85 Tina Grlj: BLENDING AS A WORD-FORMATION PROCESS: A COMPARATIVE ANALYSIS ... Tina Grlj UDK 81'373.611:[811.111+811.133.1] tina.grlj@gmail.com DOI: 10.4312/vestnik.14.85-106 Izvirni znanstveni članek BLENDING AS A WORD-FORMATION PROCESS: A COMPARATIVE ANALYSIS OF BLENDS IN ENGLISH AND FRENCH 1 INTRODUCTION Every word is once new, a neologism. It is perceived as new until it has been used or heard enough times to lose this air of novelty. Although there are several ways of creat- ing new words, the present article is concerned with the morphological process referred to as lexical blending. This is a minor word-formation process that is used to coin new words by joining together two or more source words, at least one of which is shortened. This process is found in many languages, and is clear proof of how inventive a language can be. The present article investigates English and French blend words. The first part of the article is an overview of defining characteristics of blending and blend words. The second half of the article brings an analysis of the corpus examples collected for this particular purpose. Some of the characteristics of blends that are put forward in various studies conducted by English and French linguists are tested on the corpus. The article also investigates the differences that appear between English and French blends in order to find out if there are any major discrepancies between the two languages. 2 DEFINING CHARACTERISTICS OF BLENDING AND BLEND WORDS The examination of numerous studies of amalgamation in English and French suggests that, although obvious differences exist, some prototypical features of blend words ap- pear repeatedly. As mentioned in the introduction, the majority of lexicologists agree that a blend word is the result of the fusion of two or more source words, at least one of which is shortened or truncated. The structure that is mentioned as the most typical is the front part of the first source word and the last part of the second source word (e.g. Eng. Vestnik_za_tuje_jezike_2022_FINAL.indd 85 Vestnik_za_tuje_jezike_2022_FINAL.indd 85 24. 01. 2023 09:18:44 24. 01. 2023 09:18:44 86 VESTNIK ZA TUJE JEZIKE/JOURNAL FOR FOREIGN LANGUAGES chofa < chair + sofa, Fr. gélule < gélatine + capsule 1 ) with overlap in the middle of the blend (e.g. Eng. babymoon < baby + honeymoon, Fr. franglais < français + anglais). In addition, a number of linguists add that amalgamation may include more than two source words (e.g. Eng. Xrunkopy < Xerox + drunk + photocopy, Fr. almasilicium < aluminium + magnésium + silicium), and that it is also possible that the blend word is composed of two source words shortened to their front parts (e.g. Eng. modem < modulator + de- modulator, Fr. courriel < courrier + électronique 2 ) or that a source word is infixed in the other (e.g. Eng. Sleavenia < Slovenia + leave, Fr. rajolivissant < ravissant + joli 3 ). The following table shows the typical characteristics of blends described in various publications, as listed in the bibliography section, and checks how many authors include each individual characteristic in their definitions. Table 1: Defining characteristics of blends according to different authors   Involves two SW 4 Involves more than two SW Front part of SW1 and last part of SW2 Front part of SW1 and front part of SW2 Shorten- ing of at least one SW Overlap of SW1 and SW2 Embed- ding of one SW into the other Pound (2015) +   + + +     Adams (2001) +       + +   Algeo (1977) + +     + + + Bauer (2012) + + + + + + + Cannon (2009) +   +   +     Lehrer (1996) +       +   + Plag (2003) + + + +     López Rúa (2004) + + + + + +   1 Eng. “gelatin” + “capsule”. 2 Eng. “email” + “electronic”. 3 Eng. “delightful” + “pretty”. 4 SW stands for source word. Vestnik_za_tuje_jezike_2022_FINAL.indd 86 Vestnik_za_tuje_jezike_2022_FINAL.indd 86 24. 01. 2023 09:18:44 24. 01. 2023 09:18:44 87 Tina Grlj: BLENDING AS A WORD-FORMATION PROCESS: A COMPARATIVE ANALYSIS ... Involves two SW 4 Involves more than two SW Front part of SW1 and last part of SW2 Front part of SW1 and front part of SW2 Shorten- ing of at least one SW Overlap of SW1 and SW2 Embed- ding of one SW into the other Gries (2004) + + +   + +   Brdar- Szabó & Brdar (2008) +   +   + +   Clas (1987) + + + + + + + Fradin (2000) + +     + + + Renner (2012) +   +   + +   Léturgie (2012) +   +   + + + TOTAL 14 7 10 4 14 10 5 If we summarize the information gathered in the table, we can construct the follow- ing prototypical definition of blending: Blending is a word-formation process that involves two or more source words. It involves the shortening of at least one source word, but frequently the source words dis- play some overlap. The most typical pattern of blending is the front part of the first source word and the last part of the second source word. 3 CORPUS ANALYSIS 3.1 Compiling the corpus The present article attempts to provide a clear description of significant structural proper- ties that dominate the process of blend formation in English and French on the basis of a corpus that has been compiled for this particular purpose. Blends do not exhibit a single, clear-cut rule of formation, but, hopefully, the analysis of the corpus will reveal some prominent structural patterns in both languages. Vestnik_za_tuje_jezike_2022_FINAL.indd 87 Vestnik_za_tuje_jezike_2022_FINAL.indd 87 24. 01. 2023 09:18:44 24. 01. 2023 09:18:44 88 VESTNIK ZA TUJE JEZIKE/JOURNAL FOR FOREIGN LANGUAGES The only way to get meaningful results, which would reveal any significant differ- ences between English and French blends, is to analyse blends gathered from similar sources. Therefore, the vast majority of lexical data forming the corpus of French and English blends was gathered from articles and studies included in the bibliography. Addi- tionally, some entries were found in opportunistic sources, such as contemporary Ameri- can and British online newspapers and magazines and online collections. 5 The corpus is limited to single-word blends only, excluding multi-word blends such as store d’oeuvre, created by blending store and hors d’oeuvre. Additionally, only blends with two source words are included in the corpus, eliminating entries coined from three source words. The most notable limitation, however, is the exclusion of words created by joining together the front parts of two or more source words (e.g. modem < modulator + demodu- lator). These coinages are considered by some authors as a sub-type of acronyms or com- plex clippings rather than blends. Both Renner (2006, s. p.) and Léturgie (2011b, 204) claim that a blend is composed of source words which are shortened at their inner edges (meaning that the last part of the first source word and the first part of the second source word are removed). Arndt-Lappe and Plag (2013, 5) classify modem as a representative of complex clippings that have the following structure: AC = AB + CD. They state that “AC formations are therefore often treated as a pattern distinct from blending”. On the other hand, the corpus includes examples where one source word is embedded into the other. As mentioned in section 2 on definitions of blends, Bauer (2012, 17) claims that while these “infixed blends” are rare in English, they are more frequent in French. The final corpus contains 458 English and 396 French blends. 3.2 Corpus analysis In the present article, English and French blends are analysed separately in order to see if any major differences can be found between blends from both languages. The analysis is limited to the following structural parameters: • The lexical categories of blends and of their source words entering each blend. • The presence or absence of shortening of source words: some blends are formed by juxtaposing two source words in their entirety with overlap in the middle, while oth- ers show different degrees of shortening of either one or both source words. • The type of shortening: which source word is shortened in the process of blending – first, second or both. • The structural patterns of blends: how are source words combined to form blends. 5 For example Word Spy https://wordspy.com/, Urban Dictionary https://www.urbandictionary.com/, Pinterest https://www.pinterest.com/. Vestnik_za_tuje_jezike_2022_FINAL.indd 88 Vestnik_za_tuje_jezike_2022_FINAL.indd 88 24. 01. 2023 09:18:44 24. 01. 2023 09:18:44 89 Tina Grlj: BLENDING AS A WORD-FORMATION PROCESS: A COMPARATIVE ANALYSIS ... • The presence or absence of overlap between the source words: the overlap may be perfect (with both source words present in their entirety) or only partial. • The type of overlap: the overlap may be orthographic, phonic or both. • Infixation: sometimes one source word is inserted into the other. The parameters listed above are corroborated by the statistical analysis of the cor- pus. Additionally, some parts of the analysis are compared to the findings published by Renner in “French and English lexical blends in contrast” (2018). His study is based on 97 French and 374 English blends from two corresponding dictionaries in both lan- guages, namely the Grand Robert de la langue française and the Shorter Oxford English Dictionary. Both his study and the present article analyse the lexical categories of blends and their source words, structural patterns of blends and type of overlap between source words. Upon closer examination, however, some interesting discrepancies emerge. 3.2.1 Lexical categories In English as well as in French, the majority of blends are nominal, followed in quantity by adjectival, then verbal and adverbial blends. The corpus also includes one English pronominal blend (shim 6 ). As the graph below demonstrates, the percentage of each of these lexical categories is almost identical in both languages. Nouns represent 85.4% of the English part of the corpus and 84.8% of the French part. Chart 1: Lexical categories of English and French blends Chart 2 shows how frequent each combination of lexical categories of source words is. 6 In Renner’s study shim (< she + him) is not considered a pronominal blend but rather a nominal one, even if he analysed it as being formed from two pronouns. 85.4% 9.6% 4.4% 0.4% 0.2% 84.8% 7.3% 7.1% 0.8% 0 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% NOUN ADJECTIVE VERB ADVERB PRONOUN ENGLISH BLENDS FRENCH BLENDS Vestnik_za_tuje_jezike_2022_FINAL.indd 89 Vestnik_za_tuje_jezike_2022_FINAL.indd 89 24. 01. 2023 09:18:45 24. 01. 2023 09:18:45 90 VESTNIK ZA TUJE JEZIKE/JOURNAL FOR FOREIGN LANGUAGES Chart 2: Comparison of nominal blends in both languages The most frequent combinations are the juxtaposition of two nouns (e.g. Eng. archi- tourist < architecture + tourist, Fr. confipote < confiture + compote), a noun and an adjec- tive (e.g. Eng. fakeation < fake + vacation, Fr. sublimitude < sublime + attitude) or a noun and a verb (e.g. Eng. shareware < share + software, Fr. Cracotte < craquer + biscotte). These three combinations appear in both languages. On the other hand, the English part of the corpus displays four additional combinations, those of a noun and an adverb (e.g. backronym < back + acronym), a noun and an acronym (e.g. waplash < W AP + backlash), a noun and a pronoun (e.g. himbo < him + bimbo) or a noun and an interjection (e.g. grrrl < grrr + girl). These four combinations, however, are only present in the English part of the corpus and they are a minority (their frequency is below 1%). Chart 3 demonstrates different combinations of source words in adjectival blends and their respective frequencies. Chart 3: Comparison of adjectival blends in both languages 79.3% 14.0% 4.6% 0.8% 0.5% 0.5% 0.3% 78.3% 19.3% 2.4% 0 0 0 0 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% ENGLISH BLENDS FRENCH BLENDS 70.4% 27.3% 2.3% 62.1% 37.9% 0 0% 10% 20% 30% 40% 50% 60% 70% 80% ADJECTIVE + ADJECTIVE ADJECTIVE + NOUN ADJECTIVE + VERB ENGLISH BLENDS FRENCH BLENDS Vestnik_za_tuje_jezike_2022_FINAL.indd 90 Vestnik_za_tuje_jezike_2022_FINAL.indd 90 24. 01. 2023 09:18:45 24. 01. 2023 09:18:45 91 Tina Grlj: BLENDING AS A WORD-FORMATION PROCESS: A COMPARATIVE ANALYSIS ... The list of lexical data contains 44 English and 29 French adjectives, which represent 9.8% and 7.3% of each part of the corpus, respectively. This quantity does not suffice to deduce any definitive conclusions about adjectival blends in general, but it seems safe to assume that the majority of them are formed from two adjectives (e.g. Eng. fantabulous < fantastic + fabulous, Fr. spûre < spontanée + sûre 7 ). The second most frequent com- bination attested in both languages is that of an adjective and a noun (e.g. Eng. lugged < luggage + mugged, Fr. ennuiversel < ennui + universel 8 ). Additionally, only one English blend is constructed from an adjective and a verb (e.g. brusherific < brush + terrific), while no examples with this type of structure are found in French. Chart 4 illustrates the frequency of each combination of source words in adjectival blends. Chart 4: Comparison of verbal blends in both languages Twenty English and 28 French verbal blends make up 4.4% and 7.1% of each part of the corpus, respectively. Again, this amount is not sufficient to draw any conclusions about the entire category. However, as the chart below demonstrates, the majority of ver- bal blends combine two verbs (e.g. Eng. chillax < chill + relax, Fr. pleiger < pleuvoir + neiger 9 ). The juxtaposition of a verb and a noun is found in both languages, but it is much more frequent in French than in English (e.g. Eng. prowebstinate < procrastinate + web, Fr. clavarder < clavier + bavarder 10 ). Among the English lexical data, the combinations of a verb and an adjective (e.g. narrowcast < narrow + broadcast) or a verb and an adverb (e.g. gazunder < gazump + under) are also present. 7 Eng. “spontaneous” + “safe”. 8 Eng. “boredom” + “universal”. 9 Eng. “to rain” + “to snow”. 10 Eng. “keyboard” + “to chat”. 70% 15% 10% 5% 60.7% 39.3% 0 0 0% 10% 20% 30% 40% 50% 60% 70% 80% VERB + VERB VERB + NOUN VERB + ADJECTIVE VERB + ADVERB ENGLISH BLENDS FRENCH BLENDS Vestnik_za_tuje_jezike_2022_FINAL.indd 91 Vestnik_za_tuje_jezike_2022_FINAL.indd 91 24. 01. 2023 09:18:45 24. 01. 2023 09:18:45 92 VESTNIK ZA TUJE JEZIKE/JOURNAL FOR FOREIGN LANGUAGES Adverbial blends also exist, and the corpus contains two in English (e.g. absolutively < absolutely + positively) and three in French (e.g. intelligentiment < intelligemment + gentiment 11 ). As mentioned at the beginning of this section, shim is the only pronominal blend in English that made its way into the corpus. It is a combination of the pronouns she and him. As far as lexical categories of blends are concerned, Renner’s study (2018) presents almost identical results to the analysis above, apart from the fact that Renner did not find any adverbial or pronominal blends. However, this difference is of little importance as these two lexical categories represent a minority. Table 2: Percentage of each lexical category of blends revealed by the present analysis and Renner’s study LEXICAL CATEGORIES OF BLENDS % OF ENGLISH BLENDS % OF ENGLISH BLENDS (RENNER) % OF FRENCH BLENDS % OF FRENCH BLENDS (RENNER) NOMINAL 84.8% 89% 85.4% 88% ADJECTIVAL 7.3% 8% 9.6% 8% VERBAL 7.1% 3% 4.4% 4% ADVERBIAL 0.4% 0 0.8% 0 PRONOMINAL 0.2% 0 0 0 A notable difference between the two studies can be observed in the number of possible combinations of lexical categories of source words. While the present analysis discovered 13 possible combinations, Renner discovered only eight. 12 The table below demonstrates the differences between the findings of both studies. 11 Eng. “intelligently” + “nicely”. 12 Renner’s analysis actually reveals 10 combinations, but in order to simplify the comparative analysis, I joined com - binations ‘noun + adjective’ and ‘adjective + noun’ into one category. The same goes for ‘verb + noun’ and ‘noun + verb’. Vestnik_za_tuje_jezike_2022_FINAL.indd 92 Vestnik_za_tuje_jezike_2022_FINAL.indd 92 24. 01. 2023 09:18:45 24. 01. 2023 09:18:45 93 Tina Grlj: BLENDING AS A WORD-FORMATION PROCESS: A COMPARATIVE ANALYSIS ... Table 3: Percentage of each combination of source words in both studies COMBINATIONS OF LEXICAL CATEGORIES OF SOURCE WORDS  % OF ENGLISH BLENDS % OF ENGLISH BLENDS (RENNER) % OF FRENCH BLENDS % OF FRENCH BLENDS (RENNER) NOUN + NOUN 67.7% 75.0% 66.4% 74.5% NOUN + ADJECTIVE 14.6% 21.7% 19.2% 24.5% ADJECTIVE + ADJECTIVE 6.8% 0.6% 4.5% 1% NOUN + VERB 4.6% 1.2% 4.8% 0 VERB + VERB 3.1% 0 4.3% 0 NOUN + ADVERB 0.7% 0.3% 0.3% 0 ADJECTIVE + VERB 0.7% 0 0 0 NOUN + ACRONYM 0.4% 0 0 0 NOUN + PRONOUN 0.4% 0.6% 0 0 ADVERB + ADVERB 0.4% 0 0.5% 0 NOUN + INTERJECTION 0.2% 0.3% 0 0 VERB + ADVERB 0.2% 0 0 0 PRONOUN + PRONOUN 0.2% 0.3% 0 0 In both analyses English blends display a greater number of possible combinations than French blends. In Renner’s corpus French blends are only found with the following three combinations: two nouns, a noun combined with an adjective or two adjectives. The corpus assembled for the present article also includes combinations of a noun and a verb, two verbs, a noun and an adverb or two adverbs. As Renner (2018, 6) also pointed out, this seems to indicate that, if compared to French, there is “[...] a higher degree of word playfulness of blending in English than in French.” Moreover, if we disregard the smaller discrepancies discovered in connection with the combinations that are in the minority, three major differences stand out. In contrast with Renner’s list of blends, the present corpus includes examples of blends that couple two adjectives, two verbs and even a noun and a verb, which are marginal or even non- existent in the earlier corpus. Vestnik_za_tuje_jezike_2022_FINAL.indd 93 Vestnik_za_tuje_jezike_2022_FINAL.indd 93 24. 01. 2023 09:18:45 24. 01. 2023 09:18:45 94 VESTNIK ZA TUJE JEZIKE/JOURNAL FOR FOREIGN LANGUAGES 3.2.2 Shortening of source words The analysis shows that blends from the corpus display one of the following three structures: • one source word is shortened, the other one remains intact (e.g. Eng. Dixiecrat < Dixie + democrat, Fr. jaguarion < jaguar + lion 13 ), • both source words are shortened (e.g. Eng. Chimerica < China + America, Fr. plapi- er < plastique + papier 14 ) or • both source words are present in their entirety, with medial overlap (e.g. Eng. cuten- sil < cute + utensil, Fr. féconductrice < fécond + conductrice 15 ). The chart below shows the percentage of each type of shortening in both languages. Chart 5: Number of source words shortened during the formation of blends in both languages About 50% of all the words in the corpus include shortening of one source word, whereas the other two types of shortening display a difference between the two languag- es. Shortening of both source words is more frequent in English (38.4% versus 25% in French). On the other hand, in comparison to English, French favours keeping both source words intact (25.3% versus 9.2% in English). 3.2.3 Structural patterns of blends As has already been pointed out, blends show a variety of structural patterns. The source words can be shortened or not, they can be juxtaposed or inserted or they can even be kept whole and overlap in the middle part of the blend they form. 13 Eng. “jaguar” + “lion”. 14 Eng. “plastic” + “paper”. 15 Eng. “fertile” + “forewoman”. 52.4% 38.4% 9.2% 49.7% 25% 25.3% 0% 10% 20% 30% 40% 50% 60% Shortening of one SW Shortening of both SW Both SW in their entirety ENGLISH BLENDS FRENCH BLENDS Vestnik_za_tuje_jezike_2022_FINAL.indd 94 Vestnik_za_tuje_jezike_2022_FINAL.indd 94 24. 01. 2023 09:18:46 24. 01. 2023 09:18:46 95 Tina Grlj: BLENDING AS A WORD-FORMATION PROCESS: A COMPARATIVE ANALYSIS ... Both English and French blends show the following five structural patterns: • The first part of the first source word is in juxtaposition with the last part of the sec- ond source word (e.g. Eng. chofa < chair + sofa, Fr. gélule < gélatine + capsule); • The whole first source word is followed by the last part of the second source word (e.g. Eng. furminator < fur + terminator, Fr. kebabisation < kebab + islamisation); • The first part of the first source word is in juxtaposition with the entire second source word (e.g. Eng. relationblip < relationship + blip, Fr. catapostrophe < catastrophe + apostrophe); • None of the source words is shortened, but they overlap in the middle (e.g. Eng. Ya- hooligan < Yahoo + hooligan, Fr. animalphabet < animal + alphabet) or • The second source word is embedded into the first one (e.g. Eng. Armachillo < ar- madillo + chill, Fr. s’embellemerder < s’emmerder + belle-mère 16 ). The chart below demonstrates the difference in the percentage of each structural pat- tern in both languages. Chart 6: Percentage of each structural pattern in both languages This section is the first that reveals quite a few notable differences between both lan - guages. The majority of the English blends from the corpus are composed from the front part of the first source word and the last part of the second source word (38.4% of the Eng - lish versus 25.3% of the French blends), while the majority of French blends are formed by juxtaposing the front part of the first source word and the second source word in its entirety (30% of the French versus 19% of the English blends). The entire first source word fol - lowed by the last part of the second source word is also a popular structural pattern in both 16 Eng. “to be bored” + “mother-in-law”. 38.4% 29.9% 19% 9.2% 3.5% 25.3% 17.7% 30% 22.7% 4.3% 0% 5% 10% 15% 20% 25% 30% 35% 40% 45% Front part of SW1 and last part of SW2 Whole SW1 and hind part of SW2 Front part of SW1 and whole SW2 Whole SW1 and whole SW2 Embedding of SW2 into SW1 ENGLISH BLENDS FRENCH BLENDS Vestnik_za_tuje_jezike_2022_FINAL.indd 95 Vestnik_za_tuje_jezike_2022_FINAL.indd 95 24. 01. 2023 09:18:46 24. 01. 2023 09:18:46 96 VESTNIK ZA TUJE JEZIKE/JOURNAL FOR FOREIGN LANGUAGES languages, but there is a significant difference in the percentage of blends with this pattern, namely, 29.9% of the English blends and 17.7% of the French blends are formed this way. Another dissimilarity between both languages is that French has more than twice as many blends composed from two complete source words if compared to English, namely 22.7% in French versus 9.2% in English. As far as infixed blends go, the percentage in both lan - guages is comparable (4.3% in French and 3.5% in English). Renner (2018) also investigated which structural patterns appear in English and French blends, and the following table shows the differences between the results of both studies. Table 4: Percentage of each type of blend structure in the present article and Renner’s study BLEND STRUCTURE  % OF ENGLISH BLENDS % OF ENGLISH BLENDS (RENNER) % OF FRENCH BLENDS % OF FRENCH BLENDS (RENNER) Front part of SW1 and last part of SW2 (double inner shortening 17 ) 38.4% 31% 25.3% 31% Whole SW1 and last part of SW2 (right-hand-side inner shortening) 29.9% 21% 17.7% 8.5% Front part of SW1 and whole SW2 (left-hand-side inner shortening) 19% 24% 30% 44.5% Whole SW1 and whole SW2 (haplologic blending) 9.2% 7% 22.7% 3% Embedding of SW2 into SW1 (sandwich blending) 3.5% 1% 4.3% 0 Front part of SW1 and front part of SW2 (double right-shortening) 0 14% 0 12% Other 0 2% 0 1% 17 The names of patterns of lexical shortening in brackets are from Renner’s study (2018, 6-7). Vestnik_za_tuje_jezike_2022_FINAL.indd 96 Vestnik_za_tuje_jezike_2022_FINAL.indd 96 24. 01. 2023 09:18:46 24. 01. 2023 09:18:46 97 Tina Grlj: BLENDING AS A WORD-FORMATION PROCESS: A COMPARATIVE ANALYSIS ... Firstly, Renner’s corpus contains blends which are formed from the front parts of source words, which the present article excludes. Secondly, Renner also found a small percentage of words which display other structures, as an example he lists audimat, which is coined from the first part of audimètre and the middle part of automatique. The present article excludes such formations from the category of blends, following Arndt-Lappe and Plag (2013, 5) who classify these coinages as complex clippings that have the following structure: AC = AB + CD. They state that “AC formations are therefore often treated as a pattern distinct from blending”. Secondly, both lists of blends show that the majority of English blends are formed by connecting the front part of the first source word and last part of the second source word, while French blends clearly prefer the juxtaposition of the front part of the first source word and the entire second source word. As far as the frequency of each structural pattern is concerned, the only notable difference is that the second and third places in the English part of both corpora are reversed, namely Renner’s second most frequent blend structure is the front part of the first source word followed by the entire second source word, while the third place is taken by blends coined from the entire first source word and the last part of the second source word. The analysis conducted in the present article reveals that these two places are switched in frequency. Blends that keep both source words in their entirety with overlap in the middle (e.g. Eng. affluenza < affluence + influenza, Fr. déceptionniste < déception + réceptionniste 18 ) are clearly much more frequent in the French part of the present corpus than in Renner’s, which seems strange as some linguists dub this type of coinage as the ideal blend (e.g. Kaunisto 2000, n. pag.). It is possible that such blends started gaining in popularity in recent years and have therefore not been included in traditional dictionaries, which are the basis of Renner’s corpus. The same might apply to infixed blends (e.g. Eng. Sleavenia < Slovenia + leave, Fr. Dékafkaïné < décaféiné + Kafka 19 ), where the second source word in embedded into the first one. The present corpus includes a slightly higher number of such blends than Renner’s. 3.2.4 Overlap The present section focuses on an analysis of overlap that appears in some blends and it is the second area of analysis that revealed some notable differences between both lan- guages. Before looking closely at these differences, it needs to be pointed out that even if overlap is frequently mentioned in studies on blends, it is rarely specified which type of overlap is meant. The present analysis is only interested in medial overlap, namely the string of letters or phonemes that are present at the inner edges where both source words are joined to form a blend (e.g. skinship < skin + kinship). 18 Eng. “disappointment” + “receptionist”. 19 Eng. “decaffeinated” + “Kafka”. Vestnik_za_tuje_jezike_2022_FINAL.indd 97 Vestnik_za_tuje_jezike_2022_FINAL.indd 97 24. 01. 2023 09:18:46 24. 01. 2023 09:18:46 98 VESTNIK ZA TUJE JEZIKE/JOURNAL FOR FOREIGN LANGUAGES Chart 7 presents a comparison of different types of overlap in English and French. Chart 7: Comparison of types of overlap between English and French In English, as well as in French, the majority of blends show both orthographic and phonemic overlap, but the percentage of such blends is significantly higher in the French part of the corpus (67.7% versus 47.7% of English blends). On the other hand, the percentage of blends without overlap is lower in French (25.8% versus 42.6% in the English part of the corpus). The rest of the lexical data show either orthographic (7.2% of English versus 4.5% of French blends), or phonemic overlap (2.8% of English versus 2% of French blends). The following example is very interesting as far as overlap is concerned: hangry < hungry + angry. Both source words have over 65% of letters in common, however, the first source word only contributes the initial letter h, and since the second source word does not have the same letter at its left edge, there is no medial overlap. Additionally, various types of medial overlap can occur, the source words can have orthographic (e.g. Eng. smog < smoke + fog, Fr. chiantifique < chiant + scientifique 20 ), phonemic (e.g. Eng. ballute < balloon + parachute, Fr. jeansmnastique < jeans + gym- nastique 21 ) or both orthographic and phonemic overlap (e.g. Eng. babymoon < baby + honeymoon, Fr. picoléreux < picoler + coléreux 22 ). What needs to be underlined at this point is that even if a blend shows orthographic and phonemic overlap of its source words, this does not necessarily entail that the overlap is perfect. To illustrate, from the orthographic standpoint croissandwich (< croissant + sandwich) overlaps in three letters, but looking at this blend from the phonemic angle, the overlap is not perfect because the source words overlap in only one phoneme (/krəˈsɑːnt/ + /ˈsænwɪdʒ/). The same holds for the French example créatique (< création [kʀeasjɔ̃] + informatique [ɛ̃fɔʀmatik]). 20 Eng. “boring” + “scientific”. 21 Eng. “jeans” + “gymnastics”. 22 Eng. “to drink” + “angry person”. 47.4% 42.6% 7.2% 2.8% 67.7% 25.8% 4.5% 2% 0% 10% 20% 30% 40% 50% 60% 70% 80% ORTHOGRAPHIC AND PHONEMIC NO OVERLAP ORTHOGRAPHIC PHONEMIC ENGLISH BLENDS FRENCH BLENDS Vestnik_za_tuje_jezike_2022_FINAL.indd 98 Vestnik_za_tuje_jezike_2022_FINAL.indd 98 24. 01. 2023 09:18:46 24. 01. 2023 09:18:46 99 Tina Grlj: BLENDING AS A WORD-FORMATION PROCESS: A COMPARATIVE ANALYSIS ... Before we examine overlap further, it needs to be stated that some blends do not display any overlap (e.g. Eng. procrastibaking < procrastination + baking, Fr. paponcle < papa + oncle 23 ). The two most interesting examples found in the corpus are clandestiny (< clan- destine /ˈklændəstaɪn/ + destiny /ˈdestəni/) and aiguillotine < (aiguille [egɥij] + guillo- tine [gijɔtin] 24 ). The former overlaps in six letters, but only one phoneme, while the latter overlaps in five letters but only one phoneme. In contrast, cashmiracle (< cashmere /ˈkæʒmɪr/ + miracle /ˈmɪrəkl/) has a perfect phonemic overlap and an incomplete orthographic overlap. In French, télépholie (