AB + CD = AD. However, this mechanism can be realized in a plethora of ways. First of all, a source word can either be fully retained in the blend, or be curtailed into a splinter (the term commonly used in literature, e.g., in Adams, 1973; Cannon, 1986; López Rúa, 2004; to name word parts that become blend constituents). For example, the blend chairdrobe contains its first source word chair in full, and the splinter ‑(r)drobe from the second source word wardrobe. The splinters do not coincide with existing morphs (cf. the suffix ‐‑ic in alcoholic and the splinter ‑(a)holic in the blend workaholic). A splinter, however, can be used in several blends (e.g., floordrobe) and eventually may acquire the status of a morph. Thus, a splinter, ‑(a)holic (shopaholic, chocoholic) was listed as a suffix in dictionaries in the early 21st century (Cambridge English Dictionary Online, n.d.; OED Online, n.d.). If a splinter thus becomes a productive affix, new words that it forms are affixations rather than blends.
With regard to the ways a word AB can be merged with a word CD to make a blend AD, a variety of outcomes is possible depending on: (a) whether or not there is overlap, and (b) whether or not the source words are preserved in their entirety in the blend. The classification in (5) lists the attested structural types of blends, based on Gries (2004a, p. 415), which is compatible with earlier classifications in Algeo (1977), Cannon (1986), and other studies (overlapping segments are in bold type):
Although such classification accounts for most blends, it does not exhaust the possible ways a word AB can be blended with a word CD. An example of a different (although rarely attested) structure is the blend fro(hawk) < (a)fro + (m)ohawk which contains two final splinters.
The blends in (6) are structurally similar to the examples in 5d, but in each case the first splinter retains no other phonological and graphical material than the segment overlapping with the second source word. This may complicate the processing of the blend, which is why such formations are classified as cryptoblends in Renner (2015a, p. 125).
In other attested formations, a fragment from the middle of one source word is replaced by the other source word or its part (7). Blends of this kind are referred to as infixed blends in Bauer (2012) or intercalative blends, for example, in Kemmer (2003) and Gries (2004b). This may not include any phonological or graphical overlap, as in 7a, but in a number of cases one word replaces a phonologically or graphically similar segment of another, as in 7b–h. Occasionally, only a fragment of one source word is intercalated into the other, as in 7f–g. In 7h, the fragments of an abbreviation are intercalated into the other source word, which has two similar graphemes. The examples in 7g–h illustrate the fact that blends tend to maximally preserve identical segments of their source words even if the similarity is not concentrated in one segment.
The presence of the abbreviation KPN in the blend KoPuNa is more evident in written form (capitalization is retained in the blend, which enhances recognizability). In some blends, one of the source words can be recognized only in graphic form, which is the reason why they are called graphic blends (Konieczna, 2012). Each of the blends in (8) is pronounced as one of its source words, and the presence of the other becomes noticeable due to spelling conventions (8a), or has to be made noticeable by graphical means such as parentheses in 8b or capital letters in 8c–d.
In addition, blends can be formed of more than two source forms. For example, 9a–b are three-splinter blends, and occasionally blends containing more than three splinters are attested. An example of a six-splinter blend is 9c, an inclusive term referring to the December holidays (Renner, 2015a, p. 126).
Finally, in some studies the category of blends includes items formed by concatenation of initial segments of the source words (e.g., López Rúa, 2004 and Mattiello, 2013 for English, Konieczna, 2012 for Polish), exemplified in (10). Such items may contain splinters from two (10a–c) or more (10d–e) source words. Formations retaining only one or two initial letters of each source word (10b, d–e) are very similar to acronyms, and can be classified as borderline cases between blends and acronyms (López Rúa, 2004).
In regard to the examples in (10), also classified as ‘clipping compounds’ or ‘complex clippings’, there is evidence in literature that they differ from most blends in terms of their origin (Beliaeva, 2014), structural characteristics that are related to the recognizability of the source words (Gries, 2006, 2012; Beliaeva, 2014) and processing (Beliaeva, 2016). Nevertheless, even not taking clipping compounds into consideration, blends are variable enough to conclude that the category is fuzzy. The existence of such examples as the ones in 6–9 is pointed out in some studies to support the claim that blends are “minimally predictable” (Mattiello, 2013, p. 96) and therefore must be analyzed as an extragrammatical phenomenon (Dressler, 2000; Mattiello, 2013). However, despite the observed formal diversity, blending involves a considerable degree of regularity, as will be discussed in section 3.3.
3.3 Formal Regularity
One obvious regularity that is postulated in the literature as a defining feature of blends is that most blends combine the initial part of one word with the final part of another. Other structures discussed in section 3.2 are much less frequent by comparison, as demonstrated in several empirical studies such as Kubozono (1990) and Gries (2006, 2012) for English blends, and Ronneberger-Sibold (2012) for German, Farsi, and Chinese blends. In addition, the structure of blends is “constrained by semantic, syntactic and prosodic restrictions”, as summarized in Plag (2003, p. 125). Semantic properties of blends and the most frequently attested grammatical structures were discussed in section 2.1. In terms of the prosodic structure, blends tend to conform to the phonotactic requirements of the particular language, as has been observed in studies on typologically different languages, for example, in Cannon (1986), Kubozono (1990), and Plag (2003) for English, in Bat-El (1996) for Hebrew, in Piñeros (2004) for Spanish. The phonotactic constraints on blend formation concern (1) the syllabic length of the blend in relation to the length of its source words, (2) the stress pattern, and (3) the placement of the switch point in relation to syllable constituents.
The number of syllables in the blend doesn’t deviate much from the length of the longest of its source words. That is, if both source words have an equal number of syllables, the tendency is for the blend to have the same number of syllables, as in 11a–c, or one syllable more, as in 11d–g.
If the source words are not of the same length, the longest of them determines the blend length (that is, the first source word in 12a–b, and the second source word in 12c–d). Longer blends are also attested, although, as explained in section 2, such cases are rare. For example, the blends in 12e–f are one syllable longer than the longest of their source word, which in both cases is due to the preservation of the overlapping segments.
Importantly, the correspondence between the number of syllables in the blend and in the longer source word is also relevant for structurally less typical blends such as cryptoblends in (6), intercalative blends in 7a–c, 7e–f, 7g, and graphic blends in (8) (but not clipping compounds in (10)).
The position of the stress in the blend typically corresponds to that in at least one of the source words (in languages with flexible stress: see Bat-El & Cohen, 2012 and Gries, 2004a for English; Piñeros, 2004 for Spanish; Tomaszewicz, 2012 for English and Polish). Blends tend to retain the prosodic contour (that is, the overall number of syllables and the main stress position) of the longer source word (Cannon, 1986), or the second source word (Bat-El, 1996; Gries, 2004a; Bauer, 2012). Given the considerations in section 3.1, the two cases often coincide. For example, in 13a the stressed syllables from both source words are retained in the blend, but the prosodic contour corresponds only to that of the second word. The stress pattern of the blend can repeat that of the second source word if it has more syllables than the first one (13a–c), or the same number of syllables (13d). However, in 13e the first source word is longer, and therefore it provides the stress pattern for the blend. Intercalative blends like 13f (see also 7a–c, 7e) preserve the prosodic contour of the longer source word. In blends that are longer than each of their source words, the stressed syllable of at least one of the source words remains stressed in the blend; for example, the second source word provides the stress in 13g.
As shown in (13), the stress in blends is subject to different factors such as the length of the source words, their stress, and the way they are blended. The interplay of those factors is analyzed, for example, in Bat-El and Cohen (2012) and Tomaszewicz (2012). These and other authors acknowledge that the prosodic contour of the second source word tends to be preserved more often. The blends that retain the prosodic contour of one (usually the second) of their source words are classified in Ronneberger-Sibold (2012) as contour blends, and the corpus analysis therein shows that contour blends prevail in German, Farsi, and Chinese. This is compatible with the analysis of Spanish blends in Piñeros (2004), and with the properties of experimentally induced blends (see, e.g., Arndt-Lappe & Plag, 2013 on English and Borgwaldt, Kulish, & Bose, 2012 on Ukrainian). The position of the main stress is also related to headedness. That is, the source word that provides the stress pattern of the blend tends to be interpreted as its morphological and semantic head, as demonstrated in Shaw, White, Moreton, and Monrose (2014) and Moreton, Smith, Pertsova, Broad, and Prickett (2017) for experimentally induced blends. The effect of headedness is, however, closely intertwined with that of ordering (e.g., the second source word is usually the head word in languages like English). The preservation of the prosodic contour is featured in Piñeros (2004), Gries (2012), and other studies as an important factor contributing to recognition of the source words in the blend. Thus, preserving the stress pattern compensates for the loss of the word beginning by the second source word (see, e.g., Whitney, 2001 and White, Johnson, Liversedge, & Rayner, 2008 for experimental findings demonstrating the importance of word beginning for recognition).
3.3.3 Switch Point Placement and Prosodic Structure
Switch points in blends are placed at “major phonological joints” (Kelly, 1998, p. 585). That is, they tend to fall on syllable boundaries or, failing to do that, at the boundaries of sub-syllable elements, most often between onset and rime (Kubozono, 1990; Kelly, 1998; Plag, 2003; Bauer, 2012).
If there is a phonological or graphical overlap between the source words, it is likely to define the switch point (Cannon, 1986; Bauer, 2012; Gries, 2012). As exemplified in (14), the switch point is realized as an overlapping syllable (14a–b) or a sub-syllable element (nucleus in 14c and coda in 14d; overlapping segments are in bold type). In each case, the syllables that contain similar elements are merged into one syllable in the blend.
Correspondingly, the switch point tends to fall on the syllable boundary if there is no overlap between the polysyllabic source words (15a–b). Otherwise, if the switch point falls within a syllable, the first source word provides the onset, and the second source word provides the rime. This concerns both polysyllabic (15c) and monosyllabic (15 d–e) blends. The switch point placement within the onset is also attested, for example, in 15f.
In addition to phonotactic constraints, the position of the switch point is related to the prosodic structure of the blend. In polysyllabic blends, the switch point tends to be placed within the stressed syllable of the second source word (14c, 15c), or at its left boundary (15a–b), so that the stressed vowel of the second source word is retained in the blend (see also Arndt-Lappe & Plag, 2013; Shaw et al., 2014 for more detailed discussion and experimental evidence). In the rare cases when the switch point falls within the syllable onset, blends may also violate phonotactic constraints. Thus, in 16a the complex onset contains an affricate, which is contrary to English phonotactics (Harley, 2001), and in 16b the blend starts with a consonant cluster, which is illicit word-initially in Modern Greek (Ralli & Xydopoulos, 2012, p. 43).
The fact that one can find counterexamples violating constraints is in itself in accordance with the reality of any living language. In addition, the possibility of structural and phonological transgression in blends is pointed out in Renner (2015a) as a factor that enhances the blends’ playfulness. The violation of constraints attracts attention, which is important in many contexts where blends are used (see section 4).
The regularities of blend formation discussed in section 3.3 are related to the recognizability of the source words. On the one hand, the source words are blended “so as to maximize overlap in the middle of the fusion section and maximize phonemic / graphemic similarity elsewhere as much as is still possible” (Gries, 2012, p. 164). On the other hand, the similarity should not go beyond certain limits, that is, both words have to remain recognizable. As a result, most blends are formed in such a way that one source word retains its beginning, and the other its ending and the prosodic contour, which are essential for recognizability (cf. Arndt-Lappe & Plag, 2013; Shaw et al., 2014).
The constraints concerning the number of syllables, the main stress, and the switch point placement can be extrapolated to blends of seemingly exotic types. Thus, in most infixed blends such as the ones in (7), the source words are merged where they are maximally similar to each other. In the absence of overlapping segments, the shorter source word replaces a segment of the longer source word in such a way that the blend inherits the prosodic contour of the latter. As a result, one source word retains its beginning, ending, and prosodic contour, which are crucial for recognizability, and the other is preserved in full. In the few cases when the word that is inserted is not fully preserved (e.g., 7f–g), its prosodic contour is retained.
Likewise, the structure of cryptoblends in (6) favours not only the preservation of the second source word, but also the recognizability of the first source word. In particular, the second source word in each case is clipped in such a way that the beginning of the blend coincides with the beginning of the first source word.
It is worth noting that in OT analyses of the constraints determining the structure of the blends, for example, in Piñeros (2004) and Tomaszewicz (2012), the recoverability of the source words is a key factor of the well-formedness of a blend. The results of the corpus analysis in Gries (2006, 2012) show that the switch point in blends is related to the cognitively relevant ‘uniqueness point’ marking the amount of phonological / graphical material needed to uniquely identify the word. This is compatible with the results of experimental studies in Beliaeva (2015, 2016), which demonstrate that the cognitive processing of blends involves the activation of their source words in the readers’ lexicon.
4. The Use of Blends
It has been often pointed out (Dressler, 2000; López Rúa, 2004; Renner, 2015a) that blending implies wordplay. Renner (2015a) reiterates that coining a blend is an act of wordplay in the sense that it involves “an intentional and formally ingenious way of associating the semantics of two or more words in a new morphological object” (Renner, 2015a, p. 119). It is not surprising, therefore, that blends are often used as expressive means in various domains including slang (17a), popular media (17b), political terms (17c), professional vernacular (17e–f), company names (17g), names of musical bands (17h), and other cultural groups (17j). In media contexts, blends can be accompanied by visual amalgamations illustrating their meaning.
Wordplay is not the only driving force of blend formation. Coining a blend to name an object or fact of reality can also be motivated by the nature of the object. For example, blends are often coined to name real objects or phenomena composed of two or more components, for example, animal hybrids (18a), language varieties (18b), geographical regions (18c), food or drink mixtures (18d), clothing items (18e), and the like.
More than two splinters in hybrid names like the ones in (18) reflect the nature of the hybrids. For example, 18c refers to the merged holiday including Halloween, Thanksgiving, and Christmas (Urban Dictionary, n.d.). Moreover, the structure of the blend can reflect the actual relationships between the denotata. For example, the proportions of the source words retained in the names of apricot-plum hybrids reflect the genetic profile of the hybrids: plumcot denotes a hybrid of half apricot and half plum, aprium an apricot-rich hybrid and pluot a plum-rich hybrid (Renner, 2015a, p. 124). Similarly, the order of the splinters in an animal hybrid name indicates the breeding type: liger denotes a cross between a male lion and a female tiger, and tigon a cross between a male tiger and a female lion.
5. Blends in Different Languages
As hypothesized in Brdar-Szabó and Brdar (2008), blends are possible in any language where clipping and compounding is also possible. Although blending “is mainly a feature of Indo-European languages” (Štekauer, Valera, & Kőrtvélyessy, 2012, p. 132), blends are attested in a variety of languages not limited to those mentioned in this article (see, e.g., Dobrovolsky, 2001 on Malay; Kang, 2013 on Korean; and Lepic, 2016 on American Sign Language). In different languages, the phonotactic restrictions on blending diverge; for example, Chinese blends can combine whole syllable only (Ronneberger-Sibold, 2012).
In some languages the morphological model of blending was incremented as a result of language contact. In particular, the influence of English loans is documented in the literature (see, e.g., Konieczna, 2012 on English influence in Slavic languages). Language interference also results in the manifestation of transliterated blends (19a) and cross-linguistic blends (19b–c).
6. Blending and Other Word Formation Processes
Despite the apparent formal diversity of blends, the process of blend formation involves a considerable degree of regularity and predictability, primarily in terms of the choice of the source words, their ordering, and the way they are blended. Merging two (or more) words into a blend demands maximal recognizability of each while at the same time conforming to the phonotactic requirements of a particular language. This concerns even the structurally less common formations such as the examples in 6–9. The recognizability of the source words can be regarded as a feature that defines the stance of blending in relation to other word formation processes. On the one hand, the constituents of blends are markedly more recognizable than the ones in clipping compounds or acronyms (Gries, 2006; Beliaeva, 2014, 2016). On the other hand, leaving both the source words intact would result in producing a compound instead of a blend. According to Bauer et al. (2013), blends are similar to compounds in all aspects but the phonotactic structure (although other important differences, particularly in semantic relations between the source words are pointed out in Renner, 2015b and other studies). The particularities of the structure, however, result in more cognitive effort required to process blends (Beliaeva, 2016), which makes them attention-catching, and therefore appropriate to use as expressive means of language (Mattiello, 2013; Renner, 2015a). The playful nature of many blends, in its turn, explains their dissemination in particular contexts of use such as political media and advertisement.
- Adams, V. (1973). An introduction to modern English word formation. London, U.K.: Longman.
- Algeo, J. (1977). Blends, a structural and systemic view. American Speech, 52, 47–64.
- Arndt-Lappe, S., & Plag, I. (2013). The role of prosodic structure in the formation of English blends. English Language and Linguistics, 17(3), 357–563. doi:10.1017/S1360674313000154
- Bat-El, O. (1996). Selecting the best of the worst: The grammar of Hebrew blends. Phonology, 13, 283–328.
- Bat-El, O., & Cohen, E.‐G. (2012). Stress in English blends: A constraint-based analysis. In V. Renner, F. Maniez, & P. J. L. Arnaud (Eds.), Cross-disciplinary perspectives on lexical blending (pp. 193–211). Berlin, Germany: De Gruyter Mouton.
- Bauer, L. (1983). English word-formation. Cambridge, U.K.: Cambridge University Press.
- Bauer, L. (2012). Blends: Core and periphery. In V. Renner, F. Maniez, & P. J. L. Arnaud (Eds.), Cross-disciplinary perspectives on lexical blending (pp. 11–22). Berlin, Germany: De Gruyter Mouton.
- Bauer, L., Lieber, R., & Plag, I. (2013). The Oxford reference guide to English morphology. Oxford, U.K.: Oxford University Press.
- Bauer, L., & Tarasova, E. (2013). The meaning link in nominal compounds. SKASE Journal of Theoretical Linguistics, 10, 1–18.
- Beliaeva, N. (2014). A study of English blends: From structure to meaning and back again. Word Structure, 7(1), 29–54.
- Beliaeva, N. (2015). Blends at the interface between compounding and clipping: Evidence from readers’ evaluations. Neologica, 9, 205–219.
- Beliaeva, N. (2016). Blends at the intersection of addition and subtraction: Evidence from processing. SKASE Journal of Theoretical Linguistics, 13(2), 23–45.
- Berg, T. (1998). Linguistic structure and change: An explanation from language processing. Oxford, U.K.: Oxford University Press.
- Borgwaldt, S. R., Kulish, T., & Bose, A. (2012). Ukrainian blends: Elicitation paradigm and structural analysis. In V. Renner, F. Maniez, & P. J. L. Arnaud (Eds.), Cross-disciplinary perspectives on lexical blending (pp. 75–92). Berlin, Germany: De Gruyter Mouton.
- Brdar-Szabó, R., & Brdar, M. (2008). On the marginality of lexical blending. Jezikoslovlje, 9(1–2), 171–194.
- Cambridge English Dictionary Online. (n.d.). Cambridge dictionary. Cambridge University Press.
- Cannon, G. (1986). Blends in English word formation. Linguistics, 24(4), 725–753.
- Devereux, R. (1984). Shortenings, blends and acronyms. Word Ways, 17, 210–215.
- Dobrovolsky, M. (2001). Malay blends—CV or syllable template? Calgary Working Papers in Linguistics, 23, 14–29.
- Dressler, W. U. (2000). Extragrammatical vs. marginal morphology. In U. Doleschal & A. M. Thornton (Eds.), Extragrammatical and marginal morphology (pp. 1–10). Munich, Germany: Lincom Europa.
- Fradin, B. (2000). Combining forms, blends and related phenomena. In U. Doleschal & A. M. Thornton (Eds.), Extragrammatical and marginal morphology (pp. 11–59). Munich, Germany: Lincom Europa.
- Grésillon, A. (1984). La règle et le monstre: Le mot-valise. Tübingen, Germany: Niemeyer.
- Gries, S. T. (2004a). Isn’t that fantabulous? How similarity motivates intentional morphological blends in English. In M. Achard & S. Kemmer (Eds.), Language, culture and mind (pp. 415–428). Stanford, CA: Center for the Study of Language and Information (CSLI).
- Gries, S. T. (2004b). Shouldn’t it be breakfunch? A quantitative analysis of blend structure in English. Linguistics, 42(3), 639–667.
- Gries, S. T. (2006). Cognitive determinants of subtractive word-formation processes: A corpus-based perspective. Linguistics, 17(4), 535–558.
- Gries, S. T. (2012). Quantitative corpus data on blend formation: Psycho- and cognitive-linguistic perspectives. In V. Renner, F. Maniez, & P. J. L. Arnaud (Eds.), Cross-disciplinary perspectives on lexical blending (pp. 145–167). Berlin, Germany: De Gruyter Mouton.
- Harley, T. A. (2001). The psychology of language: From data to theory (2nd ed.). Hove, U.K.: Psychology Press.
- Hrushcheva, O. A. (2011). Lingvokulturnye osobennosti blendinga. Vestnik Chelyabinskogo Gosudarstvennogo Universiteta, 218, 143–145.
- Hrushcheva, O. A. (2017). Blendy sovremennogo russkogo yazyka kak zerkalo kultury. Vestnik Orenburgskogo Gosudarstvennogo Universiteta, 202, 66–70.
- Kang, E. (2013). An Optimality-Theoretic analysis of lexical blends in Korean. Journal of Studies in Language, 28(4), 653–672.
- Kelly, M. H. (1998). To ‘brunch’ or to ‘brench’: Some aspects of blend structure. Linguistics, 36, 579–590.
- Kemmer, S. (2003). Schemas and lexical blends. In H. Cuyckens, T. Berg, R. Dirven, & K.‐U. Panther (Eds.), Motivation in language: From case grammar to cognitive linguistics. A Festschrift for Gunter Radden (pp. 69–97). Amsterdam, The Netherlands: Benjamins.
- Konieczna, E. (2012). Lexical blending in Polish. In V. Renner, F. Maniez, & P. J. L. Arnaud (Eds.), Cross-disciplinary perspectives on lexical blending (pp. 51–73). Berlin, Germany: De Gruyter Mouton.
- Kubozono, H. (1990). Phonological constraints on blending in English as a case for phonology-morphology interface. Yearbook of Morphology, 3, 1–20.
- Language Monitor.com. (2015). Top Words for the first 15 Years of the 21st century and what they portend.
- Lehrer, A. (1998). Scapes, holics, and thons: The semantics of combining forms. American Speech, 73, 3–28.
- Lepic, R. (2016). Lexical blends and lexical patterns in English and in American Sign Language. In J. Audring, F. Masini, & W. Sandler (Eds.), Online Proceedings of the Tenth Mediterranean Morphology Meeting MMM10 (Haifa) 7–10 September 2015 (pp. 98–111).
- López Rúa, P. (2002). On the structure of acronyms and neighbouring categories: A prototype-based account. English Language and Linguistics, 6(1), 31–60. doi:10.1017/S136067430200103X
- López Rúa, P. (2004). The categorial continuum of English blends. English Studies, 85(1), 63–76.
- López Rúa, P. (2012). Beyond all reasonable transgression: Lexical blending in alternative music. In V. Renner, F. Maniez, & P. J. L. Arnaud (Eds.), Cross-disciplinary perspectives on lexical blending (pp. 23–34). Berlin, Germany: De Gruyter Mouton.
- Mattiello, E. (2013). Extra-grammatical morphology in English: Abbreviations, blends, reduplicatives, and related phenomena. Berlin, Germany: Mouton de Gruyter.
- Meringer, R., & Mayer, C. (1978). Versprechen und Verlessen: Eine psychologisch-linguistische Studie. Amsterdam, The Netherlands: Benjamins.
- Moreton, E., Smith, J. L., Pertsova, K., Broad, R., & Prickett, B. (2017). Emergent positional privilege in novel English blends. Language, 93, 347–380. doi:10.1353/lan.2017.0017
- OED Online. (n.d.). Oxford English dictionary. Oxford University Press.
- Piñeros, C.‐E. (2004). The creation of portmanteaus in the extragrammatical morphology of Spanish. Probus, 16(2), 203–240.
- Plag, I. (2003). Word-formation in English. Cambridge, U.K.: Cambridge University Press.
- Pound, L. (1967). Blends: Their relation to English word formation. Amsterdam, The Netherlands: Swets & Zeitlinger.
- Ralli, A., & Xydopoulos, G. J. (2012). Blend formation in Modern Greek. In V. Renner, F. Maniez, & P. J. L. Arnaud (Eds.), Cross-disciplinary perspectives on lexical blending (pp. 35–50). Berlin, Germany: De Gruyter Mouton.
- Renner, V. (2006). Les composés coordinatifs en anglais contemporain (Doctoral thesis). Université Lumière-Lyon 2, Lyon.
- Renner, V. (2015a). Lexical blending as wordplay. In A. Zirker & E. Winter-Froemel (Eds.), Wordplay and metalinguistic / metadiscursive reflection: Authors, contexts, techniques, and meta-reflection (pp. 119–133). Berlin, Germany: De Gruyter Mouton.
- Renner, V. (2015b). Panorama rétro-prospectif des études amalgamatives. Neologica : Revue Internationale de La Néologie, 9, 97–112.
- Renner, V., Maniez, F., & Arnaud, P. J. L. (2012). Introduction: A bird’s eye view of lexical blending. In V. Renner, F. Maniez, & P. J. L. Arnaud (Eds.), Cross-disciplinary perspectives on lexical blending (pp. 1–9). Berlin, Germany: De Gruyter Mouton.
- Roig-Marín, A. (2017). The Anglicisation of word formations: Cross-linguistic blends in Spanish: What do they tell us about the state of English as an international language?. English Today, 33(4), 1–7. doi:10.1017/S0266078416000614
- Ronneberger-Sibold, E. (2012). Blending between grammar and universal cognitive principles: Evidence from German, Farsi, and Chinese. In V. Renner, F. Maniez, & P. J. L. Arnaud (Eds.), Cross-disciplinary perspectives on lexical blending (pp. 115–143). Berlin, Germany: De Gruyter Mouton.
- Shaw, K. E., White, A. M., Moreton, E., & Monrose, F. (2014). Emergent faithfulness to morphological and semantic heads in lexical blends. In J. Kingston, C. Moore-Cantwell, J. Pater, & R. Staubs (Eds.), Proceedings of Phonology 2013. The Linguistic Society of America.
- Štekauer, P., Valera, S., & Kőrtvélyessy, L. (2012). Word-formation in the world’s languages: A typological survey. Cambridge, U.K.: Cambridge University Press.
- Thurner, D. (1993). Portmanteau dictionary: Blend words in English language, including trademarks and brand names. Ann Arbor, MI: McFarland.
- Tomaszewicz, E. (2012). Output-to-output faithfulness in the phonological structure of English blends. In V. Renner, F. Maniez, & P. J. L. Arnaud (Eds.), Cross-disciplinary perspectives on lexical blending (pp. 213–232). Berlin, Germany: De Gruyter Mouton.
- Urban Dictionary. (n.d.). Thankshallowistmas.
- White, S. J., Johnson, R. L., Liversedge, S. P., & Rayner, K. (2008). Eye movements when reading transposed text: The importance of word-beginning letters. Journal of Experimental Psychology: Human Perception and Performance, 34(5), 1261–1276. doi:10.1037/0096-1518.104.22.1681
- Whitney, C. (2001). How the brain encodes the order of letters in a printed word: The SERIOL model and selective literature review. Psychonomic Bulletin and Review, 8(2), 221–243.
- Wood, F. A. (1911). Iteratives, blends, and ‘streckformen’. Modern Philology, 9(2), 157–194.